John

Harrisons
Contrivance
John Harrisons
Contrivance
Stuart Harrison
And now, if the Royal Society please, I will shew them the
Draught of the Clock which I have in great Part made, and not only the Draught of
the Pallats, as in particular, but also the Pallats themselves, in order that they may see
at least some Reason for what I found, or might as in Consequence find from such a
Contrivance of Pallats
John Harrison, Concerning Such Mechanism, 1775, p.55
This allegorical picture, engraved by Albrecht Drer in 1514 but not signed, depicts
the antagonism between the scientific method and the dogma of the Roman
Catholic Church. The influence of the Italian Renaissance swept through Europe,
introducing new concepts in architecture and the arts. The ideal of the Renaissance
man, embracing a view of the world through a synthesis of science and art, is
exemplified here. This philosophy culminated in the formation of the Royal Society
during the Restoration enabling further acceptance and free promulgation of
scientific ideas. During the eighteenth century it became known as the
enlightenment following Immanuel Kants 1784 essay Beantwortung der Frage: Was ist
Aufklrung (Answering the Question: What is Enlightenment). Natural philosophers
developed mechanics which led to the Industrial Revolution, as exemplified by John
Harrisons work.
Published by Fionchra Press, Cedar House, ORFORD, Suffolk IP12 2NQ
Typesetting and origination by Fionchra Press
Printed by Copytech (UK) Ltd.,
Orton Southgate, PETERBOROUGH PE2 6XD
ISBN 978-184426-773-6
Contents
Cottinghams 1909 Photographs iv
Preface vii
Chapter 1 John Harrisons Astronomical Regulator
Introduction 1
Tobias Mayer 7
Martin Folkes 11
Dr William Stukeley 15
Act for Commencement of Year 16
Thomas Coram 17
Analysis of Beauty 18
Chapter 2 Documentary Sources
Portraits 19
A Philosopher giving that Lecture 20
Dr William Heberden 24
Manuscripts and Pamphlets 24
Lionel Charlton 28
Nicholas Saunderson 30
Nautical Astronomy 31
Geometry used by Harrison 32
Chapter 3 Development of the Pendulum Clocks
Introduction 33
1713 Longcase Clock 34
1717 Longcase Clock 34
1722 Brocklesby Hall Turret Clock 35
1725-1730 Longcase Clocks 37
The Marine Clocks 40
Jost Burgi 44
Jeffreys Pocket Watch 46
Admiral John Campbell 47
The Marine Watch (H4) 1757-1761 47
RAS Regulator 1750-57 49
Robert Hooke 50
Chapter 4 Royal Astronomical Society Regulator
History of the RAS Regulator 51
The Analytical Society 51
i
Description of the RAS Regulator 54
Materials Used by Harrison 56
Wire and Plate Gauges 60
Restorations of RAS Regulator 60
Unfinished and Changed Features 64
R.T.Goulds Report to RAS 65
John Harrisons Measurement Standards 69
R.T.Goulds 1927 Case 70
Chapter 5 Logical Design
Development of Form 71
Rev. Edmund Gunter 75
Geometrical Conception 76
Marking Out 80
English Bellringing 82
Chapter 6 Oscillator
Introduction 83
Galileo Galilei 85
Cycloid 87
Pendulum Analysis 93
Pendulum Damping 107
Escapement Error 109
J. L. Lagrange 112
Temperature Compensation 113
Ole Christensen Rmer 117
J. B. Joseph Fourier 119
Joseph Black 121
John Shelton 123
Pendulum Thermal Compensation 124
Pendulum Suspension 126
Suspension Spring 129
Ageing of Pendulum 135
Rev. Abraham Rees 137
Harrisons Clock Escapement 138
RAS Regulator Escapement 139
Description of the RAS Escapement 144
Analysis of Harrisons Drawings 146
Escapement Drawing for RAS Regulator 149
Action of Escapement 151
Determining Pallet Arm Lengths 157
Geometrical Construction 157
Algebraic Calculation 162
ii
Escapement Error 177
Sir George Biddell Airy 184
George Grahams Escapement 185
John Sheltons Regulators 186
Chapter 7 Power Supply
Bearings and Friction Wheels 187
Roller Bearing Metal 191
Driving Weight and Pulley 191
Wheels and Maintaining Power 193
Dr. John Arbuthnot 200
Remontoire 201
Geometrical Design 206
Algebraic Method 208
Chapter 8 Integrator
Seconds 215
Motion Work 215
Day Wheel 216
Case 218
Chapter 9 Adjustment and Errors 219
Chapter 10 Conclusions 227
Chapter 11 Drawings and Construction
Introduction 231
Recorded Measurements of RAS Regulator 232
Plates 238
Friction Wheels and Roller Bearings 257
Wheels 260
Escapement 267
Remontoire 273
Pendulum 279
Pendulum Suspension Block 280
Dials and Motion Work 283
Driving Weight and Pulley 293
1728 Movement 294
Appendices
A. Subscribers to Isaac Wares Book 297
B. Musical Temperament 301

Bibliography and References 307
iii
Preface
Some Years ago I made several alterations in order to render the
Motion of Clocks more exact than heretofore, but when I came to try them by
strict observation as below, I judgd the best performance of the best Pendulum
Clock I ever saw, made, or heard of, to be incapable of this Matter, woud it go
as well in a Ship at Sea in any part of the World, as in any one fixd place on
the Land. Yet from several observations, I still endeavourd to make further
Corrections in this Motion; and in these 3 last Years have brought a Clock to
go nearer the truth than can be well imagind, considering the vast Number of
seconds of Time there is in a Month, in which space of time it does not vary
above one second, and that mostly the way I expect: So I am sure I can bring
it to the Nicety of 2 or 3 seconds in a Year. And twill also continue this
exactness for 40 or 50 Years or more; however so as not to vary above 2 or 3
seconds from what it did the Year next before; for twill not want Cleaning,
and the little it wears can but alter it insensibly little. This Nicety is owing
partly to the Matter the Clock is made of, partly to the Contrivance it is made
with, and partly to the Nice observations it is tryd by, and the convenient place
it stands in. These I will now treat of, and show how I think their application,
with some further Addition, may compose a Motion to go nearly as true in a
Ship at Sea.
John Harrison, 1730 Manuscript
When John Harrison, Inventor of the Timekeeper for the Longitude at
Sea, first met Edmond Halley he would have emphasized that no
pendulum clock made at that time other than his own could reliably
perform to the accuracy required to fulfil the conditions of the 1714
Queen Anne Act of Longitude on land let alone at sea. As Newton
stated, finding longitude at sea was dependent upon knowing a time
difference and the geographical position of a heavenly body. In 1728
tables of the ephemerides of sufficient accuracy were not generally
available. During the following years Harrison developed an accurate
pendulum clock which aided the development of his sea clocks and
produced a design which he hoped would be used to make clocks to
keep standard time at ports of departure with an accuracy of one
second in a hundred days.
This development has to be considered in the light of social
conditions and scientific thought of the mid eighteenth century.
When he moved to London in 1737 he became one of the subscribers
to Isaac Wares translation and publication of Palladios Four Books of
Architecture (reprinted by Dover Books, 1965). This subscribers list
gives a good indication of Harrisons circle of influential friends and
vii
contacts in London and together with his writings suggests how he
obtained an introduction to Edmond Halley and King George III.
Wares book was dedicated to Lord Burlington.
William Kent (1685-1748), one of Wares subscribers, was
born in Bridlington, Yorkshire, and trained as a sign painter. One of
the most important Harrison clocks is the turret clock at Brocklesby
Hall. Brocklesby Hall was refurbished by William Kent in 1710
providing a Palladian facade. For ten years Kent lived and studied in
Rome. In 1715 he met Lord Burlington, another Yorkshireman, who
encouraged him to work as an architect. He notably designed
Holkham Hall, Houghton Hall and Horse Guards (1745). Harrison
was a carpenter and land surveyor who may have been associated with
Kent at Brocklesby during its refurbishment and subsequently.
According to their Charter of 1671, all clock and watchmakers
within the City of London and within a radius of ten miles of the city
boundaries had to be a member of the Worshipful Company of
Clockmakers. Harrison was never a member of the Company. He
made or had made prototypes according to his designs. His
manuscripts illustrate his knowledge of contemporary mathematical
and physical principles, e.g. the application of Hookes Law and the
inverse square law. Harrisons last known pendulum clock, now known
as the Royal Astronomical Society regulator, is the culmination of his
work and has many innovations. The movement was made to be
looked at, not hidden in a case, and was therefore designed as an object
of beauty. It can be considered to be one of the most important
mechanical artefacts of the eighteenth century existing today.
To determine your position at sea using a heavenly body you
need to know its geographical position and its angular height above
the horizon. The geographical position is the position on the earths
surface where it is absolutely overhead. Because the earth is rotating
and the sun, moon and planets are moving along their elliptical orbits,
ephemeris tables are required giving the geographical position at de-
fined times. By measuring the angle between the heavenly body and
the overhead point, the zenith, the distance of the observer from the
geographical position can be calculated using spherical trigonometry.
Unfortunately the zenith is not marked in the sky, so that the angle
between the heavenly body and the horizon has to be measured and
subtracted from 90E. In 1731 Hadley and Godfrey separately
invented the reflecting octant, the forerunner of the sextant which
enabled angles to be measured accurately at sea. Hadleys first design
viii
was a reflecting quadrant following Newtons earlier unpublished
design and was published in the Philosophical Transaction of the
Royal Society vol 37, p.147, 13th May 1731.
Harrison developed his coach or sedan clock, H4, to prove
that time could be carried reliably at sea by mechanical means. This
enabled seamen to determine the geographical position of the heavenly
body they were observing. His pendulum clock was accurate enough
for the determination of the ephemerides as well as for setting a
chronometer at the commencement of a voyage. Nevil Maskelyne
continued Halleys work and published the Nautical Almanac giving
accurate ephemerides. Further developments of the chronometer and
octant made position finding more available to the Royal and
merchant navy.
But how the Nation [or World] must, or will fairly come at what is
much better, or truly fit for this Purpose, ought to be fairly or rightly
considered; not but that I once thought of giving a Clock to the Observatory at
Greenwich, but my bad usage proved too tedious for that; but still , as already
hinted, my next or second clock will be somewhat better than if it had been
finished sooner; and if I live to see it done, that will be my Wages in the Matter;
but one would think, a Minstrel at the Play House is much better off,, save
only this has been my great Delight, and must be more noble than his
Achievements.
John Harrison, Concerning Such Mechanism, 1775
Harrisons writings and remaining artifacts are an important
example of the development of the scientific method and mathematical
advances pioneered by members of the Royal Society following the
Restoration. The English Civil War had enabled free discussion on
ix
religious topics; free discussion on the physical state of the universe
was now possible. Careful study of Harrisons work engenders great
respect for his logical thought and craftsmanship and gives a unique
insight into the political and scientific life of eighteenth century
London.
Acknowledgements
I am most grateful to the President and Council of the Royal
Astronomical Society for permission to photograph the detail of their
regulator housed at the Royal Greenwich Observatory. This study
would have been impossible without their permission and Goulds
notes which were presented to the Society and are kept in the James
Caird Library at Greenwich. I wish to thank the many people who
have in various ways assisted with this investigation of Harrisons
work, but in particular wish to thank Jonathan Betts for his assistance
on my visits to Flamstead House, Adam Perkins, Curator of Scientific
Manuscripts, Cambridge University Library for showing me the
Bradley drawings, Rebecca and Rhys Brookes for discussion, Henry
Casson and Peter Hastings for their commenting of my script and Ray
Darnell for his help in correcting the drawings and manufacture of
prototype parts. I apologize to anyone who considers that I have not
acknowledged their help or work in this descriptive monograph
promoting John Harrisons enormous achievement in introducing
scientific and mathematical thought into horology.
x
Chapter 1
John Harrisons Astronomical Regulator
No complete narrative of Harrisons career was ever written. Only
a short notice of him appears in the Biographia Britannica, published in
1766, during his lifetime - the facts of which were obtained from himself. A
few notices of him appear in the Annual Register, also published during his
lifetime. The final notice appeared in the volume published in 1777, the year
after his death. No Life of him has since appeared. Had he been a destructive
hero, and fought battles by land or sea, we should have had biographies of him
without end. But he pursued a more peaceful and industrious course. His
discovery conferred an incalculable advantage on navigation, and enabled
innumerable lives to be saved at sea; it also added to the domains of science by
its more exact measurement of time. But his memory has been suffered to pass
silently away, without any record being left for the benefit and advantage of
those who have succeeded him. The following memoir includes nearly all that
is known of the life and labours of John Harrison.
Samuel Smiles, Men of Invention and Industry, 1884
The first biography of John Harrison (1693-1776) was
written by Samuel Smiles in 1884 in his Men of Invention and
Industry. Humphrey Quill amplified Smiles essay in his book
John Harrison, the Man who found Longitude published in 1965.
Neither writer gives a full description of Harrisons activities
after he moved to London in 1737. The Harrison family were
able to move from Wragby to Barrow in 1700 because his
father, Henry, remained in the employment of Sir Rowland
Wynn. The Settlement Act of 1662 meant that only
freeholders, indentured apprentices and contracted employees
could move from a parish without being classed as a vagrant.
London had a strict Guild system for its artisans and self
employed. Harrison moved to London, prospered, and was able
to converse readily with Edmond Halley and other members of
the Royal Society. He was able to earn sufficient money to
have a watch made to his own design by one of the leading
London watchmakers and was able to find sufficient to have the
large coach clock, H4, in the most part made by London
artisans. The money forwarded by the Longitude Commission
was insufficient to cover the cost of the materials and making of
the sea clocks, RAS regulator and H4, together with his living
expenses.
1
His early interest in clocks would have been stimulated
by Rowland Lockeys copy of Hans Holbeins Design for the More
Family Group (1526-7) at Nostell Priory, the home of Sir
Rowland Winn, his fathers employer, which shows a wall clock
prominently in the centre. The original painting by Hans
Holbein was lost in a fire in the eighteenth century but the
Nostell Priory painting is thought to be a faithful copy of the
pen and wash drawing although he added a moon dial to the
clock!
In his latter years John Harrison, (1693-1776) wished
to make an accurate regulator to present to the Royal
Observatory at Greenwich. He claimed that he had experience
of clocks that could be made to go to 2 or 3 Seconds in a Year in a
manuscript that he wrote in 1730. He repeated his claim that
this accuracy was possible in his 1775 printed pamphlet entitled
A Description Concerning Such Mechanism as Will Afford a Nice, or True
Mensuration of Time, &c. written when he was 82 years old.
whenas, if proper steps be taken, or can be taken, in or for the adjusting
my next Clock, there must then be more Reason, (and that withal, as
from my Experience in my other Clock) that it shall perform to a
Second in 100 Days,
John Harrison , Concerning Such Mechanism, 1775
Copy of Holbeins Thomas Mores Family Rowland Lockey
2
When he went up to London before 1728 to visit
Edmond Halley, Astronomer Royal, he observed that no other
existing clocks on land or sea could fulfil the conditions of the
1714 Queen Anne Longitude statute.
Edmond Halley (1656-1742),
the 2
nd
Astronomer-Royal (1720-
1742) had a diverse career. The Royal
Society in London asked him to
analyse some data collected by Caspar
Neumann, a clergyman of Breslau in
Germany, relating to the births and
deaths in that city between 1687 and
1691. Unlike the London data which
had been used by John Gaunt to
predict the onset of plague for rich
householders, the Breslau data was clas-
sified by age, and this enabled more accurate survival rates at
each age to be obtained. Halley used the data in 1693 to
construct his own life table (for individual ages, not just age
groups), which was found to give a reasonably accurate picture
of survival and became well known throughout Europe.
Most important, however, was the method which
Halley demonstrated of using his life table to work out how
much money someone of a given age should pay to purchase a
life-annuity. Halley examined each future annual instalment
of the annuity separately and used his life table to estimate the
probability that the person would survive to receive that
instalment. The resulting probability was multiplied by the
sum (obtained from a compound interest table at a specified
rate of interest) which would need to be invested now in order
to pay for that installment if one were certain to receive it.
Halley then went on to do likewise for the next instalment,
and so on. Summing these present values for all future
instalments up to the end of life then gave the value of the
whole annuity. Actuarial science had been created.
Probability theory had emerged with a publication by
Christiaan Huygens in 1657. When all Harrisons references
to the accuracy of his clocks are assessed it is clear that he
understood variance and trend as defined at that time. His
observations brought geometric observation to bear on the
laws of motion, thereby hinting at non-predictability - i.e. the
chaos theory - and pre-empting Poincar.
Sir Edmund Halley
3
Whenas, by a right Use of my Clocks, [viz. as when with
proper Conveniences and Apparatusses] even that Piece of Astronomy
[the Equation] may hereafter be corrected, because as thence the
Eccentricity of the Earths Orbit, and whatever else may pertain to the
Matter, may be the better known.
John Harrison, Concerning Such Mechanism, 1775
Halley required an accurate land clock to confirm
Keplers and Newtons work on the motion of heavenly bodies
and therefore to check the astronomical ephemera which were
being prepared at Greenwich.
From the time I speak down to the time of Halley, there was
no clock upon which reliance could be placed. Attention was called by
Baily to the fact that Halley in his observations from time to time used
a clock wound up by a string, which was probably not any better than
the alarums now common.
p. 296, The Astronomical Register, December, 1880
Halley was working in particular on the position of
the moon. The mathematics of deriving these predictions of
the three body gravity problem (sun, earth and moon) had
been solved by Leonhard Euler and was published by him in
1760. Tobias Mayer began calculating lunar and solar tables
in 1753 using Eulers method, and in 1755 he sent them to
the British Government. These tables were good enough to
determine longitude at sea with an accuracy of half a degree.
(viz. a position within 15 nautical miles at the equator or 9
miles at the latitude of Greenwich) The Board of Longitude
considered that the method was too complicated for general
use by seamen and declined to award the Longitude Prize to
Bradley and Mayer for this lunar distance method of
determining time. The British Government later sent Mayers
widow 3000 as an award for the tables.
Now it is to be understood, that I had along with me (as
affording the Principles upon which we reasoned) the Descriptions, with
some Drawings of the principal Parts of the Pendulum-Clock which I
had made, and as also of them of my intended Time-Keeper for the
Longitude at Sea. But here it must be highly worth remarking, that I
at first made Application to Dr. Halley, and as he had long been
observing for the Longitude by the Moon, as then becoming quite tired
of it, or thoroughly satisfied, as touching the Impossibility of its ever
doing any certain Good, [but not then so as to be let known to me, but
was afterwards told it me by Mr. Graham] received me the better; and
in his finding what I came about as principally touching a true
4
Mensuration of Time, viz. For that Purpose of Longitude; but as
previous thereto, viz. As in his finding my Dealings with the Pendulum
for a true Mensuration of Time, to be very much contrary to Mr.
Grahams thoughts of, or Proceedings in the Matter, advised me
therefore to go to Mr. Graham, but that Advice went hard with me, for
I thought it as a Step very improper to be taken; but he told me, that in
the Way which I was in, viz. as by Machinery [for the Longitude]
more than Astronomy, that I should certainly be sent to Mr. Graham,
and therefore advised me to go without any farther to do; certifying me
also, that Mr. Graham was a very honest Man, and would do me no
harm, viz. as by pirating any Thing from me, but that on the contrary,
would certainly do me Good if it was in his Power; but withal;,
cautioned me how to begin with Mr. Graham, viz. So as in as few
words as possible to let him to understand, that I had indeed something
worthy of Notice to communicate to him; but notwithstanding that
Piece of Advice, and my doing my best as pursuant thereto, Mr.
Graham began, as I thought it, very roughly with me, and the which
had like to have occasioned me to become rough too; but however we got
the Ice broke, the which would not bear, and the Passage over was as I
have shewn, and indeed he became as at last vastly surprised at the
Thoughts or Methods I had taken, or had found Occasion to take, and
as thence found Reason enough to believe that my Clock might go to a
second in a Month, and that as in Consequence to be, as in the first
Place, of great Service in the adjusting a Time-Keeper for the
Longitude; and indeed, as according to Dr. Halley, Mr Graham proved
a very great Friend to me.
John Harrison, Concerning Such Mechanism, 1775
In 1761 Rev. Nevil
Maskelyne (1732-1811), 5
th
Astronomer-Royal (1765-1811),
was deputed by the Royal Society to
make observations of the transit of
Venus at St Helena. During the
voyage to St. Helena he
experimented on the determination
of longitude by lunar distances using
Mayers tables, and ultimately
introduced the method into
navigation. In 1763 he published the British

Mariners Guide
which includes the suggestion that in order to facilitate the
finding of longitude at sea lunar distances should be
calculated beforehand for each year and published in a form
accessible to navigators. This important proposal, the germ of
the Nautical

Almanac was approved of by the government, and
Rev. Nevil Maskelyne
5
under the care of Maskelyne the Nautical Almanac for 1767
was published in 1766.
The Commissioners of Longitude, in pursuance of the Powers
vested in them by a late Act of Parliament, present the Publick with the
NAUTICAL ALMANACK and ASTRONOMICAL
EPHEMERIS for the year 1767, to be continued annually; a work
which must greatly contribute fo the Improvement of Astronomy,
Geography and Navigation. This EPHEMERIS contains every Thing
essential to general Use that is to be found in any Ephemeris hitherto
published, with many other useful and interesting Particulars never yet
offered to the Publick in any Work of this Kind. The Tables of the
Moon had been brought by the late Professor MAYER of Gottingen to
a sufficient Exactness to determine the Longitude at Sea, within a
degree, as appeared by the Trials of several Persons who made Use of
them. The difficulty and Length of the necessary Calculations seemed
the only Obstacles to hinder them from becoming of general Use: To
remove which this EPHEMERIS was made; the Mariner being hereby
relieved from the Necessity of calculating the Moons place from the
Tables, and afterwards computing the Distance to Seconds by
Logarithms, which are the principal and only very delicate Part of the
Calculus; so that finding the Longitude by the help of the
EPHEMERIS is now in a Manner reduced to the Computation of the
Time.
It may be proper first to premise, that all the Calculations are
made according to apparent Time by the Meridian of the Royal
Observatory at Greenwich.
But if Watches made upon Mr. John Harrisons or other
equivalent Principles should be brought into Use at Sea, the apparent
time deduced from an altitude of the Sun must be corrected by the
Equation of Time, and the mean time found compared with that shewn
by the Watch, the difference will be the Longitude in Time from the
Meridian by which the Watch was set; as near as the Going of the
Watch can be depended upon.
Maskelyne, First Edition of Nautical Almanack, 1766
Maskelyne continued the superintendence of this
invaluable annual during the remainder of his life. He further
induced the government to print his observations annually,
thereby securing the prompt dissemination of a large mass of
data of inestimable value for their continuity and accuracy.
6
Maskelyne had but one assistant, yet the work of the
observatory was perfectly organized and methodically
executed. He was awarded the Copley Medal of the Royal
Society in 1775 for his experimental work in finding the
earths density using a plumb line at Schiehallion, Perthshire.
Tobias Mayer (1723-1762), was a self taught
mathematician who worked as a cartographer in
Nrnberg. He discovered the libration of the
Moon and this gained him fame which led to his
appointment as professor of economics and
mathematics at Gttingen in 1751.
I am the more unwilling my tables should lie any
longer concealed; especially as the most celebrated
astronomers of almost every age have ardently wished for a
perfect theory of the Moon ... on account of its singular use
in navigation. I have constructed theses tables ... with respect to the inequalities of motions,
from that famous theory of the great Newton, which that eminent mathematician Eulerus first
elegantly reduced to general analytic equations.
In 1742 James Bradley (1693-1762) was appointed 3rd
Astronomer Royal, succeeding Edmund Halley. Bradleys time saw the
breakthrough in solving the longitude problem by lunar distances. Tobias
Mayer had been working at Gttingen on a theory of the Moon and had
conducted observations the equal of Bradleys in precision. In 1754 Mayer
submitted his tables to the Board of Longitude hoping for an award from the
Longitude Prize of 1714. The tables were passed to Bradley who checked
them against the Greenwich data and found them to be very accurate. In
1761 he despatched Nevil Maskelyne to St Helena to observe the Transit of
Venus that year. Maskelyne used Mayers tables to find longitude by lunars
on the voyage and returned convinced about the method. The Board of
Longitude did not accept Bradley and Mayers lunar distance method as
viable because they considered the calculations too complex. Mayers method
of determining longitude by lunar distances and a formula for correcting
errors in longitude due to atmospheric refraction were published in 1770
after his death.
The Georg-August University of Gttingen was founded in 1737
by King George II, (who was also Elector of Hanover) as a University of the
Enlightenment, thereby forsaking theology as the dominant faculty. Its
alumni now include over forty Nobel laureates.
7
Halleys comment about Harrisons demeanour and
perusal of his manuscripts suggest that he suffered from
Specific Speech Disorder, an inherited form of Brocas
dysphasia. He displays many of the other characteristics of
this affliction i.e. an excellent musical ear and enhanced spatial
cognisance. His writing is typically rambling but when
studied carefully is found to be as accurate as his
draughtsmanship. He has an exceedingly large vocabulary for
his time. He would have appeared awkward and garrulous to
his peers and this probably accounts in part for declining
election as Fellow of the Royal Society, even though he was
awarded their Copley Medal in 1749. Martin Folkes was
President giving the citation. Harrison was not a member of
the Worshipful Company of Clockmakers, would at present
only qualify to be an Associate of the British Horological
Institute and would not be interested in the Antiquarian
Horological Society!
John Harrison made a large pendulum clock at the
time that he was making his third sea clock. He most
probably used this clock to try out many of the developments
which he included in the third sea clock (H3)and the fourth
sea clock (H4) which eventually won him a large part of the
Longitude Prize. This large clock exists and is usually referred
to as the Royal Astronomical Society Regulator. It is an
improvement on the clocks with predominantly wooden
movements that he made around 1726. The proportional
design of the movement of this clock is based upon irrational
numbers and reflects the metaphysical ideas of the seventeenth
and eighteenth centuries as also illustrated by Robert Boyle,
Robert Hooke and Isaac Newton. Harrison developed his
ideas throughout his life and the RAS regulator is almost
certainly the prototype of the improved astronomical
regulator described in Concerning Such Mechanism that he
proposed to design and build for the Greenwich Observatory.
A clock made and finished according to the principles
that he expounded in his manuscripts, together with the detail
in his portrait, and examination of surviving artefacts can be
used to understand fully his ideas and to determine whether
8
the degree of accuracy in timekeeping he suggested could have
been attained in a clock made in the eighteenth century with
the available materials and technology.
Now if there were Clocks made as described above, (of Which
I have had experience, even such Clocks as may be made to go to 2 or
3 Seconds in a Year) and fixd at Sundry Ports in the World, where
Ships resort, or rather each Clock made (but however its Cycloid etc.,
corrected) at the Port where it is to be fixd (because nearer the Equator
Pendulums Oscillate slower) these would be good standards to set the
Sea Clocks by; when the Ships are ready to sail. And if the Sea Clocks
were made as here treated of (which is also practicable) I think they
would not be much inferior to the other. But if in the Ships they should
vary 4 or 5 seconds in a Month, it would not always be one way,
which makes the variation less in regard to its use, (for the mean of
what is always one way implies the want of better adjusting) and 4
seconds of Time being but 1 Minute of the Equinoctial; (or but little
more than a Mile towards the Equinoctial, and not so much towards
the Poles) such little variation cannot deceive the Sea Men much in the
Time they sail to a far Port, or to where there is another fixd Clock.
John Harrison, 1730 Document
Because the RAS regulator was unfinished by
Harrison additions have been made since to make the clock
functional. Neither are any of the 1726 series of wooden
movements in exactly the same condition as they left
Harrisons hands. This monograph sets out to explore and
prove what is original work. All the evidence has to be
assessed and an attempt made to determine Harrisons
intention for the missing parts. Studying his written work
together with examination of the methods he used in making
other clocks, particularly the sea clocks, during his long
working life is the only means by which his intentions can be
determined. Harrison considered that the provision of an
accurate time standard at the port of departure was of equal
importance to the determination of longitude at sea as the
accuracy of the chronometer carried by the ship.
John Harrison moved to London in 1737, living at
first near to and then in Red Lyon Square. This area was a
suburb of a rapidly developing London close to open fields.
The Middlesex Hospital was founded in 1745 in nearby
Windmill Street and in 1739 Thomas Coram promoted the
9
Foundling Hospital in fields near to Queen Square. James
Gibbs had built St. Mary-le-Strand church in 1717, and St
Martins in the Fields in 1726. John Harrison would have
been interested in these buildings and in James Gibbs two
books A Book of Architecture (1728) and Rules for Drawing the
Several Parts of Architecture (1728).
Harrisons name appears in the list of subscribers in
the translation of The Four Books of Andrea Palladio published
by Isaac Ware in 1738 in London. The list contains groups
of people; members of Walpoles government, tradesmen
working for or tenants of Lord Burlington, men who would
become Governors of Corams Foundling Hospital and
associates of the East India Company. Acquaintances of
Harrison listed are William Hogarth, Thomas Heath
(secretary, Royal Society) and Lord Macclesfield (p. 100,
CSM). Many of the listed subscribers to this book are
acknowledged by him in his writings as helpers in his
endeavours.
William Hogarth referred to Harrisons large sea
clocks in his book The Analysis of Beauty (1753). This most
probably stimulated Harrison to design the second and third
sea clocks and the RAS regulator according to contemporary
ideas of proportion.
John Harrison began his working life as a carpenter
and surveyor. Whilst designing, drawing and setting out the
front and back plates and dials of this regulator he would have
used straight edge, set square, dividers, Gunters scale and
either proportional dividers or a sector. His measurements
would have been relative rather than definitive, with an
emphasis on proportional division. Reconstruction of his
drawings using the same methods demonstrates the accuracy
he was able to attain. Mathematical instruments were
available at Heaths shop in the Strand and their use is
described in Robertsons 1747 Treatise of Mathematical
Instruments (Third Edition 1775 reprinted by Flower-de-Lune
Books). Instruments supplied by Jonathan Sisson in the
Strand were advertised in Samuel Wylds 1725 The Practical
Surveyor.
10
Martin Folkes (1690-1754),
President of the Royal Society 1741-
1754, and President of the Society of
Antiquaries 1749-1754. In 1739 was
elected one of the founding Vice
Presidents of the Foundling Hospital.
He was largely responsible for the
exclusion of Coram from the General
Committee of the Hospital in 1742.
Dr. Stukeley was a friend and wrote
the following description of him. It is
evident that there would not be much
affinity between Folkes and Coram because of the latters
religious views. For the same reason it is difficult to consider
that there could be much affinity between Folkes and Harrison
and this may be the reason that Harrison declined Fellowship of
the Royal Society.
He chuses the Council and officers out of his junto of
Sycophants that meet him every night in Rawthmills coffee house, or that dine
with him on Thursdays at the Miter, fleet street He has a good deal of
learning, philosophy: but knows nothing of natural history. In matters of
religion an errant infidel & loud scoffer. Professes himself a godfather to all
monkeys, believes nothing of a future state, of Scriptures, of revelation. He
perverted the Duke of Montagu, Richmond, Ld Pembroke, & very many more
of the nobility, who had an opinion of his understanding; and this has done
an infinite prejudice to Religion in general, made the nobility throw off the
mask & openly deride and discountenance even the appearance of religion
which has brought us into that deplorable situation we are now in, with
thieves & Murderers, perjury, forgery &c. He thinks there is no difference
between us & animals; but what is owing to the different structure of our
brain, as between man & man. When he lived in Ormond Street in 1720
he set up an infidel Club at his house on Sunday evening, where Will Jones,
the mathematician & others of the heathen stamp, assembled. He invited me
earnestly to come thither but I always refused. From that time he has been
propagating the infidel System with great assiduity, & made it even
fashionable in the Royal Society, so that when any mention is made of Moses,
the deluge, of religion, Scriptures &c., it is generally received with a loud
laugh.
William Stukeley, Journal
Folkes continued to support Harrison throughout his term
as President of the Royal Society.
Martin Folkes
11
Harrison travelled to London at some time before
1728 to meet Edmund Halley, Astronomer Royal.
According to the social conventions of the time Harrison
must have had an introduction to Halley. This was most
probably through connections in the East India Company.
The East India Company gave Harrison 100 towards the
development of his sea clocks. Halley had been an employee
of the Company; he succeeded Robert Hooke (1635-1703)
as Curator of the Royal Society. Harrison was introduced to
George Graham (1673-1751) by Halley. Graham worked
with his uncle-in-law, Thomas Tompion (1639-1713) and
managed the business. Tompion had worked closely with
Hooke, and under his supervision had made one of the first
English watches equipped with a balance-spring. Graham
must have learnt details of Hookes work on springs, pendula,
double balance wheels, and antifriction wheels from Tompion.
Dr. Stukeley wrote about Harrisons clock in his manuscript
journal of 1728:-
I saw his famous clock last winter at Mr. George Grahams.
The sweetness of the motion, the contrivances to take off friction, to defeat
the lengthening and shortening of the pendulum through heat and cold
cannot be sufficiently admired.
Harrisons 1730 manuscript must have been written
after Harrison had become well acquainted with Graham.
The similarity between Hookes marginal drawing and the
interlinked compound pendula in Harrisons manuscript may
have arisen after conversation between the two men. Harrison
was awarded the Copley Medal of the Royal Society (its
highest award) in 1749 for his work with his first sea clock.
Sir Godfrey Copley was a close friend of Hooke.
His third (H3) sea clock has large, heavy vertical
balance wheels linked in a similar manner to the two
preceding sea clocks. He found that these mutually
compensating large masses could not maintain reliable
oscillations in the irregular motion of a ship even when
compensated for changes in temperature and made
isochronous. As soon as he realised that the solution was a
small lightweight balance wheel vibrating 5 times per second,
the large sea clocks became redundant and he ceased
development of them.
12
Harrisons over-riding principle in the design of his
clocks is to use mechanical methods to minimise friction and
to use no lubricating fluids so that the frictional losses would
remain constant with age. There are no sliding plane surfaces
in the RAS regulator and radial sliding motion is minimised
by using roller bearings, friction wheels and lantern pinion
rollers.
Whenas, by a right Use of my Clocks, [viz. as when with proper
Conveniences and Apparatusses] even that Piece of Astronomy [the
Equation] may hereafter be corrected, because as thence the Eccentricity
of the Earths Orbit, and whatever else may pertain to the Matter, may
be the better known.
John Harrison, Concerning Such Mechanism, 1775
Some Years ago I made several alterations in order to render
the Motion of Clocks more exact than heretofore, but when I came to try
them by strict observation as below, I judgd the best performance of the
best Pendulum Clock I ever saw, made, or heard of, to be incapable of
this Matter, would it go as well in a Ship at Sea in any part of the
World, as in any one fixd place on the Land. Yet from several
observations, I still endeavourd to make further Corrections in this
Motion; and in these 3 last Years have brought a Clock to go nearer the
truth than can be well imagind, considering the vast Number of
seconds of Time there is in a Month, in which space of time it does not
vary above one second, and that mostly the way I expect: So I am sure
I can bring it to the Nicety of 2 or 3 seconds in a Year. And twill also
continue this exactness for 40 or 50 Years or more; however so as not
to vary above 2 or 3 seconds from what it did the Year next before; for
twill not want Cleaning, and the little wear can but alter it insensibly
little. This Nicety is owing partly to the Matter the Clock is made of,
partly to the Contrivance it is made with, and partly to the Nice
observations it is tryd by, and the convenient place it stands in.
John Harrison, 1730 Manuscript
In stating the accuracy of his regulator Harrison
suggests that it can vary by up to 1 second per day, but
then continues to state the trend over a year. In Concerning
Such Mechanisms he confirms that he can attain this accuracy
with the phrase one second in a hundred days. Because of
the chaotic nature of the action of a pendulum the variance
13
and the average deviation diverge rather than converge when
calculated from increasing number of observations. The Allan
variance (two-sample variance) has to be used to measure the
stability of oscillators. Woodward collected data from a
number of accurate clocks. These results are drawn on
logarithmic scales and show a horizontal line where there is
no change in rate or accuracy. The results from two Dent
clocks having Graham dead beat escapements are shown. The
1904 results are from a clock that has barometric error
compensation; the 1966 results from a clock that did not.
These results indicate that Harrison was claiming an accuracy
for his clocks that is very close to the theoretical limits for a
clock in free air and that he had mitigated the changes in rate
due to the varying viscosity of air.
The importance of Harrisons purported accuracy of
his regulators is that it was what the astronomers desired at
that time. Johann Bayer (1572-1625) published Uranometria:
omnium asterismorum continens schemata , the first star atlas and
catalogue in 1603. He developed the star naming system and
14
showed the stars as seen from earth. His atlas has a Cartesian
grid to enable star position to be depicted accurately.
John Flamsteeds Atlas coelestis was published
posthumously in 1729, four years after his British Catalogue.
This atlas and catalogue of unprecedented accuracy and
Halleys Astronomical Tables have Right Ascension accurate to
2 seconds. A clock, if it is to be relevant for long term
observations would therefore have to be accurate to 2-3
seconds in a year. This accuracy claim by Harrison, was most
probably the reason for his introduction to Halley by mutual
acquaintances in the East India Company. The clock was
then carefully examined and rated by Graham as noted in Dr.
Stukeleys Journal of 1728.
John Harrison died on 24
th
March 1776, his eighty
third birthday. He is buried in the churchyard of St. Johns,
Hampstead, approximately three and a half miles from his
home in Red Lyon Square. The church was unusable in
1744; it was rebuilt and re-dedicated in 1747 to designs by
Henry Flitcroft and John Saunderson, both subscribers to
Isaac Wares book. The tower has never had bells for change-
ringing. It is suggested that Harrison had either been closely
connected with the building of the church or he had moved to
this newly fashionable out of town health resort before he
died.
Dr. William Stukeley, 1687-1765, British
antiquarian, was born in Holbeach.
Qualified in Medicine at Cambridge and St.
Thomass Hospital, but took Anglican
orders in1729 and the living of Stamford.
Published details of Stonehenge and
Avebury circles in 1743. Together with
Halley tried to align Stonehenge with
magnetic north. Became a close friend of Sir
Isaac Newton and Martin Folkes. He was Newtons first biographer
and is responsible for the story of the falling apple.
15
An Act for Regulating the Commencement of the Year; and for Correcting the Calendar
Now in Use. 25 George II. c.30 1751
The list of subscribers in Isaac Wares translation and printing of Palladios Four Books of
Architecture includes the names of Sir Robert Walpole and many members of his Whig government. Two of
them, Philip Stanhope, 4
th
Earl of Chesterfield (1694-1773) and George Parker, 2
nd
Earl of Macclesfield
(1697-1764), supported by James Bradley (1693-1762), Halleys successor as Astronomer Royal, introduced
the Calendar Bill in Parliament in 1750. Lord Macclesfield was Vice-President and Lord Chesterfield a
Governor of Corams Foundling Hospital. Lord Macclesfield succeeded Sir Hans Sloane as President of the
Royal Society in 1752.
The Roman calendar before Julius Caesar was based on a year of 365 days. In his time it was
realised that the calendar had fallen out of step with the seasons because it had been determined that the
actual length of time for the sun to orbit the earth was nearer to 365.25 days! The calendar was corrected by
adding an extra day every four years and two extra months.
By the seventeenth century the calendar was again out of step, because 365.25 days is a slight over-
estimate of the length of a year. The cumulative error then amounted to ten days. So Pope Gregory XIII, on
the advice of Aloysius Lilius, decreed that the day following 4
th
October 1582 would be 15
th
October 1582.
To prevent the problem reoccurring the rule for leap years was changed so that the end of century years would
not be leap years unless divisible by 400. Pope Gregorys decree was not accepted by the Protestant parts of
Europe causing its introduction to be staggered over 160 years. The Gregorian year is accurate to 26secs.
The Act came into force in 1751 amid
popular disapproval and rioting. The rioting is
depicted in the background of Hogarths
painting Canvassing for Votes. The Act
changed the first day of the year from 25
th
March (Lady Day) to January 1
st
. In England
the day after 24
th
March 1642 was 25
th
March
1643. The Act changed this so that the day after
31
st
December 1751 became 1
st
January 1752.
This resulted in 1751 being a short year from
25
th
March until 31
st
December. Scotland had
changed the first day of the year to 1
st
January in
1660.
The 2
nd
of September 1752 was
followed by the 14
th
September 1752 to change
the English Julian Calendar to the Gregorian calendar, which was used throughout most of Europe. Hogarths
paintings of the 1755 Parliamentary election is the historical source of the Give us our eleven days slogan.
The beginning of the financial year was changed from Lady Day to April 5
th
so that the Government did not
lose taxable days.
Nevil Maskelynes method of Lunar Distance for finding time at sea was based on observations made
at Greenwich. This led to the publication of ephemera in the Nautical Almanac and the adoption of
Greenwich mean time as the British and American navigators standard time. With the advent of the railways
in Britain, GMT was adopted by the Railway Clearing House in 1847 but did not become the legally adopted
official time throughout Britain until 1880 (Statutes(Definition of Time) Act 1880, 43&44 Vict. c.9).
The standard time, common to every place in the world, is now Co-ordinated Universal Time
(UTC), derived from International Atomic Time (TAI) by the addition of a whole number of leap seconds
to synchronise it with Universal Time (UT1), thus allowing for the eccentricity of the earths orbit, the
rotational axis tilt (23.5E), but still showing the earths irregular rotation on which UT1 is based. Co-
ordinated Universal Time (UTC) is expressed using a 24 clock and uses the Gregorian calendar. It is used in
navigation, where it is also known by the name Zulu Time. Zulu stands for Z in the international phonetic
alphabet and is used to label longitude zero.
Following the Interpretation Act 1978 and the Co-ordinated Universal Time Act 1997 we
should refer to UTC rather than GMT in legal documents. These Acts updated the 1880 Definition of Time
Act and confirm the universal principles of time and place from Greenwich, as agreed in the First
International Meridian Conference 1884 (Washington Conference). Since the mid 1970's the Westminster
clock has been set to UTC. The maximum difference between UTC and GMT is 0.9 seconds.
Thomas Coram (1668-1751), seaman and founder of the Foundling Hospital, went to
Massachusetts for ten years as a young man building ships and working to further the spread of
Anglicanism. Driven out by religious prejudice he returned to England heavily in debt. He joined
the Master of the Hatters Company to petition Parliament against the manufacture of hats in the
plantations resulting in the 1732 Hat Act. A strong believer in womens rights and equal
opportunities for girls, he was convinced that it was due to the unique support of aristocratic women
that he was granted a Royal Charter for the Foundling Hospital. Within two years of the
establishment of the Hospital he fell out with the Governors and was ejected from the governing
body. Hogarth admired Coram and wished to excel when he painted his portrait. Notable icons in
the picture are the Tuscan column, felt hat, globe showing North America, the Charter, no wig and
the feet hardly touching the ground.
William Hogarth Analysis of Beauty 1753
C H A P T E R XI.
Our necessities have taught us to mould the matter into various shapes, and to give them fit
proportions for particular uses, as bottles, glasses, knives, dishes, &c. Hath not offence given rise to the
form of the sword, and defence to that of the shield? And what else but proper fitness of parts hath fix'd
the different dimensions of pistols, muskets, great guns, fowling-pieces and blunderbusses ; which
different characters of fire-arms, as the different shapes of men are called characters of men
We find also that the profuse variety of shapes, which present themselves from the whole
animal creation, arise chiefly from the nice fitness of their parts, designed for accomplishing the peculiar
movements of each.
And here I think will be the proper place to speak of a most curious difference between the
living machines of nature, in respect of fitness, and such poor ones, in comparison with them, as men are
only capable of making ; by means of which distinction, I am in hopes of shewing what particularly
constitutes the utmost beauty of proportion in the human figure.
A clock, by the government's order, has been made, and another now making, by Mr. Harrison,
for the keeping of true time at sea ; which perhaps is one of the most exquisite movements ever made.
Happy the ingenious contriver! although the form of the whole, or of every part of this curious machine,
should be ever so confused, or displeasingly shaped to the eye ; and although even its movements should
be disagreeable to look at, provided it answers the end proposed. An ornamental composition was no part
of his scheme, otherwise than as a polish might be necessary. If ornaments are required to be added to
mend its shape, care must be taken that they are no obstruction to the movement itself, and the more as
they would be superfluous, as to the main design.---But in nature's machines, how wonderfully do we see
beauty and use go hand in hand!
Had a machine for this purpose been nature's work, the whole and every individual part would
have had exquisite beauty of form without danger of destroying the exquisiteness of its motions, even as
if ornament had been the sole aim ; its movements too would have been graceful, without one superfluous
tittle added for either of these lovely purposes.---Now this is that curious difference between the fitness
of nature's machines (one of which is man) and those made by mortal hands : which distinction is to lead
us to our main point proposed ; I mean, to the shewing what constitutes the utmost beauty of
proportion.
Chapter 2
Documentary Sources
Portraits
Sir Joshua Reynolds, Joseph Wright, and Thomas
King painted portraits titled John Harrison. Engravings were
made of these portrait paintings by William Henry
Worthington (1790-1839), John Raphael Smith (1752-
1812) and Phillipe Joseph Tassaert (1732 - 1803).
Harrison only refers to one portrait in his will.
Joseph Wright Portrait
J.R.Smith made a mezzotint copy of Joseph Wrights
portrait of John Harrison. Joseph Wright (1734-1797) was
a pupil of Thomas Hudson for two years in 1751. He had a
further fifteen months with Hudson in 1756-6 before
establishing a portrait painting practice in Derby. His
portrait of Harrison, present whereabouts unknown, was
almost certainly painted whilst Wright was working in
London. If so Harrison would be aged 58, which is
compatible with the painting. Harrison is wearing the same
Mezzotint of John Harrison by
J.R.Smith after J.Wright
19
Joseph Wright painted A Philosopher giving that Lecture on the Orrery for Admiral
Washington Shirley, 5
th
Earl Ferrers, in 1766. Washington Shirley was an
astronomer who had observed the transit of Venus in 1761 following which
he was elected a Fellow of the Royal Society. He inherited his title and
estates after his brother, 4
th
Earl Ferrers had been indicted and hung for
murdering his steward. The picture shows the planets in the ecliptic which is
graduated with dates and points of the zodiac and an ungraduated armillary
hemisphere. The figure making notes on the left is a likeness of Peter P.
Burdett, who made the first county survey with accurate instruments. The
central figure could be John Whitehurst, a Derby instrument maker, but is
more likely a copy of Vanderbanks 1725 portrait of Isaac Newton.
Whitehurst was a notable member of the Lunar Society. The figure on the
extreme right is a portrayal of John Harrison. The notable feature in his face
observed by Wright, King and Reynolds is the dominant supra-orbital ridge.
The same style of jacket is seen in all these three labelled portraits and in the
above picture.
A quotation from Titus Maccius Plantuss drama Miles Gloriosa is
engraved below Smiths engraving.
non invenies alterum lepidiorem ad omnis res nec qui amicus amico sit magis
You will not find another man of that age more delightful in every way, nor another who is
more of a friend to a friend
The engraving, 280 x 200 mm., is suitable for a quarto book, J.R.Smith was
appointed engraver to George III. The engraving is different from other
Smith engravings, having no carefully engraved dedication below.
20
style of suit as in the King and Reynolds portraits. The
image is very similar to the right hand figure in Wrights large
painting A Philosopher giving a Lecture on the Orrery.
When compared to a photograph of the original painting the
mezzotint is seen to be a remarkably close copy. A surgeon
named John Harrison was the primary founder of the London
Hospital in 1740. He was dismissed in disgrace a few years
later aged 35 years. Another John Harrison was the foreman
at Boulton and Watts Soho works in Birmingham
Thomas King Portrait
The portrait of John Harrison, aged 73, painted in
1766 by Thomas King is strikingly similar in style to William
Hogarths portrait of Captain Thomas Coram, the instigator
of the Foundling Hospital. William Harrison, (1728-1815)
Johns son by his second wife, was a Governor of the
Foundling Hospital from about 1773. Harrison had been
awarded the first part of the Longitude Prize on October
28th 1765.
Thomas King was a pupil of George Knapton in 1735
and became an assistant to Arthur Pond, painter and engraver,
in 1744 until 1748. He specialised in draperies and
reluctantly painted portraits, mostly in the later period of his
life. He was apparently eccentric and thriftless, dying at the
Bull and Axe public house in John Street, in 1769. John
Street is very close to Red Lion Square where Harrison was
living at this time.
Harrison is shown seated on a Chippendale style
chair amidst his creations, viz, his watch made by Jeffreys, the
sea clock H3 in a gimbaled case, a pendulum within a glass
fronted long case and drawings of his escapement wheel and
maintaining mechanism. He is wearing an expensive, fashion-
able three piece suit with marine cuffs, and a wig indicating
that he is a gentleman. The triangular composition and the
sitters gaze directs us to his right hand holding the watch
made by Jeffreys.
This watch undoubtably gave Harrison the
inspiration to make his H4 sea clock with an extremely
lightweight balance wheel which has its mass concentrated on
the periphery. The momentum exerted by a ships motion is
21
therefore small compared to the angular momentum
generated by the spiral spring. Until he encountered
insurmountable problems with his large sea clocks he had
followed the ideas of Newton and Hooke that a successful
sea clock could only be made with large pendula of great
mass. It has been suggested that the Jeffreys watch is
shown in the portrait because H4 was being examined for the
Board of Longitude at the time that the portrait was being
painted. This is implausible; if Harrison had wanted to
show H4 in the portrait, King could have been shown
drawings &c. This is an iconographic painted portrait, not a
photograph.
Now in my Longitude Time-Keeper, the Strength or
Command which the Ballance-Spring has over the Ballance, as in
Proportion to the Force it has from the Wheel, is so great as 80, or even
John Harrison by Thomas King 1766
22
more than 80 to 1; a strong artificial Gravity indeed, (for so it may
be termed) as even in Effect much surpassing natural. The Limb of the
Ballance moving thereby through about 25 Inches in a Second,
notwithstanding, as in that Time (and still an Augmentation to its
Power or Dominion) its Motion being changed to contrary Direction
5 Times, as from which it must as in Consequence be as farther
allowed, and as withal considering the Arch it describes, that its Motion
must be prodigiously quick, or even violently powerful in the Middle of
each Vibration, and when, as even without that, 25 Inches in a Second,
is no less than 34 Miles a Day; so rapid and powerful is the Motion
of the Ballance, for faint sleepy Work could never do; and whereas no
such Velocity (as in this Point for the better) can be in a Pendulum,
viz. in such Arches as in the long Pendulum Way are commonly, or can
as the most properly be described, until it comes to about the length of
13 Foot, and then it will still not be in Effect so quick, nor
consequently so powerful, because of it not having as in Effect so much
Velocity, but only so much as whereby to accomplish a Vibration, or
such a Space, viz. as without any Returns or Return, in the Space of
two Seconds of Time; but here, if it should be alledged, that the Length
of such a Pendulum is much longer than the Radius of this Ballance,
it is then to be remembered, that (as in common) the Lengths of the
Pallats, etc. are to be taken into the Question; and, as already in some
Measure implied, it is as farther to be understood or remembered, that
in such of these mechanical Cases, that no Ponderosity in a Pendulum
or a Ballance, can rightly or ever make up the Want of Velocity; and
indeed Velocity was very much wanting in my three large Machines,
yea, I say, very much, notwithstanding their Weightiness of Ballances,
or as notwithstanding what Philosophers may reason in other Affairs,
or rather what Philosophy in this Affair (viz. as hitherto through Mis-
take, as from the Steel-Yard, etc. in the Matter) might teach us to
reason; but I did not then (viz. in the Designs of my three large
Machines) so thoroughly understand it, nor should I, or the World, (as
I think I may make bold to say) ever have understood it as it were
through Accident, in or by my third Machine; but as it would be more
tedious here than necessary, for me to fully reason the Matter, I shall
forbear it; but still the Knowledge of the same is highly material.
John Harrison, Concerning Such Mechanism, 1775
Harrison had to succeed with H4, his longitude
timekeeper, against a background of immense scepticism.
Seamen making the passage to and from the New World were
reticent to use a mechanical device which they considered
unreliable to confirm their deduced reckoning position. If a
simple astronomical method could be found it would not have
this uncertainty. Following the discovery of the Americas
23
seamen began sailing long distances in East-West directions
stimulating the need for an accurate and simple means of
determining longitude. The introduction of the pendulum
replacing balance wheels and foliots had improved the
timekeeping of domestic clocks from an error of thirty
minutes per day to less than a minute. The use of a balance
wheel with a verge escapement would be seen as a backward
step according to accepted wisdom. For this reason the H4 is
over-engineered and has everything included that he thinks
important viz. a verge escapement for minimal and constant
friction, remontoire control of power to the escapement,
temperature compensation, a strong balance spring with few
coils and a central seconds hand. After he had shown that it is
possible to carry time at sea sufficiently accurately for
navigation, other watch and clockmakers were able to devise
simpler suitable instruments.
His left hand lies on his thigh prominently in the
lower centre of the portrait. This hand has Heberdens nodes
on the middle joints of the fingers. Heberdens nodes are one
of the mildest and commonest forms of osteoarthritis. These
Dr. William Heberden (1710- 1801), English
physician, was born in London. At the end of
1724 he was sent to St Johns College, Cam-
bridge, where he obtained a fellowship about
1730, became Master of Arts in 1732, and took
the degree of Doctor of Medicine in 1739. He
remained at Cambridge nearly ten years longer
practising medicine, and gave an annual course of
lectures on materia medica. In 1746 he became a Fellow of the Royal
College of Physicians in London; and two years later he settled in
London where he was elected a fellow of the Royal Society in 1749,
and enjoyed an extensive medical practice for more than thirty years.
Heberden, who was a good classical scholar, published several papers
in the Phil. Trans. of the Royal Society, and among his noteworthy
contributions to the Medical Transactions (issued, largely at his
suggestion, by the College of Physicians), were papers on chicken-pox
(1767) and angina pectoris (1768). His Commentarii de morborum
historia et curatione, the result of careful notes made in his pocket-book
at the bedside of his patients, was published posthumously in 1802.
This was the last British medical paper published in Latin.
24
nodes begin to make their appearance in middle life,
particularly in women and manifest themselves as bony
outgrowths around the bases of the terminal phalanges of the
fingers. They are more prone to develop in people who work
hard with their hands. Also, the back of the hand is swollen
and the fingers have the typical fusiform swelling of the prox-
imal inter-phalangeal joints characteristic of early rheumatoid
arthritis. Rheumatoid arthritis is a collagen disease thought
to be exacerbated by fatigue and stress. Harrison in
Concerning Such Mechanism mentions that Dr. Heberden
(1710-1801) contributed ten guineas to the development of
the sea clocks.
The long case on the right hand side of the portrait is
elongated and the front fully glazed so that the detail of the
enclosed pendulum can be clearly seen above the chair. The
pendulum is drawn with careful detail. It is more complex
than any present in existing Harrison clocks and shows
features that indicate that it was designed at the time of the
portrait. The thickness of the brass rods are greater than the
steel and there are extra brass rods with a means for fine
adjustment. The driving weight is on the wrong side of the
resting pendulum. One of the regulators that Shelton made
for the transit of Venus in 1761 has a pendulum similar to
the one in the portrait but with no means of adjustment.
This portrait, following in the genre of Hogarths
Coram portrait, is a statement that Harrisons achievements
and riches have been gained by his own endeavours but at a
cost to his health.
Sir Joshua Reynolds Portrait
This like the Joseph Wright portrait is only known because of
a nineteenth century engraving by W.H. Worthington labelled
John Harrison. It shows Harrison as an elderly man. Like
the Joseph Wright portrait it does not give any information
about his creations but does give an indication of his
perceived importance at that time as it is difficult to find
another John Harrison of equivalent fame or fortune that these
eminent artists would paint or engrave.
Manuscripts and Pamphlets
The most important manuscripts written by Harrison
relating to this clock are his 1730 document and the late
25
pamphlet Concerning Such Mechanism. The 1730 manuscript is
in the archives of the Clockmakers Company (MS 6026/1).
It had been overlooked for sixty years until 1950. The
document describes the methods he has used to create an
accurate pendulum clock and how he can continue to use these
principles to make a reliable sea clock. It gives details of his
cycloidal cheeks to the pendulum suspension, his roller lantern
pinions, the setting out of his escapement, details of his
temperature compensated compound pendulum and his
proposed construction of a sea clock. The dimensions and
proportions shown in the associated drawings relate to the
1726-1728 wooden movements.
Further drawings in the Guildhall Library show the
geometrical construction of the cams of a remontoire, a
modified anchor escapement, and the geometrical construction
of a modified escapement. The diameter of the pitch circle of
the remontoire wheel in the drawing relates to the RAS
regulator.
An Explanation of my Watch published in 1763 is
important as it describes many of Harrisons principles and
their derivation (MS 3972/2) . Unfortunately the Trustees
of the Worshipful Company of Watchmakers will not give
permission for a photographic copy of this document to be
made to enable careful private study.
Concerning Such Mechanism published in 1775 repeats
most of the details about pendulum clocks given in the 1730
manuscript and describes his musical scale derived from B.
Graham had died in 1751, but he was still considered a
horological authority when Harrison wrote Concerning Such
Mechanism. Harrison, aged 82, wrote this book to present his
differing ideas so that they might be remembered in posterity.
On first reading it appears that Harrison has written the
pamphlet to denigrate Graham but he is careful to state that
he has discussed every part face to face with Graham. Parts of
this pamphlet are copied from An Explanation of my Watch. The
pamphlet also contains important details of the development
of H4, the watch that fulfilled the conditions of the
Longitude Act.
We may digress here, to observe how that modesty for which John
Harrison was conspicuous in his bright days had forsaken him when,
in 1775, at the age of eighty-two, he published his pamphlet entitled A
26
DESCRIPTION CONCERNING SUCH MECHANISM AS
WILL AFFORD A NICE, OR TRUE, MENSURATION OF
TIME: which also refers to the attempts to find the Longitude by the
Moon, and to a Scale of Music of his Invention. He would accept of
no assistance in revising this work, which is so encumbered by his
singular and undefinable manner of expressing himself in writing, as
to be unintelligible to the general reader without a translation; and
although it has valuable hints for those who are thorough masters of the
subject, it may be set down altogether as memento mori.
Johan Horrins, Memoirs of a Trait, 1835
A manuscript dated 1770 in the U.S. Library of
Congress in Washington titled A True and Full Account of the
Foundation of Musick, or, as Principally Therein, of the Existense of
the Natural Notes of Melody. This gives a fuller description of
his musical temperament based on pi as also described in
Concerning Such Mechanism, the shortcomings of integer
27
derived tuning systems, Church Services, Psalms, Parsons,
Choirs and the tuning and shape of bells using Radical
Numbers. In this document Harrison describes his
correspondence with Lionel Charlton.
Drawings were made of the sea clocks H2 and
H3when they were in Dents workshop for restoration in
1840. Most of these drawings are by Thomas Bradley.
(Cambridge University Library, MS. RGO 6/586f213-
217)
The Sotheran auction catalogue Bibliotheca Chemico-
Mathematica of 1922 in the U.S. Library of Congress has a
photograph of Harrisons manuscript and his notes of
Saundersons lectures. These lecture notes have been lost.
In 1972 a set of notes of Saundersons lectures was
transferred from the Patent Office to the British Library.
These notes have exactly the same content as the Harrison
notes photographed in the Sotheran catalogue. (Add.
57488). Another copy of notes of Saundersons lectures
is in the Bodleian Library, Oxford.
The remontoire in the RAS regulator is similar to
H3 and Bradleys drawing of the cam and detent are useful
guides. The RAS regulator was in Charles Frodshams
Lionel Charlton, (1720-1788), historian of Whitby, was
born at Hexham about 1722. Described as lame from his
youth, halting with one leg, and having one hand shrunk up.
His disability did not prevent him studying at Edinburgh
University and becoming a land surveyor and teacher of
mathematics. About 1748, Charlton settled in Whitby, in
North Yorkshire, where he would stay for the next forty years
running his school in the old town house. He wrote The
History of Whitby and Whitby Abbey which was published in York,
in 1779. Captain James Cook began his three year
apprenticeship as seaman based in Whitby in February 1747.
Cook was equipped with Kendals copy of H4 for his second
voyage to the South Seas (1772-1775); he wrote favourably
about its performance on his return. Kendals simplified
copy, K2, was used successfully by John Phipps for his
unsuccessful attempt to find a Northwest passage in 1773
and by Lieutenant William Bligh on HMS Bounty in 1787.
28
workshops in 1856 and 1860. Dents made the
Westminster clock, designed by Sir George Airy and
Edmund Dennison, Baron Grimthorpe, in 1858.
Frodshams workshop was two doors away from Dents in
the Strand.
The outer bearings between the minute and hour
hand coaxial shafts of the Westminster clock have four
roller caged bearings similar to those fitted by Harrison to
the great wheel arbor of his RAS regulator. Unfortunately
Frodshams archives contain no bench books or account
books relating to this period.
Gould made notes when he dismantled and restored
the RAS regulator in 1927. These notes are helpful but
deficient in many details. They highlight the difficulty in
adjusting the escapement if the pallet arms are not of the
correct length. Gould made accurate measurements of some
parts of the regulator when he was arranging to have a case
constructed. He lists all the parts as he dismantled the
clock and gives details of the parts that he added or had
remade.
Goulds 1927 Notebook
29
Nicholas Saunderson (1682-1739)
became blind at the age of one year after
contracting smallpox. He went to stay in
Cambridge University with his friend,
Joshua Dunn after studying mathematics
with William West. William Whiston,
Lucasian Professor of Mathematics,
allowed him to lecture in the University.
When Whiston was removed from his
chair, Saunderson was chosen to succeed
him. Notes taken at his lectures were
widely circulated and he wrote Elements of Algebra. John Harrison
made a copy of notes taken from Saundersons lectures. These
notes were sold in 1922 by Sotherans and are now lost. Sotherans
specialized in sales of books and documents to American buyers.
Saunderson communicated with Sir Isaac Newton and was a friend
of Halley and William Jones. William Jones son William (1746-
1794) was a philologist. A copy of Saundersons lecture notes
exactly similar to Harrisons was owned by another philologist,
Hannibal Evans Lloyd (1771-1847), a friend of Henry Fox Talbot.
Sotherans photograph of John Harrisons copy of Saundersons lecture notes
30
SP
GP
C
OP
0
M
Alt
am
pm
Meridian Passage
Nautical Astronomy
e
c
b
a
A
B
Decagon and Pentagon
Draw circle with centre O and diameter MN. Erect a perpendicular at O cutting circumference at
A. Bisect MO at P and with centre P and radius AP, describe an arc to cut NO at Q. Let radius =
1 then OP = . Therefore AP, the radius of the first arc = 5/2 and MQ = (5 + 1)/2 = and OQ
= 1/ .
OQ equals the side a the regular decagon completed by drawing chords equal to OQ, thereby
dividing the circle into ten equal parts.

Hippiass Quadratrix
This is a curve formed as the intersection of a radius and a line segment moving at corresponding
rates. Asquare and a circle are drawn so that one corner of the square is the centre of the circle,
and the side of the square is the radius of the circle. Bisect the right base angle and connect the
midpoints of the two vertical opposite sides. The intersection is on the curve. Then construct the
midpoints of top and bottom segments of the opposite sides and bisect the two new angles also.
The intersection of these produces two more points on the curve. The quadratrix is the locus of
the intersection points of both moving line segments. The upper endpoint is fixed but the lower
can never be attained. Using Cartesian co-ordinates the equation for the quadratrix is:-
x = y cot(y* /2)
This expression does not have any sense for y = 0. The limit for x as y approaches 0 is 2/ .

Golden Section or Extreme and Mean Ratio

Asquare drawn within a semicircle divides the diameter into three parts related by the golden
section. The ratio of a side of the square to the diameter is then 1:5
Divide a square of side AD = 1, so that AE =
By Pythagorass Theorem
DE = AD + AE = 1 + 1/4 = 5/4
Therefore
DE = EG = EH = 5/2
Draw a semicircle centered on E of diameter DE
Therefore
BG = AH = 5/2 + = (5+1)/2 =
Therefore
AG = BH = - 1 = 1/
2 2 2

All the irrational numbers that are
important for architectural proportions
are generated by simple constructions
involving the circle, square and right
angled triangle.
Pythagorass Theorem
The theorem proves the
irrationality of 2, 3, 5 and .
In right angled triangles, the square on the side subtending the right
angle is equal to the squares on the sides containing the right angle

The Square Root of 2

This is the ratio of the base of a right angles isosceles triangle to its two perpendicular sides, and
therefore of the diagonal of a square to its side, or of the diameter of a circle to the side of a square
inscribed within it.
The Square Root of 3
This is the ratio of the height of a regular hexagon to the side and the ratio of the diagonal of a
cube to the side.
Geometry Used
by Harrison



= 1 + 1/
= 1 + = 1 + 1 + 1/
= + = 1 + 2 = 2 + + 1/ = 5 + 2
2
3 2
1/2 = 2/2
Chapter 3
Development of the Pendulum Clocks
Introduction
At the end of the Thirty
Years War (1618-1648) men
returning home from hostilities in
Bohemia and North Eastern Aus-
tria (Niedersterreich)to the
Protestant Rhine Palatinate
expanded the production of
wooden clocks in the Black Forest
area (Schwarzwald). These clocks
were marketed by pedlars travelling
throughout Europe, particularly
the Low Countries by utilising
water transport on the Rhine.
Many of these clocks came to England through the ports
of Ipswich and Hull. They would not be able to enter
London easily because of the restrictions of the Charter of
the Worshipful Company of Clockmakers. Harrisons
wooden clocks have many features indicating an intimate
knowledge of these clocks (viz. tooth profiles, wheel
ratios, frame structure, and count wheel striking).
There are eight wooden framed pendulum clocks,
known to have been made between 1713 and 1730, that
have survived. The early clocks made before the birth of
his first son, John (1719-1738), are labelled J
n
o
Harrison
The clocks that he made during Johns lifetime are labelled
James Harrison and those after 1738 John Harrison.
Dr. Abraham Rees always refers to James Harrison in his
Cyclopaedia; it is most likely that he was known to his
family and close acquaintances as James, even though his
brother had the given name James.
The wooden framed clocks are:
(i) Clock, dated 1713 (incomplete movement in the
33
collection of the Worshipful Company of
Clockmakers, London).
(ii) Clock dated 1715 (incomplete movement only in
Science Museum, Kensington).
(iii) Clock dated 1717 (movement only at Nostell
Priory, Yorkshire).
(iv) Turret clock dated 1722 (modified by Harrison
after installation, Brocklesby Park, Lincolnshire)
(v) Longcase clock c. 1725-6 (movement only in
collection of Worshipful Company of
Clockmakers)
(vi) Longcase clock c. 1725-1726 (movement without
pallet frame and pendulum, and case,).
(vii) Longcase clock dated 1727 (movement without
escapement and pendulum, and case, private
collection).
(viii) Longcase clock dated 1728 (movement and case in
the collection of the Worshipful Company of
Clockmakers).
1713 Longcase Clock
Harrisons frame structure with three way mortise
and tenon corner joints seen in all his wooden pendulum
clocks is used in this clock. Central posterior and anterior
plates are attached to this frame. It has going and striking
trains. Dial work is mounted on the front of the anterior
plate. The anchor escapement is of contemporary shape;
the escapement wheel is made of brass and has 48 teeth.
1717 Longcase Clock
This movement, at Nostell Priory is in a newer
plain oak case. The face has an engraved brass chapter ring
with arabic minute numerals and standard roman hour
numerals. The chapter ring is divided in 10 minute
intervals. The corner spandrels are decorated with an
imperial crown. The calendar wheel aperture is in the
lower centre. The oak plates are 3/8th. inch thick. The
wheels are made of oak with inserted teeth. There are
34
ogival shaped pinions turned on wooden arbors. The 40
tooth escapement wheel is made of brass and mounted on
a boxwood arbor. Careful examination reveals the curved
impulse faces of both pallets. Even though they are worn
and have been repaired the pallets match the escapement
wheel tooth profile so that the escapement is designed to
behave more as a dead beat escapement rather than a recoil.
The pendulum fitted to the clock has a wooden rod with
lead bob. The dial work is mounted on the front of the
anterior plate.
1722 Brocklesby Hall Turret Clock
This turret clock follows the design of the
domestic clocks with added features to suit the greater
forces involved. The cuboidal frame has corner elbows on
its front and rear faces and extra bracing in the middle of
the side faces. The arbors are mounted onto plates that are
let into the frame with open faced mortice and tenon
joints and held by boxwood screws.
There is shouldering of the top and bottom rails
to protect the end grain when dismantling. The timber
house framers system of numbering as still used is applied
to these joints. The wheels are laminated with oak faces;
the radial segments have teeth with radiused extremities as
in contemporary millwrights practice and are probably
made of apple wood. The pinions are made of brass. The
pendulum suspension has pivoted cheeks. The drums are
grooved with a right hand helix.
Harrisons development of an escapement
continued with this clock. The torque applied by an
escapement wheel to the impulse face of the pallets is a
tangent to the tip of the tooth. On the entry pallet with a
flat face the angle between the impulse face and the
tangent at the tip of the tooth increases by 1.9 during its
action. On the exit pallet, the impulse angle reduces
slightly by about 0.4 during its action. This results in a
differing impulse between the entry and exit pallets. The
curvature on the impulse faces of the Nostell Priory clock
increases the angle of incidence of the impulse face during
35
the action of the entry pallet and reduces the angle of
incidence during the action of the exit pallet.
There are two similar drawings of an anchor
escapement made by Harrison that have a curved exit
pallet. One of these drawings incorporates tangential lines.
The curvature on the impulse face of the exit pallet
decreases the angle of incidence slightly (0.4) and thereby
gives a more equal impulse but also increases the friction.
Harrisons drawing has been copied and reversed to
measure the angles subtended by the faces. It is suggested
that this is the escapement that was originally fitted to the
Brocklesby Hall clock and which did not perform
satisfactorily most probably because of the inefficiency of
the exit pallet .
Harrison modified this escapement by hinging the
pallets. It continued to function as a recoil escapement
and the pallets were released on recoil. Composers
controlled the position of the pallets. The torque circles
36
show that this modification gave nearly uniform impulse
with the entry pallet but an increasing impulse with the
exit pallet.
The improved efficiency of
the escapement meant that the
damping of the pendulum had to be
increased by fitting a vane on the
bob to increase air resistance. The
escape wheel is made of brass, the
going and striking train have
wooden wheels and the leading off
wheels are of cast brass. The arbors
have brass pivots running originally
in boxwood bushes. The escape
wheel arbor is supported by a single
friction wheel and has a lantern
pinion with rollers. These features appear to have been
added when the pallet arms were hinged because the
escapement then had no drop to reduce friction. The
lantern pinion with rollers was also added at this time.
The clock has a count wheel striking train.
1725-1730 Longcase Clocks
These four clocks have a very similar construction
and must be the clocks with which Harrison developed his
asymmetrical escapement and gridiron temperature
compensating pendulum. It has been suggested that the
first two of these four clocks were fitted with escapements
similar to that of the Brocklesby Hall clock and that they
were changed later as part of the development of the
asymmetrical escapement. There are paired friction wheels
supporting the escapement wheel in these clocks as
illustrated in the 1730 document. A calendar wheel is a
feature that is included in all the domestic pendulum
clocks that still exist.
Two of these clocks have extended cases to
accommodate the pendulums large amplitude indicating
that changes were being made to the escapement during
37
this period.
The features of these clocks are described in the
1730 document. Lantern pinions are fitted throughout.
The bolt and shutter maintaining power and count wheel
striking train are similar to his earlier clocks. In the latest
clock the drums are grooved with a left hand helix. The
pendulum suspension is similar to the Brocklesby Hall
clock with additional adjusting screws.
Let A represent two bits of Glass, whereon each end of the
Axis of the Pendulum being a Sharp Edge of Brass moves in a little
notch in the Glass, which renders that Friction insensible, as also the
Wearing (for tis but the Axis of the Pallats, which communicates force
to the Pendulum; for the Pendulum it self is otherwise suspended, viz
by a thin Brass Spring). Let the Arch CD, and the least Vibration the
Pendulum must make to let the Wheel move forward, be similar;
therefore the Arch a.a. is described with the same Vibration by the
Center whereon the Pallats have their relative Motion, and from the
extremities of the said Arch or Vibration a.a. and with the length of
each Pallat may be described the portions of Circles ee.ff. which must
cut the Limb of the Wheel so, that there be half the space of one of the
Teeth between; because one Vibration lets but the space of half a Tooth
go. Now if these Pallats were fast at their relative Center a so that they
had no relative Motion, their Friction at the Wheel would be equal to
that of the common way, if the Arch described by Vibration was as
large, and maintaind with the same draught of the Wheels; but that
cod not be, for twoud require a deal more. But without, that, as their
38
length or Radius is at least 60 Times the semidiameter of the Pin, on
which they relative move, I may safely say (as in sect. 5) the friction is
40 Times less than the other, and the wearing is still less, especially in
regard to the Motion being alterd thereby; for the Wheel pressing
against the End of the Wood (or the Pallats being made lengthway of
the Wood) there will scarce be any impression made in 40 or 50 Years
for the Wheel has no drop, and its draught is but very little, in
comparison to what it would require to have with the other Pallats.
But suppose it does make a little impression; since the extremity or
Edges of the Teeth has not to do, as the common way has, i.e. to slide
out of the little hole thats worn, and then over it; for here the tooth
takes the Pallat along with it, and never departs from the same relative
place, until the other Pallat take hold, and then it departs from it
perpendicular to the surface of the Pallat; therefore no sliding, and
consequently no alteration of the Motion on this account. But again, if
we suppose the Pallats to alter in length the quantity of a little
impression and a little wearing at their Center of relative Motion; yet
this, as they are applyd like Tangents to the Wheel, can make no
sensible difference in the impulse at the Pendulum in many Years. But
these we cannot have in the Sea Clock but we may have the like in
Nature, as explaind,
John Harrison, Concerning Such Mechanism 1775
The escapement is now asymmetrical with both
pallets mounted on the same side of a balanced pallet
frame. This pallet frame has an adjustable counterweight
1728 Escapement with both pallets in contact with escapement
wheel just before release of exit pallet
39
on the other side from the pallets. This construction
means that the escapement wheel is pushing the entry
pallet and pulling the exit pallet. Although Harrison made
a drawing, he does not discuss the mechanics of this
escapement, which he refers to as his contrivance in other
parts of the 1730 document and the Concerning such
Mechanism pamphlet.
The Marine Clocks
Harrisons 1730 document is a description of how
he could construct a pendulum clock that would keep time
accurate enough to fulfill the conditions of the 1714
Queen Anne Longitude Act. He had made his first clock
in 1713 and it had taken him until 1728 to sufficiently
improve its accuracy. No other clockmaker up to 1730
had made a timepiece of equivalent accuracy. Having
surmounted that hurdle he continues by suggesting that he
can use similar methods to produce a sea clock using
linked twin pendulums. His description of the proposed
sea clock is divided into two parts by details of his
temperature compensated pendulum. He describes the
action of the pallets and states that this method cannot be
applied to single pendulum. There is a thumb-nail sketch
in one of Robert Hookes notebooks suggesting a similar
method of construction. Burgis vertical crossbeat
escapement clock has similar features.
Let AA be two Axes for two Ballances (or such artificial
Pendula as explaind below Fig. 6) we need not signify whether they
be Horizontal, or Vertical, but parallel. Let them have at each end in
the Centers of the Pevets a Brass Wire, of a competent thickness and
length, fixd fast, and then stretchd stark in the same direction with the
Axis, by which means the Axes cannot shove end-way, and the Wires
will twist with an elastick force to the Vibration. But these Axes will
be drawn towards each other, (as below Fig. 6) from which they must
be suspended by Rolls at their Pevets, (as mentiond in the last section)
but here let them be only portions of Roll of great Radii, so the Friction
at their Centers will not be felt, and we may suppose these Ballances to
move nearly as free as a Natural Pendulum suspended by a spring. But
we cannot communicate em to the Wheel, as treated of sect 23.
Therefore, as one Axis or Ballance is to move one way, whilst in the
40
same Time the other moves the other way, (as will be shewn Fig. 6) let
there be fixd to each Axis at b a fine small Brass spring, and let it be
exactly fitted to the portion of the Circumference of the Axis from b to
c, leaning against the Axis; then at c let it go off in a Tangent (to the
Axis) to t, and supposing the Ballances at one extremity of their
Vibration, one of the springs will take hold of the extremity of the Tooth
of the Wheel 0, where tis also a Tangent to the Wheel; then the Wheel
Moving forward, or towards F, the space of half a Tooth, and drawing
so much of the spring from off the Axis, and the Ballances performing
one Vibration as from D to C, (for we will suppose that Arch and the
Vibration similar) then the other spring will take hold of the Tooth d,
and there drawing the Wheel a little back, the other spring will fly off
to z, i.e. twill return into the position it was in, before the Wheel drew
some of it from off the Axis. Now these Pallats will have less Friction
than the other, Fig. 2. But they cannot be applyd to a Natural
Pendulum. And the Wearing will also be insensible; for here is no
sliding nor droping, but they take and leave as the other do.
Let AA be 2 Axes, on which 2 Ballances or the 4 Pendulum
Balls BBBB are supported. Let the Balls be all of an equal Weight
exactly; and also at equal distance from their Axis or Center of Motion
to that of Vibration or Gravity. Let them be communicated to their
Axes with Wires, such as in Fig the 5. So that they be nearer the
41
Center of Motion when Warmer; but rather somewhat neare than in the
Natural Pendulum; because some part of their support, (or rather a
bigger part of it than in a Natural Pendulum) will not by this means
be brought nearer. But of their support in particular I shall not here
enlarge: but supposing it both possible and practicable for any two of
them, both singular and together with what supports :em to he made so
exact an Equilibrium as required, I shall proceed. Let the two portions
of Circles ab and cd be of equal Radii, and fixd to each Axis with their
Centers exactly in the Center of Motion. Let two small Brass Wires,
but flat (Or thin like the Pendulum or Spiral Spring of a Watch) and
Elastick; be stretchd from a of one Arch to c on the other, and from b
of one Arch to d on the other, by which means one Ballance cannot
relative move one way without the other move the same space or
quantity the other way, so consequently if the whole Body of the Clock
be turned one way (whether it be slowly, or faster than ever the Ship
can turn it,) any portion of the Circle, (whether coinciding with the
plain in which the Ballances Vibrate, or inclining thereto) it cannot
alter the relative position of the Ballances, but they will still remain as
if the Clock had not been moved. Let SS be two Worm Springs, and let
each be suspended by the middle at f let them have fixd at each end at
e a small Wire, such as was spoken of last, by which from g, in the
portions of the Circles g h I, let the springs be stretchd to a convenient
degree, that so as the Ballances Vibrate, one one way, and the other the
other way, one spring will be stretchd out, whilst the other draws it self
in, and the thin Wires will be constantly applyd to the Arches g h I,
and (the Ballances performing two Vibrations in one Second of Time)
the said Arches will be portions of Circles of about 10 Inches Radius:
So the Angle of Contact will he very small, and as the Wires are thin
and Elastick, the Motion will he very little obstructed by their
application to the Arches; therefore almost as free as that of a Natural
Pendulum. And the wearing may be compard to that of the spring
which suspends a Natural Pendulum in a Cycloid which is insensible.
But supposing the Motion not so free as a Natural Pendulum, if it do
not alter in degrees of freeness, it is the same in respect to the truth of the
Motion, as if it was entirely free; and we must suppose that if it do
alter, it will be as the degrees of Heat or Cold. Therefore it may be
attributed (in regard to the Motion) to heavier or lighter Air, which is
accounted for above. But this may seem to be super added to the Airs
resistance, I own it is, yet both together will not as performs 2
Vibrations in one second, as 1 Pendulum Second) meets with from the
Air alone, supposing em to describe similar Arches: But whether it be
more or less, the difference in its resistance may be accounted for in the
ordering and adjusting the Wires (sect 26. 20) But greater or less
42
Vibrations may not be performd in equal Times, no more than in a
Natural Pendulum without a Cycloid. And here whether greater
Vibrations take more or less Time than less ones, they may he reduced
to exact equality by part of the Arches gh I, viz, that part from h to I
being made a portion of a Circle, respecting some other Center k, or if
not so, some other Curve. But if it be objected, that the springs SS
which are treated of above will be rather weaker when Warmer, I
answer the Arches g h I (which are part of the support of the Balls, but
not fixd fast to em) will be farther from the Center of Motion at the
same time; Therefore will almost countervail what wants in the springs;
and the Complement as above; viz the Balls rather somewhat nearer
than in the Natural Pendulum; etc.
In the last, sect is shewn, that the Ballances cannot alter their
relative position by turning the Whole Clock into another position, but
that is supposing the Clock not .in Motion, so consequently, if it were
in Motion, the Motion would not be alterd thereby. But the Clock may
be suspended from the Ship with Hinges, in Nature to them of the
Mariners Compass, and so that it alter its position but very little, tho
the Ship Toss much, nor receive any great Shocks from the Waves.
John Harrison 1730 Document
The First Marine Clock (H1) 1729-35
The first marine clock made by Harrison has two
large compound straight bar balances mounted on
anti-friction arcs, which are connected by cross-wires
running over brass arcs. They swing, as if geared together,
to minimise the effect of a ships motion on their period
of oscillation. They are controlled by four helical springs,
in tension. A triple grid-iron of brass and steel rods
varies the tension of these springs to compensate for
changes in ambient temperature affecting the springs.
The wheels (except the escapement wheel) are
made of oak with the teeth morticed into the rim so that
their grain is radial. They are all mounted on anti-friction
wheels. There is no remontoire. Two mainsprings drive a
single central fusee, which has a double helix to
accommodate the drive chains. The fusee has an internal
maintaining spring. The clock goes for 38 hours on
winding. Each balance has a single escapement pallet which
is pushed by the escapement wheel. Dials enclosed by a
43
Jost Burgi (1552-1632), a Swiss
mathematician, worked closely
with Kepler. He derived
logarithms independently of
Napier from Keplers exposition
log(x.y) = log(x)+log(y)
f(x.y)= f(x)+f(y)
He was appointed clockmaker to
the Landgraf William IV of
Hesse in 1579. The Landgraf
measured the Right Ascension of
stars by measuring the time
elapsed between the meridian
passage of the sun or other
known star and the star in
question. This method required
more accurate clocks than those
available with the foliot or
horizontal balance. Burgi
developed his vertical cross beat
escapement to fulfil this
requirement. A clock, attributed to Burgi, now at Dresden has a 360 tooth
escapement wheel with the pallets in the same plane. The two compound
pendula are geared together. The clocks action was described by Hevelius and
may account for the accurate results obtained by him and Tycho Brahe. It may
be considered to be one of the earliest pendulum clocks and may have been seen
by Halley when he visited Hevelius in 1679. This unsigned clock, if made by
Burgi is a remarkable innovation; it has a remontoire, count wheel striking, anti-
backlash spring and hour, minute and seconds indication. Harrisons first two
sea clocks have exactly the same principles of operation; he added temperature
compensation, replaced the interconnecting gears with metallic bands and
replaced the verge pallets with his non-sliding pallets.
Now these Pallats will have less
Friction than the other, Fig. 2. But
they cannot be appld to a Natural
Pendulum. And the wearing will also
be insensible; for here is no sliding or
droping, but they take and leave as the
other do.
John Harrison, 1730 Manuscript
44
ellipsoidal surround indicate hours, minutes and seconds;
it has a calendar dial like his pendulum regulators. For use
at sea, it was enclosed in a wooden case suspended by
springs on a gimbal frame.
Second Marine Clock 1737-1739
This clock was built after Harrison had moved to London.
It is basically similar to his first marine clock but made stronger
and heavier. It has a rectangular frame which has the proportions
of a golden rectangle. There is no calendar dial. All the wheels are
made of brass and the grid-iron temperature compensator is
simplified and made adjustable. A remontoire is fitted which
rewinds every 3 minutes (16 times each hour). It is fitted with a
spring maintaining mechanism. A silver dedication plate with
extremely fine engraving is attached to the top of the front plate.
The Third Marine Clock (H3) 1740-1757
The compound pendulums (or bar balances) of the two
previous sea clocks are replaced by two large balance wheels, which
are connected by cross-wires. the period of oscillation is markedly
increased and they are controlled by a single balance spring, whose
effective length is varied by a simple bi-metallic compensation curb
made of brass and steel. The remontoire rewinds every thirty
seconds and produces an absolutely constant torque at the
escapement wheel by its helical springs acting on a cycloidal cam.
There is a poorly understood isochronicity arm which is pivoted
on roller bearings. Temperature compensation is by means of a bi-
metallic strip operating on the balance spring. The plates are in the
proportion of a golden rectangle.
And indeed Velocity was very much wanting in my three
large Machines, yea, I say, very much, notwithstanding their
Weightiness of Ballances, or as notwithstanding what Philosophers may
reason in other Affairs, or what Philosophy in this Affair [viz. as
hitherto through Mistake, as from the Steel-Yard, &tc. in the matter]
might teach us to reason; but I did not then [viz. as in the Designs of
my three large Machines] so thoroughly understand it, nor should I, or
the World, (as I think I may make bold to say) ever have understood
it, had it not been, or had I not discovered it as it were through
Accident, in or by my third Machine.
John Harrison, Concerning such Mechanism, 1775.
45
The Steel-Yard, or Stahlhof, was the self-governing,
enclosed enclave of the German Hanseatic League in the City of
London. The land where it stood is now occupied by Cannon
Street station. The Hanseatic League provided communication
between Northern Germany and Britain during the Napoleonic
Wars and was a means by which Hooke and Harrison could have
become acquainted with Jost Burgis work. There is a canister
clock in the foreground of Holbeins 1532 painting of George
Giese, a Steel-Yard merchant.
Jeffreys Pocket Watch 1752
Harrison must have prospered well after he moved to
London in 1737, for in 1752-3 he was able to have a very
expensive silver pair-cased pocket watch made to his own design by
John Jeffreys. Jeffreys died in 1753 and so would not have
contributed to the making of H4. Larcum Kendal, who had been
apprenticed to Jeffreys, was asked by the Board of Longitude to
make a copy of H4 which he finished in 1769. It may be that
Kendal was asked to undertake this work because he had been
engaged by Harrison in the making of parts for H4.
The most noticeable feature of this watch as seen in the
King portrait is the sweep seconds hand, emphasised by its shadow
from light from the left whereas the rest of the portrait is lit from
above. Graham had also made a watch with a centre seconds hand.
Other notable innovative mechanical features of this watch are
maintaining power and thermal
compensation. Internally cut teeth
on the third wheel allow the train
to recoil easily with the
escapement and the centre seconds
hand to move clockwise. The
verge escapement has modified
diamond pallets which allow a
much larger amplitude than was
usual at this time because both
sides of the pallets contact the
crown wheel.
This watch was inherited by his grandson, who later
related that it was always carried by his grandfather apart from
when Admiral John Campbell (1720-1790) had borrowed it to
determine longitude. The performance of this watch, together
46
with the use of vertical balance wheels in the third large sea clock,
is almost certainly the accident that stimulated Harrison to create
his Longitude timekeeper H4.
The smaller watch intended for the pocket, and the original
of the chronometer used for finding the Longitude at this day, is now in
the Authors possession. It was made under the Inventors inspection
by a clever workman, whom he allowed to put his name on it, viz. John
Jeffreys, which is repeated on the cap, with the addition of the date, and
this being 1753, shows it to have been constructed two years prior to its
being brought forward. It was always John Harrisons pocket watch,
except when Admiral Campbell borrowed it to find his Longitude by,
for which it answered nearly as well as the larger but more expensive
Timekeepers.
Johan Horrins, Memoirs of a Trait in the Character of George III
The Marine Watch (H4) 1757-1761
The introduction of the vertical pendulum controlled
by gravity had improved timekeeping by an order of
magnitude greater than that obtainable with a foliot. In
order to convince the Board of Longitude of the practicality
of carrying time on a ship by a mechanical device he had to
over engineer a watch that he would present for a trial. After
he had demonstrated that it was possible, succeeding watch
and clockmakers were able to successfully simplify the
Admiral John Campbell (1720-1790) was the first navigator to use
Hadleys quadrant to measure lunar distances in 1747. His
observations were favourably compared to those of James Bradley, 3
rd
Astronomer Royal. Bradley suggested to the Board of Longitude
that Campbell should trial Mayers tables whilst commanding HMS
Essex in the blockade of Brest. He used lunar distances to correct
time by his watch and then determined longitude by equal altitudes.
Following this trial he suggested that the arc of the octant should
be extended to 60. John Bird (1709-1776) used Campbells
recommendations to produce the prototype of the seamans sextant
having a 20" radius arc and open brass frame.
47
Diamonds have a face centered cubic crystal structure and
readily produce octahedral shapes. They have perfect cleavage
in four different directions. The lines of cleavage result from
the diamond crystal having fewer chemical bonds along the
plane of its octahedral faces than in other direction. Diamond
cutters take advantage of lines of cleavage to facet gemstones.
Harrisons pallets are cuboidal with one curved facet which is
cut across the lines of cleavage. This curved facet will exhibit
less friction and wear better than any other known material.
mechanism.
This large watch (a coach or sedan clock, ger.
Satteluhr or Kutschenuhr) must have been very expensive for
Harrison to produce. The enamelled dial is over 120 mm. in
diameter, the largest produced at that time. He fitted
diamond pallets with a curved shape to ensure isochronicity.
His early pendulum clocks illustrates his experimentation
with the shape of pallets. He was able to fit a remontoire
with a spiral driving spring. This did not provide uniform
torque during its cycle but the average power delivered is
uniform. Friction losses in the drive train were reduced by
extensive use of pierced ruby bearings having diamond
endstones.
The Construction of the Pallats of my now small Machine,
Watch, or Time-keeper for the Longitude, and as especially in my last
Drawing, viz. So as wherein or whereby exactly to suit in each other
or all Respects, as well as my other Pallats, to the Nature or Property
of a Pendulum, [viz. More properly than as in my first Watch, nay,
somewhat better, or more to the Purpose, than as at present in my
second Watch] and still as not to pass
unnotified, The Materials of which the
Ballance-Wheel and Pallats are made, viz.
the Wheel of Steel, quite hard, and the Pallats
of Diamond, and as whence, so far as I am
hitherto able to judge, they will hold their
Figures for Ages; neither will the Watch in
any other Respect [but as chiefly from my last
Drawing, and where the Pallats will also be
somewhat easier to be done] hardly ever be out of Order, but as above,
to point out the Time [and that whether at Sea or Land] to a Second
in a Fortnight.
John Harrison, Concerning Such Mechanism, 1775
48
Royal Astronomical Society Regulator 1750-57
This large brass clock was most probably made by
Harrison at the same time that he was making his third sea
clock; there are many similar features in the design of these
clocks. The escapement is a further development of the
asymmetrical design found in the four 1725-30 wooden
framed clocks. The RAS regulator could also have been used
to test the suitability of a spiral spring powered remontoire
for H4. Many of the features of this clock are described in
Concerning Such Mechanism and it is almost certainly the
prototype of the clock he intended to build for the Royal
Greenwich Observatory
But as farther, upon due Reflection, it is certain that a Clock may still
come nearer the Truth than any present Clock, the which I have here
been speaking of; and, as towards the matter, I have for some Time has
such a Clock to the Purpose in great Part made; but as not designing
to fix it up in the not rightly convenient Place or House in which I live,
I did not hasten in finishing; as wanting withal [viz. as lately had
come in my mind] some other or farther Experiments trying with my
present going Clock, and they as to the better Completion of my other
Clock, or of any such hereafter.
John Harrison, Concerning Such Mechanism 1775
49
Robert Hooke (1635-1703), English natural scientist and architect, worked
as an assistant to John Wilkins, John Willis and Robert Boyle. He was
appointed Professor of Physic at Gresham College, London in 1665 and
Secretary of the newly formed Royal Society of London for the Improvement
of Natural Knowledge in 1667. After the Great Fire of London he was
appointed a City Surveyor and designed the Bedlam Hospital in Moorfields
and the first Montagu House. He joined Sir Christopher Wren to design
the Monument as a zenith telescope.
The critical component providing accuracy in 17th and 18th century
clocks was the pendulum, which itself was dependent on gravity as part of its
driving/controlling force. Robert Hooke voyaged to the West Indies in
1662, discovering how gravity changes (less strong on the equator) and
humidity interfered with the accuracy of a
clock. He also realised the movement of the
ship added additional errors to the
pendulum's swing. Henry Sully published
Regle Artificielle du Temps in1727 and says that
Hooke produced a double balance in1675
which he had devised in 1658. This was
described further and altered by Thiout and
Du Tertre. He then designed a marine
chronometer employing these improvements
and created a pocket watch utilising his
compensating devices. The new timepiece was
demonstrated on 20th Feb. 1668 to the Royal Society.
Hooke anticipated Newtons inverse square law in gravitation and
is thought to have played a significant part in
the development of the anchor escapement for
clocks. He demonstrated the linear relationship
between stress and strain in a helical spring,
which is now known as Hookes Law, and
demonstrated simple harmonic motion in the
oscillations of a weight attached to a spring.
Using a heavy bob suspended on a long ribbon
draped over curves having classical conic form
he demonstrated the elliptical motion of the
planets. There is a drawing in the margin of one of his notebooks showing
coupled compound pendula linked by helical springs - a system similar to
that used by Jost Burgi in his Dresden clock.
50
Chapter 4
Royal Astronomical Society Regulator
History of the RAS Regulator.
The Royal Astronomical Society regulator was inherited
by William (1728-1815), John Harrisons son, by his second
wife, Elizabeth Scott, whom he married in 1726. Williams
1814 will bequeaths my said wife the use during her life of my two
valuable clocks, telescope etc., and after her decease I give them to my grand-
daughter, Elizabeth Barton. Williams daughter, Ann, pre-deceased
him and her death was probably the stimulus for William
writing his last will. Ann had married Sir John Barton and they
had a daughter, Elizabeth Adelaide.
Sir John Barton was Deputy Comptroller of the Royal
Mint, secretary and treasurer to William, Duke of Clarence, and
later to Queen Adelaide, his wife. He invented a differential
screw micrometer and was granted a Patent in 1822 for his
method of engraving steel buttons with 200 lines per inch.
This was the forerunner of the diffraction grating. He died in
1834. The regulator was given to the Royal Astronomical
society in 1836-7 by W.H. Barton of the Royal Mint.
A description of the regulator by H. Cottingham,
F.R.A.S., was published in the Monthly Notices of the Royal
Astronomical Society in 1909 and reprinted in the Horological
Journal in 1910.
There is evidence that only the unfinished movement
came to the Royal Astronomical Society. The clock front plate
The Analytical Society. Charles Babbage, John Herschel and
George Peacock were family friends of the Bartons. They were
founder members of the Analytical Society formed in 1812 to
promote Liebnizian calculus. The Society became the
Cambridge Philosophical Society in 1832, incorporated in a
Royal Charter granted by William IV. Notable students of
George Peacock (1791-1858), Lowndean Professor of
Astronomy 1837-1858, were Augustus de Morgan, Arthur
Cayley and George Biddell Airy.
51
bears the scratched inscription John Beresford 1856 and 60 on the
front and Lord Beresford CFs 84 Strand on the back. This refers
to C. Frodshams workshop. Charles Frodshams great
grandfather, William, was a journeyman who had worked with
Justin Vulliamy before setting up his own business at the age of
68 at 12 Kingsgate Street, Red Lion Square. He is known to
have been friendly with John Harrison and Thomas Earnshaw.
Charles Frodsham (1810-71) was apprenticed to his father and
became a Fellow of the Royal Astronomical Society in 1823.
He set up in business on his own in 1834 and purchased John
Arnolds business at 84 Strand in 1843. The firm was known
as J. R. Arnold, Charles Frodsham until 1858 when Arnolds
name was dropped. After Frodshams death his business was
continued by his third son, who had been named Harrison Mill
Frodsham.
Lord John George Beresford (1773-1862), protestant
Archbishop of Armagh, appointed and supported Rev. T. R.
Robinson, Director of the Armagh Observatory. He was a rich
and generous patron who took a great deal of interest in the
Observatory and science. He contributed several thousand
pounds of his own money to refurbish the Observatory with the
best possible instruments that were available. The Observatory
has two regulators made by Thomas Earnshaw which were fitted
with jewelled, small arc Graham dead beat escapements (0.5
amplitude). The first originally fitted with a nine-bar Harrison
gridiron pendulum of steel and brass rods having the same
configuration as the pendulum in the King portrait. It was
replaced in 1830 with a mercury-compensated pendulum. The
second has a five bar compensating pendulum of steel and zinc.
Lord Beresfords donations purchased instruments for
the Armagh observatory; was he paying for the RAS regulator to
be made functional for the Royal Astronomical Society or for
the Armagh Observatory?
Cottingham got the regulator working in 1909 but
shortly afterwards it was broken by someone attempting to start
it with the remontoire unwound. Lieutenant Commander
Rupert Gould restored it in 1927 and wrote a short description
in the Horological Journal in 1932 as part of his account of the
restoration of Harrisons third marine timekeeper (H3). Gould
corresponded with Cottingham at the time of this restoration,
and used Cottinghams 1909 photographs in his 1932
description.
Goulds notebook describes the order in which he
dismantled the regulator in 1928 and gives an inventory of the
52
parts that he cleaned and those that he had remade. He
measured parts of the clock to arrange for a case to be
constructed so that it could be moved and used at the Royal
Astronomical Societys premises in Piccadilly. He added a
stopping mechanism and changed the crutch to a heavy
Continental style brass one with a different method of
adjustment, thereby altering the balance of the pallet frame of
the escapement. He experimented and fitted a gridiron
pendulum and renewed the weight and pulley. He also had to
make changes to the escapement when he changed the design of
the crutch. Part of the remontoire mechanism and of one of the
roller bearings had to be remade because they were broken when
he reassembled the regulator. The remontoire fly was also
replaced at this time.
The clock was transported on 31st January 1929 to
Burlington House. After being exhibited at the Annual
(Anniversary) Meeting of the Royal Astronomical Society in
London on 8th February 1929 it was stopped and taken down
to repair the defective pendulum that had been made by J. Ager
Baugh for Gould. About three weeks after the clock had left
Goulds workshop, the latter was partially wrecked by fire (18th
February 1929). The floor collapsed within a foot of the spot
where the clock had been standing. It had been there for nearly
two years and had not been insured.
The regulator was examined by Quill and Laycock for an
exceedingly short time when it was moved from Herstmonceux
in 1972 to the National Maritime Museum. A new hardwood
case was made to house it for the exhibition in 1976 commem-
orating the death of Harrison. The acidic atmosphere within
this case caused efflorescence of the lead counterweights of the
escapement (Lead acetate trihydrate, Pb(CH
3
COO)
2
.3H
2
O).
The regulator was in the National Maritime Museum
until 1980 when it was moved to Flamsteed House. It is now
housed in a glass cabinet as suggested by Gould in
correspondence to the Royal Astronomical Society in 1927.
Further changes have been made to the driving weight. In 2006
the gallery in which the Harrison clocks are housed was
refurbished and the pendulum mounting block firmly mounted
on a brick wall as Harrison originally intended.
The Pendulums being suspended by the Clocks, and Clocks by
the Cases (as all commonly are) tho the Cases were very good and
firmly fixd to a Brick Wall and the Pendulums near the back of the
Cases; yet the Cases certainly yielded more to the swing of the
Pendulum, when the Weather was warm and dry (by which the Clocks
53
went slower) than when moist; tho never at any Time could the Cases
be discernd to stir in the least But this I accidentally discovered by one
of the Weights swinging a little, which was, when its string (measuring
from its center of Gravity to that of suspension) was equal to the
length of the Pendulum. This I compard to an Unison in Musick, but
iudgd it to proceed more from the stirring of the Case, than from the
Air; because the center of suspension of the Weight was much lower than
that of the Pendulum, so likewise its center of Gravity or Body.
Therefore that I might not be deceivd by the suspension of the
Pendulum, I made two holes in each Clock-Case Back, through which
into the seams of the Bricks I drove two strong Irons, without touching
or depending on the case at all, to which I screwd the plate by which the
Pendulum and Cycloid are suspended.
John Harrison, 1730 Manuscript
Description of the RAS Regulator.
The Harrison RAS Regulator is a large, single train,
eight day, weight driven plated pendulum clock with a one
second beat; it has classical proportions. The pillars are riveted
to the rear plate in the traditional English manner. The
workmanship throughout is outstanding. It is constructed
principally of free cutting cast and plate brass. The dials are
engraved, filled with black wax and silvered. The minute hand
is double ended, sweeps arabic numerals and makes a complete
revolution in two hours. Hours are indicated with Roman
numerals on a revolving dial below the minute hand. Above the
minute hand seconds are indicated by a moving four armed
pointer over arabic numerals in a protected cutout of the main
dial. A further window indicates the day of the month in arabic
numerals. Harrison indicated the days of the month in a
similar way in his earliest extant clocks. The mechanics of the
clock are divided into the three logical divisions of power
supply, oscillator and integrator. The power supply, that is
drum gearing and remontoire is confined between the plates.
The oscillator, that is the pendulum and escapement is mounted
posterior to the rear plate, and the integrator, that is the dial
work is mounted anterior to the front plate. Harrison has
applied different but appropriate engineering solutions to each
of these divisions.
The great wheel at the bottom of the clock on the centre
line runs in roller bearings, the front having a larger diameter
than the rear, and each having four rollers. The relatively large
54
surface area of contact of the rollers accommodates the driving
weight. No lubrication is used and they are fitted with covers
to keep out dust and dirt. Torque can be applied to these
bearings in both directions to enable rewinding. A substantial
girder, supported by Tuscan pillars is a safety device beneath
the drum in the event of failure of the roller bearings. The
stopwork is mounted on this girder.
The centre wheel is fitted with a helical maintaining
spring adjacent to the rear plate and is pivoted on brass friction
wheels. The two pawls acting on the maintaining power ratchet
wheel are gravity operated and like the pawls on the great wheel
act a half tooth apart. The arbor is made of gunmetal. The
friction wheel pivot axes are nearly equidistant from the arbor
and the three axes describe a right angle as drawn in the 1730
manuscript. The bisection of this right angle is parallel to the
line of force applied to the arbor. Accurate depthing of the
wheels can be easily achieved by adjusting the diameter of the
friction wheels. The friction wheels are mounted on wooden
collets which run on brass pins. The arbors are held against the
friction wheels by the torque of the driving weight or the
maintaining spring and will therefore only allow torque to be
applied in one direction. The dial work mounted on the front
of the front plate is driven by the centre wheel.
The third wheel is offset to the right hand side of the
clock and is mounted on friction wheels. It drives the
remontoire wheel to which a fly is also attached, mounted on the
left hand side of the clock. The remontoire wheel arbor is
co-axial with the escapement wheel arbor and they are mounted
on separate friction wheels pivoted so that the major forces are
the mass of the remontoire and escapement wheels.
The wheels have lantern pinions fitted with rollers made
of lignum vitae. They have a distinctive grain, pink colouring
and a matt finish. The ratios of the pinion to the wheel of the
great wheel and the third wheel are not integers, resulting in the
same rollers and teeth engaging throughout the clock only after
four revolutions of the great wheel.
The posterior end bearing of the great wheel is made of
lignum vitae. The end bearings of the centre, third and
escapement wheels are blued steel, but there is evidence that
these are replacements. They were most probably originally
made of lignum vitae. The remontoire arbor is co-axial with the
escapement arbor and its side thrust is taken by the friction
wheels running in grooves in the arbor. Forward movement of
the centre arbor is prevented by the side faces of its friction
55
wheels running in grooves.
The pallet frame and escapement is pivoted in the
midline with blunt bronze knife-edges onto v-shaped glass
mounts. These pivots are designed to accommodate side thrust.
The pallet frame has a lead counterweight.
The pendulum mounting bracket with its radiused
cheeks is made to be mounted on a substantial wall. It has two
tapered mounting pins to ensure the clock can be easily
separated from the pendulum for adjustment and be replaced in
exactly the correct position. The suspension can be adjusted
laterally. There are 5/32 x 56 t.p.i. capstan headed indexed
adjustment screws.
Materials Used by Harrison.
Lignum vitae, a resinous wood derived from an
evergreen tree (Guaiacum officinale)indigenous to the West
Indies is used for rollers and bearings. When freshly cut it has
a reddish brown colour with pale yellow sapwood. As it oxidises
the colour turns to a deep green, often with black details. The
grain is highly interlocked, making it difficult to work but it
machines well and takes a high polish. K.G.McLaren and D.
Tabor of the Research Laboratory for the Physics and
Chemistry of Solids, Cavendish Laboratories, Cambridge
studied the frictional properties of lignum vitae. (1961,
Br.J.Appl.Phys.12,118-120). In the dry state the friction of
this material is comparable to P.T.F.E. and experiments show
that this is due to the lubricating activity of the wood waxes
expressed from the wood during sliding. The rolling friction
due to hysteresis losses is small. The sliding friction can be
explained satisfactorily in terms of an adhesion mechanism.
Guaiacum resin is a highly polarised di-azo alkaloid which was
prescribed as a cure for syphilis in the
sixteenth and seventeenth centuries
and used as a sensitive test for blood
in the twentieth.
In 1548 Anton Fugger set up
an endowment for a hospital called the
Holzhaus in the Fuggerei, Augsburg.
The hospital was specifically built for
treating syphilis with guaiacum. The
Fuggerei, founded in 1516 by Jakob
Fugger, is the earliest example of
social housing in Europe.
(W.A.Mozarts great-grandfather was
Anton Fugger by Holbein
56
a resident).
Snakewood is a dense, extremely close grained, resilient
hardwood derived from a small relatively rare tree (Piratinera
guianensis) found in the forests of Central and South America.
The name is inspired by its characteristic snakeskin-like
figuring. Initially deep red in colour, its mottling changes
colour on being exposed to air, which makes the wood turn
reddish brown eventually. It was used extensively during the
eighteenth century for the manufacture of bows for Baroque
string instruments. Harrison played the bass viol (viola da
gamba) and would almost certainly use a snakewood bow. Its
close grain and resilience make it the most suitable wood with
which to construct the pallets. Barak Norman (1678-1740)
master viol maker worked from the Bass Viol in St. Pauls Ally.
Brass is used extensively. Abraham Darby (1678-1717)
developed a scientific understanding of the brass making
process. He went to Holland to recruit skilled Catholic brass
workers who knew many of the industrial secrets of brass
production. He worked at the Bristol St. Pauls works before
moving to Shropshire. Brass was prepared using the calamine
cementation process. (Job, ch. 28, v.2 Iron is taken out of the earth
and brass is molten out of the stone.) Before the 18th century, zinc
metal could not be made since it melts at 420C and boils at
about 950C, below the temperature needed to reduce zinc
oxide with charcoal. In the absence of native zinc it was
necessary to make brass by mixing ground smithsonite ore
(calamine) with metallic copper and heating the mixture in a
crucible. The heat was sufficient to reduce the ore to metallic
state but not melt the copper. The vapour from the zinc
permeated the copper to form brass, which could then be melted
to give a uniform alloy.
Calamine was mined in Somerset, Nottinghamshire and
Flintshire (Rhosesmor). It is found associated with lead ores
which were separated by washing and hand picking. After
calcining the ore it was powdered and mixed with coal dust and
small pieces of copper in pots and heated in a furnace. When the
brass had formed it was poured between two stone moulds to
produce a plate weighing about 30 - 40 kgms. This method of
making brass from calamine was carried on until as late as 1858
and made in this way it had an appreciable lead content, which
made it easy to machine but also prone to degradation with age.
It is impossible to obtain a zinc content greater than 31-32% by
this method of manufacture.
In 1738 the Champions of Bristol developed a method
57
of extracting zinc metal from the calamine ore. By 1758 they
had shown how this metal could then be alloyed directly with
copper to make a better quality brass. They found that 35%
zinc content made the hardest brass
Ingots were cut into seven or eight strips and stretched
on the rolling mill to a desired thickness. They were then drawn
through iron draw plates to produce brass wire of different
thicknesses. This was used to make pins. Around 1700 the
casting process for brass developed.
The London Brass Wire was thicker than the Sheffield, and
the Sheffield than the Holland; but it seems from whats above that
different thicknesses do not cause different Extension but the degrees of
hardness, and different Mixture and Nature of Metals do. For the
Sheffield Brass was the hardest of the 3, yet the Holland Brass alterd
more, after Nealing: Still not so much as the Sheffield tho hard.
John Harrison, Manuscript, 1730
In 1711 The Bristol Brass & Wire Company, with other
brass producers, petitioned Parliament against the importation
of foreign brass and brass ware from Holland. They claimed that
they were providers of employment of the poor in England and
users of British copper and calamine to produce Brass. They
protested against laws which protected imports from Holland
- laws which came from a time when there had been no English
brass industry. This petition was denied by Parliament which
had set up a Committee of Enquiry to investigate the claims of
the English producers and the merchants who imported and
re-exported foreign brass.
Steel is used for his pendulum, crutch rod and springs
in the clock. It is not used on any sliding surface. Quality steel
was produced in England in the eighteenth century by smelting
Swedish ore. The Swedish ore contained very low levels of
common impurities, leading to higher quality irons and steels
from otherwise identical techniques applied to other ores.
Swedish iron was packed into stone boxes in layers with charcoal
and heated in a furnace for at least seven days. The surface of
the bars of iron became uneven and blistered. The bars of this
blister steel varied in quality, so a number were bound together,
heated and hammer forged to mix them together and even out
the carbon content. The resulting product was called shear steel.
A clockmaker, Benjamin Huntsman, developed an
improved method in 1738 at Handsworth, near Sheffield. He
heated crucibles in a coke fired furnace at 1600C until they
were at white heat and then charged them with blister steel and
58
a glass flux to produce a slag of the impurities. After about
three hours in the furnace the slag was skimmed off and the
steel poured into ingots. The steel produced by this method
contained a higher carbon content than our present day mild
steel. Harrison would have used this steel to make springs and
tools to cut the wheels.
The roller bearings are made of a high tin content
bronze, most probably 60 % tin, 40 % copper). Isaac Babbitt
patented similar antifriction metals, characterised by their
resistance to gall in 1839, in the United States. High tin
content alloys with copper consist of a relatively soft solid
matrix of tin in which hard copper-tin needles are distributed.
As the bearing wears the harder crystals are exposed by the
matrix eroding to provide microscopic high spots which are the
actual bearing surfaces. Harrison would know about the
crystalline structure of bell metal (80% Cu, 20% Sn)from his
experience of bell hanging and tuning. Bells were tuned in the
seventeenth and eighteenth centuries by chipping away the
periphery of the sound-bow with a cold chisel, thus revealing the
crystalline structure of the metal. Increasing the tin content in
bell metal alters its crystalline structure and lowers the
temperature of its melting point. Speculum metal (50% Sn,
50% Cu) was in common use for telescope mirrors and other
scientific instruments in the seventeenth and eighteenth
Wire and Plate Gauges
The early attempts to standardize gauges into a single standard can be
traced back to Peter Stubs and Charles Holtzapffel. In 1847,
Holtzapffel attempted to standardize the gauge used by Stubs into a
decimal gauge using decimal sub-divisions of the inch. He was
followed in this attempt by Joseph Whitworth (in 1857) and Latimer
Clark (in 1867). However, these proposals failed to get the industry
to adopt a single standard wire gauge.
In the 1870s, the Society of Telegraph Engineers (STE) revived
Clarks proposals with a view of using standard sizes to purchase
telegraph wire. They based their standard on the current Lancashire
and Birmingham gauges. The Standard Wire Gauge (SWG)was
introduced in 1883 by the Board of Trade and became the only legal
wire gauge in Britain. It was based predominantly on the base metal
gauges in common use in Lancashire, Sheffield and Birmingham. The
brass plate thicknesses in the RAS regulator are similar to SWG sizes
and suggest that Harrison used brass plate from Sheffield which he
considered harder than other sources.
59
centuries because of the ease by which it be polished and
figured. Harrison describes tempering spiral springs in molten
90% lead, 10 % tin in his manuscript Explanation of my Watch.
Restorations of RAS Regulator
Cottingham Restoration 1908
E.T. Cottingham, F.R.A.S., restored the regulator in
1908, forty years after it was given to the Royal Astronomical
Society. He described the clock in the Monthly Notices of the
Society in 1909. This paper was reprinted in the Horological
Journal in 1910. The photographs illustrating this paper are of
an extremely high standard; they have a large depth of focus and
a flat field. They were almost certainly taken with a plate camera
with tilting back and a Ross Homocentric lens (introduced in
1902). His photograph of the pallet frame, pallets and crutch
show these parts incorrectly assembled (see p. 145).
Cottinghams 1909 description of the mechanism is copied:
The escapement has a large but very light escape-wheel of 120
teeth, which turns once in four minutes, one tooth of which escapes with
every two beats of the seconds pendulum.
Eight fine pins placed round the centre of the wheel at about
60
0.25 of an inch radius engage in turn with a delicate brass-forked lever,
which releases the other wheels and allows the minute-hand to move
forward at half-minute intervals.
The axis of the crutch (which imparts motion to the
pendulum) usually carries the pallets, and the two have a common
centre of motion. In this clock the crutch carries a rectangular frame,
marked A in the figure below. In this frame the two wood pallets and
two small brass gravity arms work on a common axis at B, 0.9 of an
inch from the centre A. The pallet axis B therefore has a circular
motion equal to that of the pendulum. The crutch D has just completed
its swing to the right and the pallet E engages the tooth F silently and
without friction, causing the wheel to recoil slightly. As soon as the
wheel recoils, the pallet G, being slightly weighted at its opposite end 1,
rises clear of the tooth H until it meets the heavier brass arm I, which
is momentarily lifted from its stop K, by the slight impact. The tooth F
now forces the pendulum to the left until the pallet G meets the tooth J,
when it acts as a toggle joint in giving a slight recoil to the wheel again,
and releases the pallet E until checked by contact with the heavier arm
C. The tooth J, engaging with pallet G, forces the pendulum to the right
again, to complete the escaping cycle of one tooth. The pallet E enters the
wheel because it is in contact with C, but in the drawing they are
separated as they will be when the crutch D has nearly completed its
swing to the left. The crutch axis works on knife-edges resting in agate
Vs.
In other clocks the release of the escape-wheel teeth from one
pallet to the other causes the tick; but in this it will be seen there is no
such impact, as the pallets meet the teeth with a very slow motion, and
the beat of the clock is practically inaudible, so that, from an
astronomers point of view, it needs a seconds contact and a sounder.
There is no rubbing friction on the pallets, and that disturbing
factor, oil, is not required. Although there is a very slight recoil of the
escape-wheel, it would be a great injustice to Harrison to compare it
with that commonly known as the recoil escapement, for the working
conditions are so different; but it is much more delicate and needs great
care in adjustment, which is doubtless the reason why it died with the
inventor.
The function of the remontoire is to maintain a practically
uniform pressure at the escape-wheel teeth. The escape-wheel pinion
works through a hollow spindle or pinion, and the two are connected by
a watch-spring, which means they must make the same number of
revolutions per hour. Every half-minute the escape-wheel unlocks the
61
remontoire, and allows the wheel on the hollow pinion to move forward
half a minute, and in doing so, winds up the watch-spring from the
outer end.
The inner end of the watch spring unwinds an equal amount
by turning the escape-wheel one half-minute when another rewinding
takes place. Thus the action of the remontoire cuts out the variable
gearing or train friction from the escapement. It must not be lost sight
of that the latter has to release the remontoire, which, if not well
designed, creates much friction.
All the arbors in the clock are hard gun-metal, and pass
through enlarged holes in the plates, to rest on pairs of light
friction-wheels with lignum-vitae bushes, working on fine brass pins on
the outside of the plates. The brass pins on which the bushes turn show
no signs of corrosion, and were quite bright when removed, although the
clock has been standing for many years under the unfavourable
conditions of a London atmosphere.
The main wheel works in beautifully finished roller-bearings,
the rollers of which are made from a very hard non-rusting alloy, and
are kept from touching each other by working on the brass pins in a
revolving brass carriage. this represents in design the latest practice in
engineering of to-day, although made a hundred and seventy years ago.
This is one of the striking features of the clock, as it is generally believed
that the roller bearing, with the rollers pivoted in a revolving carriage,
was invented much later than 1740, and until further evidence is
available, this must be taken as another of Harrison s inventions.
The clock goes eight days, the double minute hand revolves
once in two hours, and as the escape wheel revolves only once in four
minutes, a four-armed seconds hand is used, one arm of which reads
from 1 to 30 seconds, and then passes from view, when the following
arm comes view on the other side of the dial, and reads from 30 to 60
seconds.
The wood rod pendulum has a 5 lens shaped brass bob, and
swings between adjustable cycloidal cheeks through an arc of 12.5, and
was started going at the beginning of November last, after a rest of forty
years.
Rupert Gould 1927 Restoration and Additions
Rupert Gould made a short note in the Horological
Journal (1932 May, vol. 74, p151) about his work, using a
62
1909 Cottingham photograph as illustration. His notebooks
and correspondence give details of this work and the changes
that he made to the clock. They are listed:
i. He renewed an escapement pallet. Altered pallet arms.
Replaced the escapement wheel.
ii. Made new parts for the roller bearings. Electro-plated all
the silvered parts.
iii. Fitted new enlarged remontoire fly according to the
Bradley drawing of H3.
iv. Altered the shape of the remontoire operating cam
wheel.
v. Added additional parts to the back plate and drive drum.
vi. Replaced crutch to adjustable Continental regulator
style made in heavy brass. Made pallet frame angle
adjustable.
vii. Replaced wooden pendulum with a copy of the gridiron
pendulum in the Harrison 1728 Guildhall regulator.
viii. Replaced the driving weight with a heavier weight.
ix. Replaced the driving weight pulley.
x. Milled elongated fixing holes in the base girders.
xi. Fitted a stopping mechanism to prevent complete
unwinding of the remontoire.
xii. Shortened stopper arm.
xiii. Fitted packing pieces between suspension plates to
accommodate Ager Boughs wide pendulum suspension
spring.
xiv. Replaced maintaining power spring.
His notebook gives an inventory of the original clock
parts that he disassembled and cleaned and accurate
measurements of the clock to enable a wooden case to be made
for use at the Royal Astronomical Societys premises.
Unfinished Features
The remontoire is very similar to that fitted on H3. The
63
release mechanism has the same proportions in both clocks.
Instead of the helical springs found in H3 the RAS regulator is
fitted with a spiral spring. This is most probably a either
temporary fitment whilst the pendulum and escapement were
being correctly fitted or an experimental fitment used for the
design of H4. Space is allowed for fitting helical springs and
the associated cams as in H3. There is a drawing of the
geometrical construction of the cams for the helical springs for
a remontoire having the same size wheel as this (208 mm) in the
Guildhall manuscripts.
Other Changed Features
The end bearing on the
remontoire fly arbor is a blued steel
screw. Beside it are two small holes
suggesting that it was originally
fitted with a lignum vitae end
bearing similar to that fitted on the
rear roller bearing. Other end
bearings have been replaced in a
similar fashion; this work was most
probably done either when the clock
was in Frodshams workshop or by
Cottingham.
Gould wrote a report in 1927 in the Monthly Notices
of the Royal Astronomical Society describing the work that he
had done and the changes that he had made to their regulator.
The Societys Harrison Clock.
These notes upon the recent overhaul of the Societys Harrison
clock should be regarded as supplementary to the account given by Mr.
E. T. Cottingham in the Monthly Notices for 1909 November, which
gives an excellent description of its mechanism. It may be noted that his
account contains the first correct drawing of the grass-hopper
escapement ever published.
The clock was entrusted to me by the Society in 1927 April,
I having undertaken to give such time as I could spare to cleaning and
repairing it. The Society, on its side, undertook to defray my out-of-
pocket expenses, the cost of such repair-work as I could not accomplish
Remontoire End Bearing
64
with my limited resources, and the expense of making a new case and
pendulum.
When it came into my hands, the clock had not been going
since 1910. It may be noted that prior to Mr. Cottingliams work on
it in 1909 it had been stopped since about 1870. It appears to have
been going from at least 1856 until that date, as I found scratched on
the plates the name of one John Beresford, with the dates 1856 and
1860, and the address 84 Strand, indicating that Beresford was a
workman in the employment of Charles Frodsham.
Having temporarily repaired the wooden pallets, which were
broken, and fitted a cardboard fly to the remontoire train in place of the
small and heavy copper fly which I found in it, I got the clock going and
made a number of experiments to determine the most suitable driving
weight, adjustment of the remontoire spring, etc. The clock was then
taken down for cleaning. During this process, which occupied over a
year, it was re-assembled from time to time to test various adjustments
and repairs. Altogether, I have dismantled and re-assembled it five
times. It may be noted that on the first occasion it took me nearly a
week to get it togetherand on the last occasion three hours in all. It
contains some 540 parts.
The movement was found to be very dirty, dust and grime
having made their way into the case in considerable quantities. As
regards mechanical defects, in addition to the pallets and fly referred to
above, the maintaining gear was found to have jammed, and the crutch
to have cracked almost completely through. The remontoire was out of
adjustment, and the fork of its detent cracked. There was a film of
oxidation and soot over almost every part of the movement, and the
cavities behind the various wooden rollers of the lantern pinions were
partly filled with dust and dirt.
After some preliminary cleaning, the main plates of the
movement were polished and lacquered and the dials re-silvered, this
work being carried out by the Gold and Silver Plating Co., Bayswater.
The broken pallets were replaced by new ones, which I cut out of lignum
vitae after determining their length and angles by an enlarged model of
the escapement. The pallets found in the clock were of mahogany, but I
do not consider that they are the original pallets. Harrison habitually
used (as in other parts of this clock) lignum vitae, in conjunction with
hard polished brass, to form a bearing or pivot which needed no oil; and
I see no reason why he should have departed from this practice in his
pallets. It may be noted that these lignum vitae-brass pairs of surfaces
are to be found all through this clock, and that in consequence no oil
whatever is required for any portion of its mechanism.
65
It was found necessary to make an addition to the clock. If (as
was ultimately the cause of its stopping in 1910) it is allowed to run
down, it cannot, as originally designed, be restarted by merely
rewinding it. The clock is not driven by its driving-weight, but by a
small remontoire spring which that weight rewinds every half-minute.
When the weight reaches the floor of the case, it has, of course, no longer
any power to rewind the remontoire spring; but the latter continues to
keep the escape-wheel turning until it has itself run down. When,
therefore, the clock ultimately stops, the remontoire spring has run down
to such an extent that it will no longer keep the pendulum swinging.
Merely rewinding the weight does not rectify matters - to do this, the
detent of the remontoire has to be let off repeatedly when the escape-
wheel is in the correct position for doing this. Such an adjustment
requires expert knowledge of the clock mechanism.
As it was fairly certain that, if kept continuously going, the
clocks weekly winding would in all human probability be occasionally
overlooked, I designed a mechanism (based partly on one used by
Harrison himself, and partly on one used by Ferdinand Berthoud in his
second marine timekeeper) which should stop the escape-wheel shortly
before the weight touched the bottom of the case. It is operated by the last
turn of gut as this unwinds from the barrel, and jumps" a small
detent lightly into the path of the escape-wheel teeth, thus stopping the
clock with the remontoire spring at its correct tension. As soon as the
clock is rewound, the detent is automatically withdrawn, and the clock
is ready to be restarted. The mechanism of this stopping-gear is
practically concealed within the movement of the clock, and it is difficult
at first sight to see that any addition has been made.
Apart from the fitting of this stopping-gear and the new
pallets, the mechanism of the clock has undergone practically no
alteration. The profiles of the cam which operates the remontoire detent
have been slightly smoothed to give an easier lift to the detent, the detent-
fork given a screw adjustment, and the fly replaced by one copied from
the fly used in Harrisons No. 3 marine timekeeper. It may be noted
that there is a very close resemblance between the mechanism of this
time-keeper and that of the RAS. clock - in fact, they may be regarded
as two similar movements, one controlled by a pair of balances and the
other by a pendulum. In one respect, however, No. 3 is mechanically
superior. Its remontoire gives an absolutely uniform torque at the
escape-wheel, while the torque afforded by that of the clock varies, every
half minute, in the proportion of n to n - I, where 1/8 n is the number
of complete turns constituting the initial tension of the remontoire
spring. Personally I am of opinion that while a remontoire was an
absolute necessity in No. 3, the RAS clock would perform practically as
66
well with its remontoire out of action, especially since it has the most
perfect train I have ever seen in a timekeeper.
The wooden pendulum found in the clock has been replaced by
a gridiron pendulum of brass and steel rods, the gridiron being copied
from that of an actual Harrison clock. It is obvious that Harrison, who
invented the gridiron pendulum, would not have fitted a wooden
pendulum to a clock of his own making so perfect in all other respects
such a proceeding would be analogous to completing a Rolls-Royce
chassis by a body made of packing-cases. It is quite safe to assume, as
I have assumed, that the wooden pendulum is a replacement, and that
the original pendulum was a gridiron.
The grandfather case of the clock has been discarded. I am
exceedingly doubtful whether it can be the original case. In the first
place, its sides have been cut away in two places almost to paper-
thickness to allow for the swing of the bob - a proceeding which does not
exactly suggest that the case was originally designed to house the
movement. And, secondly, there are various features of the movement
and of the plate carrying the pendulum, which strongly suggest that they
were originally mounted on a wall.
In consequence, I designed a new case for the movement on the
lines of a museum show-case with glass front, sides, and top, the
movement being mounted on brass brackets screwed to an iron back-
plate firmly secured to the wall of the Grove-Hills Library. This case
admits of the movement being examined in detail from almost every
point of view, and should be practically dust-tight. Provision is made
for winding the clock (and, if necessary, setting the hands and restarting
the pendulum) without removing the glass cover. I am hopeful that the
clock will continue to go for many years in this case without any
further attention except its weekly winding and occasional resetting of
the day-wheel, whose operating mechanism does not provide for the short
months.
I have not been able to determine the exact date at which
Harrison made the clock. The only indication of a date (if it can be so
regarded) which the clock affords is an inscription scratched on the
maintaining spring - which, being rusty, has been replaced by a
duplicate. This inscription reads May 827. I can make nothing of
it. From the clocks general lay-out, and by comparison of its
mechanism with that of Harrisons No. 2, 3, and 4 marine
timekeepers, I should be inclined to say that it was not made before
1740 or later than 1750, and that its most likely period of
construction is 1745 - 1748.
Apart from its unique mechanism, the clock is of very
67
considerable historical interest. There can be no reasonable doubt that
it is the regulator which Harrison made for his own use. Both in his
published writings and in an unpublished MS., of which I possess a
copy, he constantly extols the merits of my clock . . ., a clock such as
mine, etc., in comparison with the common regulators of his day, for
the purpose of adjusting his marine timekeepers. The RAS clock is the
clock in question, and in consequence it must have played in its makers
lifetime a very considerable part in solving the problem of finding
longitude at sea. The Society is to be congratulated upon possessing, in
practically mint-state, so valuable a relic of the man who was
Englands greatest horologist and one of the worlds benefactors.
In conclusion, I should like to pay a tribute to Mr. J. H.
Agar Baugh, who undertook the making of the case and pendulum; to
Mr. W. Buck, who was responsible for the re-conditioning of the plates
and dials; and to Mr. R. J. Hopgood, who made the automatic
stopping-gear and various small pieces for me. Their ready co-
operation, and that of the Societys officials, has contributed very greatly
to the success of the work.
R.T. Gould, MNRAS 1929, Vol.89, pp. 398-401
68
John Harrisons Measurement Standards
Precise and agreed standards of measuring length and quantity are vital to the
development of an ordered society. Anthropomorphic mensuration is generally believed to be
the basis of most historical measures. In the middle of the tenth century, Edgar, the Saxon
king made The Yardstick at Winchester the standard of length within his realm.
Subsequently, in 1588 Elizabeth I established a standard yard which remained the legal
British yard for the next 236 years. The current Imperial Standard yard, commissioned by the
Royal Society in 1742 and introduced by Act of Parliament in 1824 under George IV, only
differs from the remaining Elizabethan and earlier standards by a few hundredths of an inch.
Gunters chain was defined as 22 standard yards and divisions of the standard yard defined
the foot and inch.
The Right Reverend John Wilkins (1614-1672), Bishop of Chester was the
Chairman at the founding meeting of the Royal Society and its first secretary. Although
originally a Royalist he married Oliver Cromwells sister, Robina. He recognised the ideal of
a universal measurement standard in 1688 in An Essay Towards a Real Character and a Philosophical
Language. He discusses basing the standard on the circumference of the earth, but thinks that
this is too difficult and inconvenient to be practical. He then describes a method which he
says was suggested by Sir Christopher Wren; the length standard can be based on a time
standard by using the length of a pendulum with a defined period.
Wilkins directed that a pendulum should be set up with the heaviest, densest
possible spherical bob at the end of the lightest, most flexible cord and that the length of the
cord be adjusted until the period of the pendulum was as close to one second as possible.
....which being done, there are given these two Lengths, viz. of the string, and of the Radius of the Ball, to which a
third proportional must be found out; which must be as the length of the String from the Point of Suspension to the
centre of the Ball is to Radius of the Ball, so must the said Radius be to this third which being so found, let two
fifths of this Proportional be set off from the Centre downwards, and that will give the Measure desired. Using
this definition Wilkins found his standard unit of measure to be 39 inches and a quarter.
From Wilkins italicised description and
2
, = =
d r r
x
r x d
2
2 2
,
5 5
= + = +
r
l d x l d
d
The moment of inertia of a sphere about its centre derived by integration is 2mr
2
/5. When
suspended by a cord the moment of inertia of Wilkins pendulum becomes as he
2
2
5
+



r
m d
d
suggests and uses to derive the length of an equivalent simple pendulum.
In August of 1793 the Republican Government of France, on advice from the
Academy of Science decreed that their standard unit of length was to be 10
-7
of the earth's
quadrant passing through Paris and that the unit be called the metre. They chose the meridian
definition in preference to the pendulum because gravity varies with geographical position.
Harrison uses inch mensuration in the parts of the RAS Regulator that are derived
from his early clocks i.e. motion and date work and for the dimensions of work requiring
tools, i.e. holes. He works by proportion from a datum length (one link of Gunters chain)
for the main plates and dial, and by proportion of the standard 1000 mm. pendulum
69
Chapter 5
Logical Design
Development of Form
Geometry has two great treasures; one is the Theorem of
Pythagoras; the other, the division of a line into extreme and mean
ratio. The first we may compare to a measure of gold; the second we
may name a precious jewel.
Johannes Kepler (1571-1630)
John Harrison began his working life as a carpenter
and surveyor. When he moved to London in 1737 he joined
the subscribers list for publication of an English translation
of Palladios Four Books of Architecture (published Isaac Ware
1738, reprinted Dover Publications 1965). This book,
containing accurate reproductions of the original Italian
illustrations was promoted by Richard Boyle, third Earl
Burlington. Thomas Heath, secretary of the Royal Society,
William Jones, mathematician, Samuel Johnson, lexigrapher,
Thomas Gray, poet and William Hogarth, engraver, were also
subscribers. The subscribers to this book are drawn mainly
from four groups, members of the Whig government,
associates of the East India Company, governors of Corams
Foundling Hospital and artisan employees of Lord
Burlington. The book had a great deal of influence on
eighteenth century thought.
Proportion was considered a major factor in the
perception of beauty and was linked to geometrical and
organic form. It had been illustrated by Leonardo da Vincis
interpretation of Vitruviuss description of symmetry in his
Ten books on Architecture (First Century, B.C.). Bramante,
Michelangelo, Palladio, Vignola and earlier architects were
careful students of the work of Vitruvius. Eighteenth
century mathematicians were still puzzled by the occurrence
of irrational numbers but proportions based on them were
extensively used in architecture and the design of musical
instruments (viz. the shape of the soundboards of
harpsichords and string instruments).
71
The design of a temple depends on symmetry, the principles of
which must be most carefully observed by the architect. They are due to
proportion, in Greek, "<"8@(4" . Proportion is a correspondence
among the measures of the members of an entire work, and of the whole
to a certain part selected as standard. From this result the principles of
symmetry. Without symmetry and proportion there can be no principles
in the design of any temple; that is, if there is no precise relation between
its members, as in the case of those of a well shaped man. Therefore,
since nature has designed the human body so that its members are duly
proportioned to the frame as a whole, it appears that the ancients had
good reason for their rule, that in perfect buildings the different members
must be in exact symmetrical relations to the whole general scheme.
Hence, while transmitting to us the proper arrangements for buildings of
all kinds, they were particularly careful to do so in the case of temples of
the gods, buildings in which merits and faults usually last forever.
Vitruvius, Ten Books of Architecture
72
Harrisons portrait by King was painted after he had
received a relatively large part of the Longitude Prize money
(7000) and shows him as a well dressed gentleman.
During the years 1737 until 1762, whilst living in London,
he had received 1250 from the Longitude Commissioners
for his personal expenses in the development of his
timepieces which is equivalent to 50 per year. The East
India Company also gave him 100. His style of living in
London must have cost him more than 90 per annum.
Whilst living in London he is known to have definitely
worked as a coach-maker and would also have continued his
work as surveyor. The two executors of his will were
merchant venturers; he could have earned the money to
support his living expenses and the making of his watch by
Jeffreys with similar activities. Insufficient money was
forwarded by the Board of Longitude to enable the
development of H4 as illustrated by the published cost of
making a copy.
Jrme Lelande (1732-1807), French astronomer
and writer, who corrected Halleys tables of planetary
motion, visited London with Ferdinand Berthoud (1727-
1807), French clockmaker, in 1763. He wrote a diary
describing his visit.
Mr Harrison was a carpenter making carriages and carts,
9 May 1763.
We went to see Mr Harrison and his three longitude clocks.
The seconds clock which he made in 1726 is still used as the regulator;
the Ambassador of Spain wanted to give him 2,000 pices (3,600)
for it. Mr Berthoud found these pieces very beautiful, very clever, very
well executed; and though the regularity of the watch was quite difficult
for him to believe, he was even more impatient to see it after seeing the
three clocks.
Saturday 14th May
I had a long discussion with Lord Morton and I left
convinced that we will not see the watch. I took Mr Condamine,
Camus and Berthoud. We also went to Lord Macclesfield and from
there to Mr White at the Foundling Hospital. He told us that at the
next Parliament the commissioners would be reprimanded for vexing
Harrison and that it was necessary to have six votes to settle it, but he
did not promise us, as before that we would see the watch.
Jrme Lalande, Diary of a Trip to England, 1763
73
William Hogarth cites Harrisons clocks as an
example of good design in his book Analysis of Beauty.
The proportions of the plates of the second sea clock are
the first example of Harrisons use of the golden rectangle.
The RAS regulator continued this theme by using two other
irrational numbers, B and the square root of two.
There are natural Causes of Beauty. Beauty is a Harmony of
Objects, begetting Pleasure by the Eye. There are two Causes of Beauty,
natural and customary. Natural is from Geometry, consisting in
Uniformity (that is Equality) and Proportion. Customary Beauty is
begotten by the Use of our Senses to those Objects which are usually
pleasing to us for other Causes, as Familiarity or particular
Inclination breeds a Love to things not in themselves lovely. Here lies
the great Occasion of Errors; here is tried the Architects Judgment: but
always the true test is natural or geometrical Beauty.
C.Wren, Parentalia: or, Memoirs of the Family of the Wrens 1750
Whilst designing, drawing and setting out the front
and back plates and dials of his regulator he would have had
available straight edge, set square, dividers, Gunters scale,
and either proportional dividers or a sector. The
proportions of the clock are based on ratios of the irrational
numbers, N(the Golden Ratio), and B. At the time 2
when there were many mean tone temperaments and the
equally tempered musical scale was beginning to be
introduced throughout Europe (cf. J.S.Bach, Das
Wohltemperierte Klavier, 1722). Harrison devised a scale based
on major thirds having the proportion of B , using
logarithms accurate to five decimal places. An analysis of
Harrisons scale compared to that devised by J.S.Bach is
given in Appendix B.
Where he used B in the layout of the plates he could
have either calculated the dimensions or drawn them
geometrically. William Jones, another of the subscribers to
Wares book is credited with the first use of the symbol B
for the ratio of the circumference of a circle to its radius.
Leonardo da Vincis 1492 drawing of Vitruvian man
illustrates his keen interest in proportion. Da Vinci
believed that the workings of the human body are an analogy
74
of the universe and he symbolised the material existence by
the square and the spiritual existence by the circle.
Albrecht Drers in his four books Vier Bcher von
Menschlicher Proportion (1528) wrote about the proportions of
the human body. In the third book he attempted to define
ideal beauty. He believed truth was hidden in nature and
that there were rules which ordered beauty which
exemplified the blend of art and science during the
Renaissance even though he found it difficult to define the
criteria for such a code. However unlike his predecessors,
da Vinci and Alberti he tried to understand the abstract
notion of beauty, and how an artist could create beautiful
images.
Drer considered that an artist builds on a wealth of
visual experiences to inspire ability to enable him to image
beautiful things. Harrisons wooden clocks appear as well
proportioned cubes. The Brocklesby Hall clock is a delight
to the eye as well as being well proportioned to withstand
the forces imposed upon it. The proportions of the plates
of the second sea clock are the first example of the influence
of the golden rectangle, most probably as a result of
discussion with Hogarth. The many visitors to the
Harrison gallery at Greenwich appreciate the beauty of his
large sea clocks even though he could not make them fulfil
the conditions and use they were designed for.
Harrison began setting out his final regulator by
drawing a circle with a radius of one link of Gunters
surveying chain (7.92"). Gunter introduced the first
decimal mensuration into England with his surveying chain.
He divided the furlong into ten chains and then subdivided
each chain into one hundred links. The chain of 22 yards
Rev. Edmund Gunter(1581-1626), was educated at Westminster School, and
Christ Church, Oxford. After graduating bachelor and master of arts and in
1615 proceeded to the degree of bachelor in divinity. However he was more
interested in mathematics and was appointed Professor of Astronomy at
Gresham College, London in 1619. In 1620 he published his Canon
Triangulorum, a table of logarithmic sines and tangents for every degree and
minute of the quadrant. He introduced the terms co-sine and co-tangent for
sine and tangent of the complementary angle of a triangle.
75
or 20.12 metres then provided a usable base line length
when surveying land. Later it was notably used as the length
of a cricket pitch and the distance between telegraph poles.
The measure of area of land, the acre, was then defined by
statute as ten square chains, thereby continuing the
introduction of the decimal system.
Harrison dropped perpendicular tangents from this
circle to define the maximum width of the base placing the
centre of the circle 2D/B above the baseline. The width of
the plates is the same as the height of the centre. A square
is drawn within the circle which defines the square part of
the front dial thus making use of . A golden rectangle is 2
drawn above the centre point to establish the height of the
plates. The long length of this golden rectangle is marked
below the centre to from the lower line of the cutout. We
therefore have a large golden rectangle occupying the
majority of the front plate. The spacing of the plates is the
width divided by . The height of the top of the dial
2

square is virtually the diameter of the origin circle.

Geometrical Conception
Many of Harrisons marking out lines remain having
not been polished out by subsequent restorers. These lines
together with the dimensions enable the main plates to be
marked using dividers and straight edge.
76
Layout Sequence
266.70 mm
0 10 30 50 70 90 10
405.13 mm
257.91 mm
286.47 mm
286.47 mm
G
D
F J
C
K
E
B
I
H
A
R 7.98"
a b
c d
L P Q M
R202.57
466.34

2
3
8
.
1
3
1. Draw a circle of one link radius centered on C.
(201.17 mm., 7.92 ins., 100 links = 1 chain, 22
yds.)
Harrison drew a circle of 202.57 mm., 7.975 inches
radius; an error of 1.40 mm. or 55 thou.
2. Draw a square, abcd, within this circle to produce the
dimensions of the square part of the dial. The
dimension of the sides of this square have the ratio of
%2 to the diameter of the circle giving 284.50 mm.
(Measured length of sides of square Gould = 285.50
mm. - an error of +0.96 mm., 39 thou).
3. Drop perpendicular lines from the centre and sides of
this circle to A, L, and M.
4. Width of base, LM, equal to the diameter of the circle,
(405.13 mm.) as measured by Gould.
5. Centre height of circle above base (C) equals twice the
diameter of the circle divided by B. (This is the
limiting point for x as y = 0 in the quadratrix of
Hippias.)
( ht. of centre = 2 D / B ) (257.91 mm.)
6. Width of plates equal to centre height above base of
primary circle. (257.91 mm.) (PQ=CA)
7. Lines from centre of circle to outer points of base, CL
and CM, mark position of the supporters on the side of
the plates.
8. Mark position B, half the width of the plate below the
centre (128.96 mm.). Draw horizontal line through B.
The centre of the great wheel arbor is marked 4.46 mm.
(3/16") below B. ( 125.48 mm. 4.94 inches above
baseline)
9. Mark position of escapement wheel arbor (D), half the
width of plate (128.96 mm.) above the centre C. Draw
horizontal line through D.
10. Find the mid-point between points C and D on the
centre line and draw a horizontal line. Using the point
F on the intersection of this line and the centre line as
the centre point draw an arc from the edge of the plate
at the horizontal line passing through D to the centre
line at G. Draw a horizontal line through the point G
to determine the height of the plates 466.57 mm. (
Goulds measurement 466.34 mm., an error of +0.23
78
mm.). This produces two adjacent Golden Rectangles
above the centre line
11 Find the mid-point E between points B and C and draw
a horizontal line. Draw an arc centred on E from the
junction of line B with the outer edge of the plate to the
mid-line below to mark point H. A horizontal line
drawn through point H marks the lower edge of the
cutout. The cutout delineates the lower margin of an
approximate large golden rectangle occupying the whole
of the plate.
12. The distance A - I equals B - H. A horizontal line
through I marks the upper margin of the cutout. The
half width of the upper edge of the cutout equals the
distance of this margin from the lower edge of the dial
plate, c-d. Divide the extended width of the plates into
three parts. Mark this distance from the midline on the
base line R. A line joining the great wheel arbor centre
to this point marks the lateral extremities of the cut-out
on the front plate.
13. Draw a line from the top centre of plate G to the right
side adjacent to centre wheel C. Draw a perpendicular
from this line to centre wheel C. The origin of this
perpendicular is the position of the third wheel arbor
(J).
14. Divide the distance between C and D to mark point F.
Draw a line from the top centre G to the left side
adjacent to centre wheel C. The remontoire fly arbor K
lies on the horizontal line through point F transecting
the line from G to the lateral point of the horizontal
line through C. The top arch of the front dial is drawn
with radius D-K.
15. The spacing of the main plates is the width of the
plates divided by N
2
. 98.51 mm. (Goulds
measurement 98.55 mm. an error of 0.04 mm. from
measured. ) (N
2
is equal to 1 + N.)
16. On the front dial draw a horizontal line through the
escapement axis D. Draw the dial arch, of radius D-K
(R=43.4 mm), centred on the escapement axis. Draw
perpendiculars of length 0.4R on each side from the
horizontal line to the semicircle . Draw lines from the
centre to these vertices to mark the breaks in the dial
79
arch. This angle is 23.58E (very close to the latitude of
the Tropic of Cancer in the eighteenth century). Draw
the inverted supporting arcs of radius 0.5R centred on
the incline line to the upper corners of the dial. If
Harrison wished to fit an inscription plate similar to
that fitted on H2, he would most probably have used
this feature in its design.
17. Harrison increased the diameter of the great wheel by
lowering its centre by 4.46 mm. (3/16"). He also
increased the width of the plates by 8.92 mm. (3/8") to
facilitate fitting the friction wheels of the third wheel.
He was then able to accommodate 78 teeth on the great
wheel.
18. The width of the plates was increased to 10 inches,
266.79 mm. This enabled him to accommodate the
outer friction wheel supporting the third wheel arbor.
19. Draw a lines from the centre wheel C to the outer
points of the base, M and L. The top of the
supporters is positioned where these lines cross the
outside of the plates.
Suggested Method of Marking Out
Harrison could therefore describe and set out this clock
without a drawing in the same way that eighteenth and
nineteenth century carpenters and shipwrights laid out their
work but more importantly he was able to use dividers and
straight edge to mark out his work, thereby limiting the number
of measured points he would have to determine.
1. Prepare suitable rectangular pieces of c inch thick
brass 18 x 16 inches. (475 mm. x 410 mm. x 3.175
mm).
2. Scribe a full width straight line 0.5 - 1 mm. away from
the bottom margin.
3. Find the middle of this line using dividers and centre
punch. A
4. Measure 7.92 inches, 201.17 mm to one side of this
80
centre point and centre punch. Transfer this distance
with dividers to the other side. L and M
5. Set the dividers to a large radius and scribe small arc in
the centre from the outer points L and M.
6. Scribe the centre line perpendicular from A through the
intersection of the arcs.
7. The width of the plates can be determined by geometry
using a sector or Gunters scale or more conveniently by
calculation and measurement. Set the dividers to
128.96 mm (5.077 inches) and scribe an arc on each
side of the central perpendicular on the baseline and on
the perpendicular. P and Q.
8. Mark a further 128.96 mm. on the centre line
perpendicular C and centre punch for the position of
the centre wheel.
9. Mark a further 128.96 on the centre line perpendicular
D and centre punch for the position of the escapement
arbor
10. Draw perpendiculars from the marks P and Q. Draw
a horizontal line through D to position where it crosses
the perpendicular from Q labelled R. Bisect the
perpendicular line between C and D to determine point
F. Set the dividers to the hypotenuse F-R and scribe
the arc across the centre perpendicular at G to
determine the height of the plates.
11. Scribe a horizontal line through G to mark the upper
margin of the plates.
81
Harrison valued his monochords as mentioned in his will. He played the viol, an instrument which is difficult to tune as the strings are
tuned like the lute in intervals of a fourth. The violin family is much easier to tune as the strings are a fifth apart. The incidence of the minor third in
a bell probably stimulated Harrisons interest in musical temperament. In the painting by Walter Shirlaw,( (1838-1909, Toning of the Bell the
tones are created by harmonics on a violin. The violinist is playing harmonic g on the Gstring; the bell fundamental is Gand it weighs approximately
12cwt.
Fabian Stedman published Tintinnalogia in 1668, describing
English change ringing and hanging bells for full circle ringing.
And if a bell youoverthrow,
Praypaysixpence before yougo.
-as at EastlingChurch, Kent
Bell ringing became very popular in the early eighteenth century
particularly in London. Bands of ringers learnt how to raise their bells in
peal, ring changes and then lower their bells. This method is still
practiced in Devon and Cornwall. Modern bells are fitted with stays and
sliders so that the bell can be stood in the upset state. The photograph
shows three bells ringing on the right and bells on the left balanced in the
upright position ready to be pulled off to ring. When raising and lowering
the bells in peal the ringers have to compensate for the weight of the bell
that they are ringing. In this way Harrison learnt practically about the
characteristics of a compound pendulum.
A bell has five major characteristic harmonics or overtones -
humnote, fundamental, tierce or minor third, quint or fifth, and nominal.
Jacob Van Eyck, a carillonneur and relative of Christian Huygens,
interested the bellfounders Francois (1609-1667) and Pieter (1619-
1680) Hemony of Zutphen (The Netherlands) in his investigations into
bell harmonics resulting in the founding of the first well tuned carillon
(Utrecht Cathedral). Christiaan Huygens described how Van Eyck
demonstrated by whistling the right tone near to a wine glass, the partials
of the glass can be made to sound. He was also able to demonstrate the
partials in the bells of Utrecht Cathedral in the same manner. These
findings were also known to the East Anglian bellfounders, particularly
John Darbie of Ipswich, even during the Anglo-Dutch Wars.
English Ringing.cdr
Chapter 6
Oscillator

Introduction
Of the Nature of a Pendulum, as primarily implying in
itself; and secondarily, as when according to any particular
Manner [good or bad] in which it may be applied to the Draught
of the Wheels of a Clock
Opening Chapter Heading, Concerning Such Mechanism, 1775
The accuracy of a regulator is primarily dependent upon
the stability of its oscillator. Its oscillating element, the
pendulum, must be capable of compensating for changes in the
environment within which it operates and be unaffected by the
ageing of its components. The oscillator has to be driven by a
power supply and control the integrator. The escapement fulfils
these two functions but in doing so should have a minimal
effect upon the frequency and stability of the pendulum.
The pendulum was first introduced in the tenth century
by Ibn Yunus al-Masri, an Egyptian (d. 1009)mathematician
83
and skilled astronomical observer. He studied and documented
the oscillatory motion of the pendulum and devised a method of
solving the spherical triangle by addition before the advent of
logarithms.
The introduction of the vertical pendulum driven by a
verge escapement enabled much more accurate timekeeping. It
replaced the foliot and was most probably introduced
simultaneously in a number of European centres. Christiaan
Huygens (1629-1695) published an account of a working
pendulum clock with drawings in his Horologium in 1658, but
his contempories accused him of plagiarism. Jan Babtist van
Helmont, Galileo Galilei, Johannes Babtista Riccioli, Johannes
Hevelius, Tycho Brahe and Dr. Seth Ward with Robert Hooke,
all employed 'pendulum timing devices' before 1657. Richard
Harris of London is said to have converted a clock to pendulum
control in 1642. Ahasuerus Fromanteel (1607-1693) had
made pivoted pendulum clocks before he made his 1662 long
pendulum clock with a Huygens crutch and Hookes spring
suspension. Jost Burgis crossbeat verge escapement clock with
vertical compound pendula of 1605 is an example of the manner
in which the Augsburg clockmakers were moving from the
horizontal foliot to the vertical pendulum. Burgis
astronomical clock in Kassel has a horizontal crossbeat
escapement; his later clock in Dresden has a vertical crossbeat
escapement. This may be the earliest example of the thought
process leading to the driven vertical pendulum.
Johann Bayer (1572-1625) published Uranometria:
omnium asterismorum continens schemata , the first star atlas and
catalogue in 1603. He developed the star naming system and
showed the stars as seen from earth. His atlas has a Cartesian
grid so that star positions can be depicted accurately. John
Flamsteeds (1646-1719) Atlas clestis was published in 1729,
four years after his British Catalogue. This atlas and catalogue,
of unprecedented accuracy, and Halleys Astronomical Tables
tabulate Right Ascension accurate to 2 seconds. A clock for
long term accurate astronomical observation would therefor have
to be accurate to 2-3 seconds in a year, which is what Harrison
claimed for his clock in his 1730 document. Harrison was most
probably introduced to Halley by mutual acquaintances in the
East India Company on account his claims of his clocks
84
accuracy; the clock was then carefully examined and rated by
Graham as noted in Dr. Stukeley journal of 1728.
Galileo first studied the pendulum in 1588 after it is
suggested watching and timing with his pulse the swinging of
the sanctuary lamp suspended from the highest point of Pisa
Cathedral. The suspension chain would have been 25 metres
long giving a time of each swing of five seconds. For an
amplitude of 1 the lamp would be moving over 800 mm. He
began serious investigations in 1602 publishing his findings in
Dialogue Concerning the Two Chief World Systems and Dialogues Con-
cerning Two New Sciences. He determined that:-
< Pendulums nearly return to their release
heights.
< All pendulums eventually come to rest, with the
lighter ones coming to rest faster.
< The period is independent of the bob weight.
< The period is dependent upon the amplitude.
[Europe believed Galileo's pendulum was
isochronal (or tautochronous); but first
Galileo, then Godfrey Wendelin in 1643,
observed it was not.]
< The square of the period varies directly with the
length.
Galileo Galilei (1564-1642), Italian (Tuscan)
physicist, mathematician, astronomer and
philosopher, taught at the Universities of Pisa
and Padua. He discovered the four largest
moons of Jupiter after improving the telescope
and made significant discoveries in kinematics
and the strength of materials. He introduced the
concept of inertia which led to Newtons first
law of motion. Because he supported Copernicuss theory of helio-centrism
and challenged the Churchs theological teaching they banned publication
of his books and forced him to live under house arrest. On 31st October,
1992 Pope John Paul II expressed regret for how Galileo had been treated
and officially conceded that the earth is not stationary.
85
Vicienco Viviani, Galileos friend and biographer wrote:
One day in 1641, while I was living with him at his villa in Arcetri,
I remember that the idea occurred to him that the pendulum could be
adapted to clocks with weights or springs, serving in place of the usual
tempo, he hoping that the very even and natural motions of the pendulum
would correct all the defects in the art of clocks. But because his being
deprived of sight prevented his making drawings and models to the desired
effect, and his son Vincenzio coming one day from Florence to Arcetri,
Galileo told him his idea and several discussions followed. Finally they
decided on the scheme shown in the accompanying drawing, to be put in
practice to learn the fact of those difficulties in machines which are
usually not foreseen in simple theorizing.
Many of the physical and mathematical principles that
Harrison used were later documented by others. An indication
of his grasp of mathematics is his description of the pendulum
chops as a involute of a cycloid in his 1730 document.
Harrison was frustrated by the conservative ideas of his
contemporary clock and watch makers. This stimulated him to
write Concerning Such Mechanism.
Now, whether my Style of Writing in this Affair, be right proper
to the Purpose or not, I thought it must be better that the contents
of this Book should be Danger of sleeping in Oblivion; yea,
86
notwithstanding what I had - as verbally communicated to the
World.
John Harrison, Concerning Such Mechanism, 1775
Galileos findings were mathematically quantified by Sir
Isaac Newton and published in his Philosophiae Naturalis Principia
Mathematica, generally known as The Principia. The motion of
the pendulum posed the problem of whether the circular arc of
its bob or the straight line of an inclined plane was the fastest
motion from a higher to a lower point or whether it should be
a cycloid. Newton notes that Sir Christopher Wren had
discovered the properties of the cycloid.
Huygens was satisfied with the extent of his research on
Saturn by 1659, and his book, Systema Saturnium, was printed and
ready for distribution by July of that year. Huygens dedicated
the book to Prince Leopold of Medici, a powerful and
influential patron of science. Leopold had been upset with
Huygens after the latter's publication of his discovery of the
pendulum clock in Horologium, in 1658. Leopold had been led
to believe that his fellow Tuscan, the highly revered Galileo
Galilei, had invented the pendulum clock years before and that
Huygens was an opportunistic plagiarist. Huygens publication
did however include the use of a crutch thereby separating the
A cycloid is a curve generated by a point of a circles
circumference rolling on a plane. The cycloid possesses
interesting physical properties. It is brachistochronous and
tautochronous. Brachistochronous, because it represents the
path completed in the shortest time between two points for a
given type of motion (such as a fall under the effect of gravity)
and tautochronous, because a body made to oscillate along a
cycloid will always take the same time to cover it, whatever the
amplitude of the oscillation. Galileo mistakenly believed
circular oscillations to be tautochronous. The brachisto-
chronous property of the cycloid was demonstrated by Jacques
Bernoulli (1654-1705) in 1697, while Christiaan Huygens
(1629-1695) proved its tautochronism in 1659.
87
oscillating pendulum from its driving force. This is a great
advance on the pendulum clocks of Galileo, Burgi, Fromanteel
and East.
From the above explanation the
principle of the whole invention (the further
details in the figure will be mentioned later) will
be perceived, noticing in the first place that if the
rod SIT did not pass through the crutch-fork R,
or was not present at all, the crutch QR would
them undoubtedly be caused to oscillate violently
by the force of the weight D, is obedient to the
motion of the pendulum, yet for a short time in
each vibration, it also assists this movement. In
this way the continual motion of the pendulum
is accomplished which, unless the rod were
connected with the mechanism, would soon fail
and come to rest; but with each swing of the pendulum the pallets
M and N receive corresponding impulses from the contact of the
teeth of the wheel L. These are indeed the details of my mechanism
which require precise explanation because the point of the whole
invention turns on them.
Christiaan Huygens, Horologium, 1658.
Unfortunately Huygens states in the beginning of
Horologium oscillatorium sive de motu pendularium published in
1673:-
Sixteen years ago we published a little book, after having
invented the pendulum clock. Truly, observing that as we have
discovered many things in that time along the way leading to the
completion of the work, it seems fitting that the individual
discoveries should be set out here in this book. Indeed, not only are
these discoveries pertinent to the perfection of the clock we have
invented, and as such they can be considered the most important
part of it, but also these observations are relevant for the
fundamentals of all such clock mechanisms, which had earlier been
in a very poor state.
This statement has led many to believe that Huygens
was the first to devise a pendulum clock.
In 1660 Hooke had described his conical pendulum
where the bob is suspended by a long flexible ribbon over a curve
defined by the intersection of a circular conical surface by a
plane (a hyperbola, ellipse or parabola). He managed to make
88
its motion isochronous by confining the bob to a parabola of
revolution. The suspension ribbon was hung over a curved
surface that altered its effective length according to the
amplitude of the bob. He used this conical pendulum to
illustrate the combination of rectilinear motion and a central
force, and to apply this to analysing the motions of planets and
comets - a subject which eventually caused the great enmity
between Newton and he. This work was later published by
Huygens without reference to Hooke (Horologium Oscillatorium,
Part V, p.157).
In Horologium Huygens describes a weight driven
pendulum clock constructed with a device to ensure that the
pendulum was isochronous by forcing it to swing in an arc of a
cycloid. This was accomplished by placing two evolutes of
inverted cycloid arcs on each side of the pendulums point of
suspension against which the pendulum was constrained to
move. Unfortunately, friction along the arcs causes a greater
error than that corrected by the cycloidal path.
Graham utilised Galileos model for his dead beat
escapement. Harrison, being a bellringer, used the model of an
English bell hung for full circle ringing. Whilst raising to and
lowering from the upright position the bell is controlled by the
rope and wheel and the timing of each swing determined by the
Description of Conical Pendulum in Hookes Notes
89
clapper striking the bell.
. . . so as to afford from what is, or ought to be due to their
Hanging, the best Foundation for your Ringing in general, and
most especially therein for fine Raising and Ceasing
John Harrison A True and Full Account of the Foundation of Musick 1770
From raising and lowering bells in peal i.e. sounding a
descending diatonic scale with the heavy and light bells striking
evenly, Harrison would have known that a compound pendulum
does not exhibit simple harmonic motion and therefore their
amplitude has to be closely controlled to ensure isochronicity.
Pendulums having a small amplitude where the sine of
the angle is nearly equal to its measurement in radians, (i.e.
where the arc and the chord are nearly equal) maintain their
isochronicity with slight variations in amplitude. Therefore
Graham designed his escapement for a pendulum with a heavy
bob which had a small amplitude and minimum circular error,
whereas Harrison favoured a pendulum with a lighter bob
swinging with a greater closely controlled amplitude.
But, both from reason and a little experience, I find it not conve-
nient for the Pendulum of a Clock to Vibrate in small Arches, not
if such Arches, tho a little different, were performd in equal
Times (but it is demonstrated that when a little different they are
but performd very nearly so) for it is impracticable for a Clock to
maintain the Motion of a Pendulum in exceedingly small Arches,
without disturbing its truth. And if the Arches be a little bigger,
twill then to bring it to exact truth require a Cycloid, but still
impracticable. For if we suppose the Pendulum Wheels in sundry
Clocks to have the same draught, at any the same distance from
their Centers, and each Pendulum Wheel to have the same number
of Teeth, or we suppose em to have different draughts directly as the
number of Teeth; we must then allow that the Weight of their Pen-
dulum Balls ought to be reciprocally as the Square of the Arches
they then describe, in regard to their Motion being maintaind,
checkd, and resisted by the Wheels For let us instance in two
Clocks whose Pendulum swing seconds, and supposing the
Pendulum Ball of one to be 2 Pound Weight, and to Oscillate in
an Arch of 12 Gr, and the other Pendulum so orderd by its
communication to the Wheel, as to Vibrate but 3 Gr; tis plain., in
the latter there will be as much command to irregulate the Motion
of a Ball of 32 Pound Weight, as in the other of its 2 Pound;
because the Wheel can impress 4 times the force upon it, when as
at the same time the Ball has but I of the force arising from descent
that the other has; or (which is the same thing) but the same
90
Momentum, was the Ball 8 Pound Weight. Therefore to suspend
a Ball for a little Vibration, will require a strong spring at the
Top, which to apply for truth to such a very little portion of the
Cycloid, as will then be made use of, is impracticable. Nay, if the
spring was weak, tis not so practicable, as in a larger Vibration.
But a strong spring and heavy Ball are scarcely practicable with a
Cycloid, tho the Vibration was larger, except the Pendulum was
proportionably longer. But we have here supposd the draught of
each Pendulum Wheel the same: But if we suppose of that draught
to do for the lesser Vibration, twill then vary more in its draught;
because there is more difference geometrically (as we may almost
compare it) betwixt O and 1 than betwixt 3 and 4, but this
difference in draught will not (in such contrivance as treated of
Sect, the 5) irregulate the Motion of the Pendulum; only twill have
as much, or more need of a Cycloid But we must here suppose the
Ball to be 8 Pound Weight, so consequently with its spring not
practicable so as to be regulated as above. That there will be more
variation in a lesser draught of the Pendulum Wheel., than in a
greater, may thus be more clearly shewn. Let the Pendulum swing
seconds as before, and let the Pendulum Ball for the lesser
vibration be cut 2 Pound Weight, i.e. equal in Weight to the other
Pendulum Ball; then the Weights which draws the Wheels in each
Clock, ought to be in duplicate Ratio of the Arches the Pendulums
describe: And as this lesser Arch is here supposed to be cut 1 of the
other, the Weight which draws these Wheels, ought to be equal to
1/16 Part of the Weight which draws the other Wheels; for then the
Pendulum would but have as much power to regulate the motion,
as in the other. Now I know by experience yet much less than 1/16
Part of the Weight, which is able to maintain the large Vibration,
will in fine Summer Weather cause the Pendulum Wheel to move
forward; but 1/16 Part of the Weight is not sufficient in all sorts
of Weather even to cause if to move forward much less to maintain
this less Vibration, with as constant a draught , as the larger
Vibration is maintained; except the works cod be made to move
16 times freer, the which I cod fairly demonstrate to be impossible:
Therefore a Clock cannot be made to go so true, with a little
Vibration, as with a larger. Now his variation or different draught
of the Pendulum Wheel proceeds from the communication of all the
Wheels, which in such contrivance is very little, when the Wheels
are drawn by a convenient Weight, and the same in regard to the
Pendulum, as if the Weight which draws the Wheels was increased
of decreased; but not so in the other Clocks. But from what is said
below there may seem to spring a small objection, which I think
convenient here to answer; and i.e. that the Pallats drawing the
Pendulum Wheel a little back every beat, the difference in the
Friction amongst the Wheels cannot be the same in respect to the
91
Pendulum, or different Weight to draw the Wheels; because when
the Friction is most, the Pendulum Wheel has less course forward,
and not so willing to be drawn back.
John Harrison 1730 Manuscript
Harrison resolved the problem of circular error by
greater regulation of the amplitude of the pendulum with his
asymmetrical escapement. In his explanation of the required
difference in weight of a bob swinging with large amplitude
compared to one with small amplitude he has described the
inverse square relationship of the maximal velocity and
associated kinetic energy of the pendulum.
But further, a Cycloid applyd to a Clock Pendulum,
must not be exactly as tis demonstrated for a Pendulum to move
in Vacuo. Because tis here Maintaind by the draught of the Wheels
against the Airs resistance, which requires the Curve or Evolute
of the Cycloid, instead of being the Curve AB, to be the Curve
AC. Supposing em but very little different. This need no more
explaining. Now a Pendulum, whose Motion is maintaind
against the Airs resistance by the draught of such Wheels as above,
moving in such a Cycloid as treated of in the last section, and also
Naturally making it self rather shorter when warmer, will perform
its reciprocations, whether described in greater or less Arches,
exactly in equal Times. For, let us suppose the Clock Pendulum
to describe any certain Arch, and the Air at any certain Weight:
Now if the draught of the Wheels be increased, so that the
Pendulums Motion is maintaind in a greater Arch, when the Air
remains the same, then with such a Cycloid as above the Clock will
go true. And again, suppose the Air is lighter, and the Pendulum
describes a greater Arch without any more force from the Wheels,
the Clock will still go true; because the same Natural cause, which
makes the Air lighter, will also make the Pendulum shorter: if
orderd as explaind,
John Harrison 1730 Manuscript
Harrison repeated his interpretation of the action of the
pendulum in Concerning such Mechanism
92
Pendulum Analysis
. . . . a heavy body tends downward with as much force as it is
necessary to lift it up.
Galileo, De Motu 1589
Galileo analysed pendulum motion as a series of
tangential motions down inclined planes. This concept was
further developed by Newton when he wrote in Principia about
the cycloidal motion of a pendulum. During the eighteenth
century there was a dependence upon proportional reasoning,
continuing belief that pendulum motion was identical to that of
a body sliding under the action of gravity, along a smooth
circular surface and an understanding of the near co-incidence of
the arc of circle and its chord for small angles. Reading the
notes of Saundersons lectures helps us to understand these
concepts.
Harrison played the viol and in his will identified his
monochords as part of his valued possessions. He compared
the properties of the pendulum to that of the vibrating string.
There is an example in his documents of the application of
Hookes Law. From his experience as a bellringer he would be
able to envisage the ellipsoidal nature of these vibrations. Using
eighteenth century reasoning we can describe the motion of a
pendulum as follows and interpret Harrisons documents.
Galileo published the relationship between the length of
a pendulum (L) and its period of oscillation (T) in Dialogue on
the Two Principle World Systems in 1632. In1637 he suggested that
the square of the period is proportional to the length of the
pendulum for small amplitudes (T
2
% 1).
The conventional mathematical modelling of the simple
pendulum is helpful in understanding the forces involved but in
the real world the bob is not a point mass and the suspension
rod is also rigid with mass. The angular frequency and all other
related quantities will be dependent upon the moment of inertia.
Gravity acts through the centre of mass of a rigid body. Hence
length of pendulum used in the equation for the period is equal
to the distance between the pivot and the centre of mass.
93

I
T 2
mgh
=
The above relationship has no indication of the
amplitude and therefore assumes that the pendulum it represents
is tautochronous, i.e. it oscillates with the same frequency
whatever the amplitude.
This was discussed by John Wilkins in 1688 in his An
Essay Towards a Real Character and a Philosophical Language. Moment
of inertia was introduced by Euler in his book a Theoria motus
corporum solidorum seu rigidorum in 1730. In
this book, he discussed at length moment
of inertia and many concepts, such as
principal axis of inertia, related to the
moment of inertia.
In order to determine the
relationship between the length of a
pendulum (L) and the period of oscillation
(T) we need to consider that when a
pendulum swings, its mass drops through
a vertical distance, h. The velocity, v, of a
mass undergoing uniform acceleration , a,
at the end of one second will be equal to a,
and a velocity equal to a will be added
during each subsequent second. Therefore
at the end of t seconds we have v = at and
the average velocity is
av
1 1
v v at
2 2
= =
So that the displacement,
2
a
1 1 1
s v t vt at t at
2 2 2
= = = =
h
H
L
O
B
D
C
L
q/2
q/2
q/2
94
From v = at, ,
2
2
2
v v
t or t
a a
= =
giving
2 2
2
1 v v
s a
2 a 2a
= =
or
2
v 2as =
For a mass falling freely from rest under the action of gravity
the terminal velocity will be
2
max
v 2gh =
The vertical distance, h, is found by determining the chordal
length, H, from the two equilateral triangles OBD and OCD and
approximating in radians for sin for the small angles.
H 2Lsin
2
q
=
which for small angles
H L = q
It then follows that
h Hsin 2Lsin . sin
2 2 2
q q q
= =
1/4 1/4
h 2Lsin . sin
2 2
q q
=
which for small angles

2 2
1/4
L
h 2L.
4 2
q q
= =
95
For a simple pendulum, the distance s travelled from the top of
the swing to the equilibrium point, (the point of maximum
velocity)the distance along the arc of swing, is
1/4
s L = q
Therefore
s L
v
T T
q
= =
Substituting in we obtain
2
max
v 2gh =

2
2
L L
2g.
T 2

q q

=

which simplifies to give the expression

2
L
T
g
=
Because the expression above is
dimensionless and relates to only one of four quarters of the
period it should be more correctly written as
L L
T or T k
g g
=
It illustrates Galileos assertion that the mass of the bob
does not affect the period but does not include the amplitude
because of the simplification sin = . But Harrison knew
from his experience in raising and lowering bells that the period
was dependent upon the amplitude and that the motion of a
compound pendulum is not purely sinusoidal.
The period is more correctly given by the elliptic
function:-
2 2
2 4 6
1 1 3 1 3 5
1 sin . sin . . sin . . . . .
2 2 2 4 2 2 4 6 2
l
T
g

+ + +

| | | | | |
=
| ` | |
\ . \ . \ .
)
96
Harrison copied notes taken at Saundersons lectures
explaining Newtonian mechanics. Newton had stated his laws of
motion in Principia . His wrote his second law as :-
LAW II
The alteration of motion is ever proportional to the motive force
impressed, and is made in the direction of the right line in which that
force is impressed.
If any force generates a motion, a double force will generate double
the motion, a triple force triple the motion, whether the force be
impressed altogether, and at once, or gradually or successively. And
this motion, (being always directed the same way with the generating
force), if the body moved before, is added to or subtracted from the
former motion, according as they directly conspire with or are directly
contrary to each other; or obliquely joined, when they are oblique, so
as to produce a new motion compounded from the determination of
both.
Sir Isaac Newton, Principia
Lagrange dispensed with the Newtonian concept of force
and reformulated motion in terms of kinetic and potential
energy. The total energy of a simple pendulum is the sum of its
kinetic and potential energy.
= =
1 1
2 2
mv mgh giving mv mgh 0
2 2
Harrison had shown that he comprehended this when he
compared the maximum velocity of balance wheel vibrating 5
times per second with the maximum velocity of a 13 foot
pendulum.
97
Yet from several observations, I still endeavour'd to make further
Corrections in this Motion; and in these 3 last Years have brought a
Clock to go nearer the truth than can be well imagin'd, considering the
vast Number of seconds of Time there is in a Month, in which space of
time it does not vary above one second, and that mostly the way I
expect: So I am sure I can bring it to the Nicety of 2 or 3 seconds in
a Year.
John Harrison, 1730 Manuscript
Harrison was aware of the problem of circular error when
he devised his escapement and also the relationship of the mass
of the bob to the total energy of the pendulum. (One second per
day error is 11.6 parts per million) The graph illustrates that
if the amplitude of the pendulum varies by 1/30th the circular
error of the Graham escapement is 0.309 parts per million and
with the RAS regulator amplitude is 9.29 parts per million.
Harrisons escapement must either compensate for circular error
or control the pendulum more strictly if his claims are to be
The Limb of the Ballance moving thereby through about 25 Inches in a
Second, notwithstanding, as in that Time (and still an Augmentation to its Power
or Dominion) its Motion being changed to contrary Direction 5 Times, as as from
which it must as in Consequence be as farther allowed, and as withal considering
the Arch it describes, that its Motion must be prodigiously quick, or even violently
powerful in the Middle of each Vibration, and when, as even without that, 25
Inches in a Second, is no less than 34 Miles a Day; so rapid and powerful is the
Motion of the Ballance, for faint sleepy Work could never do; and whenas no such
Velocity (as in this Point for the better) can be in a Pendulum, viz. in such Arches
as in the long Pendulum Way are commonly, or can as the most properly be
described, until it comes to about the length of 13 Foot,
John Harrison, Concerning Such Mechanisms, 1775
25 inches per sec. = 2.08 feet per second
For a 13 foot pendulum with 12E amplitude (2 = 6E)
h=2L(sin 2/2)
2
= 0.0710
v
2
= 2gh = 64h = 4.55
v = 2.13 feet per second as Harrison suggested.
98
realised.
A strong Proof indeed, that the Force of Draught of the
Pendulum-Wheel of my Clock, and as in this first Place to be
understood, with its right duly adapted Proportion, and that as partly
for its Number of Teeth, corresponding to its Revolution of 4 Minutes,
but thence in chief, or as indispensably so to be acquired thereby, [viz.
from which the said Proportion] such a Qualification as whence, by my
Construction of the Pallats the said Force so to be disposed of, as to give
to the Pendulum no more Irregularity in this Maintenance of its
Motion in Air, than as if it went, or could go by itself, so as to be
observed to what Truth it had continued to measure Time, viz. as by
or from its going by itself, and that for a long Time in Vacuo; and
therefore I may ask, if such a Matter be not highly worthy
Encouragement, what other Sort of Ingenuity or Discovery in the
World must be so?
John Harrison, Concerning Such Mechanism, 1775.
According to Newton's laws the inertia force F
I
(i.e., mass times
acceleration) has to be equal to the applied force. In our case,
the applied force is the restoring force F
R
caused by gravity G.
From the geometry of the problem, it is clear that
99
R
F = -G sin = - mg sin , q q
where m is the mass of the pendulum
and g is the acceleration of gravity. Note
that the negative sign is caused by the
fact that the restoring force F
R
wants to
bring the pendulum back to equilibrium
(i.e., = 0).
Next, we have to express the
inertia force F
I
in terms of the angle 2.
Assuming a rigid pendulum (i.e., its
length l is fixed), the mass can move
only on a circle with radius l. The
position (i.e., the spatial coordinates ) along this circle is given
by l2. Note that the angle 2 is measured in radians (i.e., 180
corresponds to pi). The acceleration is therefore given by .
2
2
d
l.
dt
q
Thus, from Newton's law we get
2
2
d
ml = -mg sin .
dt
q
q
Dividing by ml and moving the term on the right-hand side to
the left-hand side leads to the equation of motion of an
undamped and undriven pendulum
2
2
g d
+ sin 0
dt l
q
q =
The mass m of the pendulum does not appear anymore in the
equation of motion. Galileo was the first who discovered this
effect. From the viewpoint of Newton's laws, there is no reason
that the inertial mass (i.e., the m in F
I
) has to be the same as the
gravitational mass (i.e., the m in F
R
). It was the ingenious idea
of Albert Einstein (1879-1955) to take this equality as a
physical principle. From this equivalence of gravitational and
inertial mass, he developed a new understanding of gravity which
led to his general theory of relativity.
100
But Harrison relates the energy in the pendulum to its
amplitude and the mass of the bob.
And first, as letting the proper Circumstances, Quality, or Efficacy
of the Pallats for a Pendulum as above treated of, and as when together with
such their other unparallelled Properties as below, to be fairly or rightly
considered, it must as in Consequence, and that as here in the first Place follow,
[viz. as from the Construction of the said Pallats, and distance inversely from
the Centre of Motion of the Pendulum, to where the Force or Draught of the
Pendulum-Wheel is by them communicated to the said Pendulum] that the
Power a Pendulum must as thence have, so as whereby to regulate a Clock,
must, I say, as in supposing the Bob of a certain Weight, to be about as the
Square of the Arch it describes; therefore, as by Corollary, it must then follow,
that if a Bob of 3 Pounds Weight be sufficient to regulate a Clock, when the
Pendulum describes an Arch of 12 Degrees, 48 Pounds must be required to do
the same Execution, or to be the same in Power, whereby to regulate, if the
Pendulum describes but 3 Degrees.
John Harrison, Concerning Such Mechanism, 1775
The solution to may be written as
2
2
g d
+ sin 0
dt l
q
q =
0 0
sin( t ) q =q v +j
where is the angular amplitude and is the initial
0
q
0
j
phase angle whose value depends upon how the pendulum was
started i.e. its initial conditions.
The angular velocity is therefore
0 0
d
cos( t )
dt

q
= q =q v v +j
These two relationships are plotted as a time series graph
and are readily seen to be 90 out of phase with each other.
This may also be shown with a phase plane diagram. where the
equations for angle and angular velocity are considered to be
parametric equations for which the parameter is proportional to
time. The orbit of this phase trajectory is the ellipse
2 2
2 2
0 0
1
( )

q q
+ =
q vq
101
As we have assumed that this motion has no friction or
forcing, energy is conserved on this phase trajectory. Therefore
the sum of potential and kinetic energies at any time can be
shown to be constant.
In the linearized approximation
2 2 2
1 1
E ml mgl
2 2

= q + q
giving the energy at maximum displacement as :-
2
0
1
E mgl
2
= q
as described by Harrison when he relates the mass of the
bob to the amplitude.
The equation of motion of an undamped and undriven
pendulum is nonlinear because it depends non-linearly on the
angle 2.
It therefore cannot be solved algebraically but has to be
analysed numerically.
From what has been said we may gather, that the draught of
the Wheels, or the force impressd on the Pendulum of such a Clock is
nearly always alike, i.e., not only always nearly equal, when the Wheels
are drawn by a certain Weight (for in adjusting a Clock, whose
Pendulum moves in a Cycloid, there is occasion to increase and
decrease the draught of the Wheels by adding to, and taking from the
Weight, as may be seen below): But I mean in a more especial manner
equal, no obstruction or irregularity arising from Friction, be the
draught of the Wheels, or chiefly that of the Pendulum Wheel more or
less, which is so predominant in other Clocks, that tho their Pendulums
movd in Cycloids, and were such as describd below; yet they could not
be brought to truth.
As shown by the quotation above Harrisons concept for
his pendulum is that of a damped harmonic oscillator driven by
constant torque. Damping and driving are caused by two
additional forces acting on the pendulum: The damping force
and the driving force. The equations of motion of a damped
and driven pendulum is an extension of the equation of the
102
undamped and undriven case. The relevant forces for a forced
and damped pendulum are conceptualised by:-
Acceleration + Gravitation + Damping = Escapement Force
The main causes of damping in Harrisons pendulum are
the insignificant constant frictional losses in the suspension
spring and the viscous damping of the bob moving through air.
Now by experience I cannot find, that the Pendulum ever alters in its
Vibration above 1/30 part of the whole Vibration, and part of what it
alters must be owing to the Air, and the other part we cannot well
attribute to difference in Friction in such Pallats as treated of above;
For certainly when the Friction increases in other Joynts cf the Clock,
there will less force come to the Pallats; and consequently, as they are
not Oyld, nor slide at the Wheel, the Friction must there rather
decrease. So the less force the Wheel impresses on the Pallats or
Pendulum, with less the Pendulum can draw it back, as Nature
requires.
John Harrison, Concerning Such Mechanisms, 1775
Thus, the equation of motion of a pendulum can be
represented by:-
2
2
2
d
ml mgl sin
dt
q
= q
which can be simplified to
2
2
g d
+ sin 0
dt l

=
and including damping becomes
2
2
g d d
+ sin 0
dt l dt

+ =
where ( is the damping constant.
If a sinusoidal horizontal periodic force is used to drive
the pendulum the equation for a driven and damped pendulum
then becomes.
103
( )
2
2
g d d
+ sin Acos t 0
dt l dt

+ =
where A and Tt are the amplitude and the phase angle of the
driving force. The viscous damping term is linear, but the driving
term is linear only in the case of a constant external periodic
force as used by Harrison.
And to prove or adjust the Cycloid to Vibrations performd
in different Arches as required I cause the Pendulum to describe such by
increasing or decreasing the draught of the Wheels, and that by adding
to or taking from the Weight: by which I can make 8 or 10 times more
difference, than Nature ever will, and yet the effect be nearly the same
as if Nature itself had alterd the Weight of the Air so much.
John Harrison, 1730 Document
Driving puts energy into the system
which is dissipated by the viscous damping.
On average, there has to be a balance
between the energy input of the escapement
and the dissipated energy. When Harrison
improved the efficiency of the escapement in
the in the Brocklesby Hall clock he had to
control damping by fitting a vane to the
pendulum bob.
From what has been said we may gather, that the draught of
the Wheels, or the force impressd on the Pendulum of such a Clock is
nearly always alike, i.e., not only always nearly equal, when the Wheels
are drawn by a certain Weight (for in adjusting a Clock, whose
Pendulum moves in a Cycloid, there is occasion to increase and
decrease the draught of the Wheels by adding to, and taking from the
Weight, as may be seen below): But I mean in a more especial manner
equal, no obstruction or irregularity arising from Friction, be the
draught of the Wheels, or chiefly that of the Pendulum Wheel more or
less, which is so predominant in other Clocks, that tho their Pendulums
movd in Cycloids, and were such as describd below; yet they could not
be brought to truth.
The angular velocity of the forcing, , may be
D
v
different from the natural frequency of the pendulum and so the
equation of motion may be rewritten in dimensionless form as:-
104
( )
2
D 2
d 1 d
+ . +sin Acos t
dt Q dt

=
in Liebnizian notation or
D
1
. sin Acos( t)
Q
q + q + q = v

in Newtons notation, where Q is the damping or quality
parameter and is the drive frequency.
D
v
If a variable is introduced as the phase of the driving f
term the equation of motion can be rewritten as a set of first
order equations:-
These equations fulfil the necessary conditions for
describing chaotic motion, that is
a. a system with at least three independent dynamical
variables, , , w q f
b. equations containing non-linear terms that couple
the variables; sin and Acos clearly non linear. q f
The amplitude of the pendulum is therefore never
completely stable unless it is rotating through 360 but
approaches regularity when the energy losses (damping) are equal
to the energy input (escapement force).
This is a deterministic system which exhibits aperiodic
behaviour that depends on the initial conditions.
D
d
dt
d
dt
d 1
. sin Acos
dt Q
f
= w
q
= w
w
= - w - q + f
105
Poincar developed a geometrical approach for this
analysis by defining trajectories in a phase space. Thus for a
known initial condition where x
1
(t) represents the position and
x
2
(t) represents the velocity of the pendulum then the solution
(x
1
(t),x
2
(t)) corresponds to a point moving along the trajectory.
The phase space is completely filled with trajectories, since each
point can serve as an initial condition.
A sample phase
diagram in three dimensions
is shown, which gives an
indication of the appearance
of a regular activity with
random variation. This is
what is observed with the
pendulum. Before this had
been properly explained
using the chaos theory it was
thought that this apparently
random variation was caused
by frequency modulated and
amplitude modulated noise.
The initial driving force imparted to the pendulum when
the clock is started may be different in amplitude or phase from
the equilibrium state. A difference in amplitude will produce
amplitude modulation
where is arbitrarily small
(t) 0
f A cos t =
(t) 0
A A
and a difference in time will produce frequency modulation
where is arbitrarily small.
0 (t)
f A cos t =
(t) 0

106
Pendulum Damping
The main factors damping a pendulum are frictional
losses in the suspension spring and losses due to air viscosity and
drag.
All gas molecules are in continuous random motion.
When there is bulk motion due to flow, the bulk motion is
superimposed on the random motions. It is then distributed
throughout the fluid by molecular collisions. The stronger the
random motion such as caused by increasing in temperature, the
stronger is the resistance to the bulk motion of flow. Theoretical
analyses based on these simple considerations predict that gas
viscosity is proportional to the square root of the Absolute
temperature of the gas.
J. Clerk Maxwells 1866 Bakerian Lecture published as
On the viscosity or internal friction of air and other gases (Philosophical
Transactions of the Royal Society of London, 156, pp. 249-268)
gives experimental details and a mathematical model of the
damping force of a vibrating disk suspended by a thread (i.e. a
pendulum). He determined that:-
Viscosity is independent of pressure in ambient conditions
and
Viscosity increases as temperature increases.
The viscosity of air depends mostly on the temperature.
At 15.0 C, the viscosity of air is 1.78 10
-5
kg/(m.s) or 1.78
10
-4
Pa s. and at 23 C it is 1.834 10
-4
Pa s. At the
atmospheric temperature range it increases by 4.93 10
-7
Pa s per
C. (W.N.Bond 1937, Proc. Phys. Soc. 49, pp. 205-213).
But a Pendulum moving in a Cycloid to Regulate a Clock
true, which is not in Vacuo, must not always keep its same Length,
unless the Air gave always the same resistance; which agrees both with
reason and experiment, viz. the Pendulum must be rather shorter in
Warm than in cold Weather, which is contrary to the operation of the
wire.
John Harrison, 1730 manuscript
Harrison over-corrected his pendulum length for a rise in
temperature to compensate for the increase in air viscosity. In
addition to the frictional losses due to friction there is losses
caused by pushing a mass of air out of the way of the moving
107
pendulum bob, proportional to the area of the bob on a plane
perpendicular to the direction of motion and the density of the
air. Lord Rayleigh, successor to J. Clark Maxwell devised the drag
equation which can be applied to gases as well as fluids:-

2
D D
1
F u C A
2
=
where
F
D
is the force of drag
is the density of the gas
u is the velocity of the object
A is the cross sectional area, and
C
D
is the drag coefficient - a dimensionless constant dependent
upon temperature.
F.W. Bessels paper Untersuchungen ber die Lnge des einfachen
Secundenpendels (Investigation into the Length of a Simple Seconds
Pendulum) published in 1826 showed from theoretical
considerations the necessity of considering the inertia of air as
well as buoyancy in reducing the effective gravity. this was further
reviewed by Sir George Gabriel Stokes in On the Effect of the Internal
Friction of Fluids on the Motion of Pendulums, Trans. Camb. Phil. Soc.
(1850) vol IX, p.8.
Changes in barometric pressure also affect the density of
the gas and therefore the buoyancy of the bob. This accounts for
the documented changes in the amplitude of a pendulum in a
vacuum (e.g Shortt regulator). However the thermal insulation
provided by the vacuum must not be neglected when assessing the
improvement of using an evacuated chamber for an accurate
pendulum. Harrison made no correction for changes in
barometric error, considering that the effect was in part
responsible for the variance of rate per day and that these
variations would average out.
Therefore ambient air affects the period of the pendulum
in three ways. Firstly by Archimedes principle the effective weight
of the bob is reduced by the buoyancy of the air it displaces The
mass (inertia) remains the same, so that there is a reduction in
the effective gravitational force which increases the period of the
108
pendulum. Secondly, the pendulum carries an amount of air with
it as it swings, and the mass of this air increases the inertia of the
pendulum, again reducing the accelerating gravitational force and
increasing the period. Thirdly, viscous air resistance slows the
pendulum's velocity requiring more energy from the escapement
to maintain the amplitude of the pendulum
Escapement Error
Having lately had occasion to investigate the disturbance
produced in the motion of a pendulum by a small external force, and
having found by a very general investigation a result of great simplicity,
I perceived that the general theorems for the alteration in the time and
extent of vibration produced by the difference between cycloidal and
circular arcs, by the resistance of the air, by the friction at the point of
suspension, &tc. could be made to depend upon it; and that these
alterations could in fact be found with greater facility from this general
theorem than from the independent and unconnected investigations. I
found also that the same principles could be applied with great ease to
that important practical subject, the escapement of clocks.
George Biddell Airy, On the disturbance of Pendulums 1827
Escapement error is the quantisation of the difference in
period of the forced pendulum and its natural period.
And for my strict trying the Motion, I have two Clocks, but of that by
and by.
John Harrison, 1730 manuscript
. . . . . . by my Construction of the Pallats, the said Force so to be
disposed of, as to give to the Pendulum no more Irregularity in the
Maintenance of its Motion in Air, than as if it went, or could go by
itself, so as to be observed to what Truth it had continued to measure
Time, viz. as by or from its going by itself, and that for a long time in
Vacuo;
John Harrison, Concerning Such Mechanism, 1775
In his writings Harrison indirectly refers to the Q of his
pendulum. The concept of Q relating to pendulums is from an
analogy with resonance in electronic tuned circuits. Q can be
defined as the ratio of the maximum potential force to the
necessary restoring force. If we start a pendulums motion at a
109
known amplitude then the number of vibrations before it comes
to rest is dependent upon its Q. Therefore the Q of a
pendulum is dependent upon its amplitude. Harrison requires
a pendulum that has a bob of small mass having large amplitude.
He understood how to determine the accuracy of his
clocks. He rates them long term by using sidereal time as
described by Jeremy Thacker (Dr. Arbuthnot), but also
compared the going of two similar clocks.
I have two ways for trying the truth of the Motion, the which
together is very Compleat. One is by the apparent Motion of the fixed
Stars with a very large sort of an Instrument, of about 25 Yards
Radius, composd of the East side of my Neighbours Chimney, (which
is situated from my House towards the South) and the West side of an
exact place of some one of the upright parts of my own Window
Frames; by which the Rays of a Star are taken from my sight almost in
an Instant: And I have another person to count the seconds of the Clock,
beginning a little before the Star Vanish: So I observe what second is
Mentions when it Vanisheth; and I have a Table Calculated to show
how much sooner any such Star is to Vanish every night, or before the
24 Hours of the Pendulum Day is expired.
But my other way is the better part of the Completion: and
that is the two Clocks placd one in one Room and the other in another,
yet so, that I can stand in the Doorstead, and hear the beats of both the
Pendulums, when the Clock Case heads are off; and before or after the
hearing can see the seconds of one Clock, whilst another Person count
the seconds of the other: by which Means I can have the difference of the
Clocks to a small part of a second. And in very Cold and Frosty
Weather, I sometimes make on Room very warm, with a great Fire,
whilst the other is very Cold. And again the contrary. And sometimes
the like in Summer by the Suns Rays in at the Windows of one Room,
and also a Fire, whilst the other is shut up and Cool. Thus I prove the
Operation of the Pendulum Wires, and adjust the same to what is
treated of above. And to prove or adjust the Cycloid to Vibrations
performs in different Arches as required I cause the Pendulum to
describe such by increasing and decreasing the draught of the Wheels,
and that by adding to and taking from the Weight: by which I can make
8 or 10 Times more difference, than Nature ever will, and yet the effect
be nearly the same as if Nature itself had alterd the Weight of the Air
so much
John Harrison, 1730 Manuscript
110
He notes that the rate of his clocks varied by up to one
second each day, but that there was a drift of one second in a
hundred days. W.M.H.Greaves, Astronomer Royal for Scotland,
presented a paper in 1943 ( The Short-Period Erratics of Free
Pendulum and Quartz Clocks W.M.H.Greaves and L.S.T. Symms,
MNRAS 1943 vol. 103 pp.196-209) giving the results of
comparing three clocks thereby eliminating changes in rate
caused by Coriolis and tidal forces. The method used was then
further developed by D.R.Allan (AVAR). If an attempt is made
to derive Gaussian variance it is found that the results diverge
rather than converge due to the erratics described by Greaves
and Symms. However if a long sequence of data is taken,
divided into bins based on the averaging time, t, the average of
the results found, then the difference from the mean of each bin
can be found. The square root of the mean of the squared
differences (Allan Deviation) gives a quantitative measure of
how much the average value changed at that particular averaging
time.
2 2
1
1
AVAR ( ) ( )
2( 1)
+
=

n n n
t y y
n
Using one of his clocks Harrison could carefully time
the non- forced pendulum at its working amplitude, in the same
manner that Kater timed his compound pendulum using one of
Sheltons regulators. He could then determine the difference in
rate caused by his escapement when this pendulum was driven in
its clock. Adjustments could then be made to reduce escapement
error to a minimum. By comparing the rate of two similar
clocks when he was adjusting the escapement he was able to
eliminate changes due to ambient temperature and diurnal
changes in gravity.
As first, the bare length of a Pendulum can be no otherwise
rightly considered or esteemed, but as only to what it bears, or may bear
in Proportion to the Length of the Pallats, and as together with such
improper Powers or Circumstances therunto belonging, or may, as
farther therunto belonging; i.e. in other Words to the equivalent
Distance from its Centre of Motion, to where the Pallats, according to
their Construction as in the common Way touch or are applied to the
Wheel; the Pendulum as with respect to the length of the Pallats, and as
here in the first Place to be notified, being no more than as about14
111
J.L.Lagrange (1736-1813) was an Italian astronomer and
mathematician who succeeded Euler at the Prussian
Academy of Sciences in Berlin. He had come to the notice
of Euler with his essay on the tautochrone using maxima
and minima. His Mecnique Analytique published in 1788
was the most comprehensive treatise on classical mechanics
since Newton. This work is notable for having no diagrams
and its use of differential equations. After Frederick II died
he left Berlin to become a member of the Acadmie des
Sciences in Paris. As a foreigner he was tried by the French
Revolutionaries but released on the testimony of Lavoisier.
He was unfortunately unable to aid Lavoisier when he was
convicted and guillotined but made the infamous comment
It took them only an instant to cut off his head, but France
may not produce another like it in a century.
to 1, so the which in Effect is not better, or can in Power be no better,
than as a long Pendulum rendered a short one.
John Harrison, Concerning Such Mechanism, 1775
Harrison played the bass viol (viola da gamba), a bowed
and fretted string instrument. A bow is most effectively applied
in all string instruments, (violin, viola, violoncello etc.) 1/14
to 1/14 proportionally distant from the end of the string (the
bridge). A string player moves the bow closer to the bridge as
he makes the effective length of the string shorter by stopping.
Harrison investigated the behaviour of strings using his
monochord and was probably the first to compare the pendulum
to a vibrating string.
112
Temperature Compensation
Harrison describes his gridiron pendulum constructed
of steel and brass rods in the 1730 manuscript. A pendulum,
of this form is fitted to the 1728 Harrison regulator owned
by the Worshipful Company of Clockmakers. This is most
probably the only gridiron pendulum made by Harrison in
existence. A similar pendulum was made for the RAS
regulator by J. H. Agar Baugh for Gould in 1928 to replace a
wooden rod pendulum. A gridiron pendulum in a dis-
proportioned clock case is painted in the upper right hand
corner of the King portrait. In this painting the clock case is
elongated, so that the details of the pendulum can be seen
above the chair on which Harrison is sitting.
Changes in temperature of the rods occurs by
conduction and convection from the surrounding air, and
radiation. Radiation is minimised by situating the clock in a
shady place and the outer case minimises the transfer of heat
by convection and conduction from the surrounding air.
The 1728 pendulum has two sets of compensating
rods; one above and one below the crutch rod. Close
examination of this pendulum has shown that the rods are
pinned differently in the upper set compared to the lower,
suggesting that they were made at different times and a
longitudinal slit is provided for the crutch. In his 1730
drawing he has shown the compensating rods starting a
significant distance below the suspension and describes
making this part of the rod in brass.. The expansion due to
heat of this length of the rod can be accommodated by the
lower compensating rods. This part of the pendulum would
expand and contract to the same degree as a crutch rod made
of similar metal and so there need not be any sliding friction
between the crutch and the pendulum if their centres of
oscillation are on the same axis.
Harrison would not have designed a pendulum with
compensation above and below the crutch unless he required
excessive temperature correction. He always looked for the
simplest effective solution to a problem; compensation
required for the upper part of the pendulum above the crutch
can be provided by the gridiron in the lower part of the
113
pendulum below the crutch as he described in the 1730
manuscript. Calculations show that if gridiron compensation
is fitted above the crutch action point there is markedly over
correction of the pendulum. He also knew that mass in the
upper part of the pendulum reduced its effective length.
The description in the 1730 manuscript shows
overcompensation of the length of the pendulum. Harrison
further emphasises the necessity for this in Concerning Such
Mechanisms.
Let s be a thin Brass spring suspending a Natural Pendulum
in a Cycloid, and B the Ball of the Pendulum; and let the thick upright
Lines represent Wires of the Sheffield Brass (as in sect 9) and the small
Lines Wires of Steel. These are fixed fast in Over-thwarts. or Bridges;
of which there are 3 at the Upper end, and 2 at the Lower, and this
Pendulum will make itself shorter when Warmer, as required sect
14.15; tho all the Wires be then lengthened. See the Computation as
followeth. And first, let us suppose it at a certain Length, and the
Weather at a certain Temperature, and let us suppose the Brass spring
(etc.) at the Top to be 1/5 of the length of the other part of the
Pendulum, viz the Wires. And then supposing the Weather to be
Warmer, and each Wire to be lengthened (as in sect the 9) viz Steel
12.6 indefinitely small parts, the Brass 23 such parts, and the Brass
Spring at the Top 1/5 of that i.e. 4.6. But the parts will be bigger or
less according to the degrees of Heat.
Brass at the Top 4.6
The two out most Steel Wires from the highest Bridge at the upper end,
to the lower, at the lower end 12.6
So this lowest Bridge will be too low 17.2
Then the two outmost Brass Wires from the lowest Bridge, to the
Middle-most at the Top 23.0
So this Middle Bridge will be too high 5.8
Then the next two Steel Wires from this middle Bridge at the upper
end, to the upper Bridge at the lower end 12.6
So the upper Bridge at the lower end will be too low 6.8
Then the other two Brass Wires from this upper Bridge, at the lower
end; to the lowest Bridge at the upper end 23.0
So this lowest Bridge at the upper end will be too high 16.2
114
Then the Steel Wire down the Middle, from this lowest Bridge at the
upper end, to the Center of the Ball about 13.3
So the Pendulum is shortened about 2.9
John Harrison 1730 Manuscript
The description and drawing in the 1730
document show that the compensating part of the
pendulum occupies 5/6 part of the rod. The drawing
also shows positions for the crutch pin which are at 1/9,
1/8, and 1/7 part of the length of the pendulum away
from the suspension. The compensating part in the
Guildhall pendulum is 4/5 of the length of the rod. The
arrangement of the rods is the same as in the 1730
document. The middle steel rod is fitted with an
adjustment to enable part of it to be made of brass.
This allows adjustment of the compensation, particularly
necessary if the top 1/5 of the rod is made of steel.
Without this compensation and the upper part made of
steel the pendulum would shorten excessively with
increasing temperature.
The pendulum illustrated in the portrait is a further
development. The disposition of the rods in this pendulum
are reversed with the outer steel rods fixed to the bob rather
than to the upper suspension. Confirmation of this is the
number of bridges shown at the lower end compared to his
description in the 1730 manuscript.
Fine adjustment of the behaviour of the gridiron is
made with pins through a sliding crosspiece at the lower end
to the outer steel rods and an outer short brass rod. The
length of the outer brass rods gives a greater range of
adjustment than is possible with the Guildhall arrangement,
although the holes shown in the portrait only allow a small
range. Harrison wishes to over-compensate for changes in
temperature to allow for alteration in the resistance offered
by air. Air resistance is increased by and increase in
temperature and by increased barometric pressure. There is
no direct compensation for changing barometric pressure in
the clock or pendulum. He associated high barometric
pressure with warm weather and conversely low barometric
pressure with cold weather. To allow for this vague
115
correlation and the change in air viscosity he shortened his
pendulum with increase in temperature. He must have
determined the degree of compensation required by
experiment and introduced a means of adjusting it in his
clocks.
Harrison measured the coefficients of thermal
expansion. In his 1730 document he gives relative figures for
the expansion of parts of his pendulum. When these figures
are compared to the associated parts they give coefficients
that are remarkably close to our modern accepted values. In
his units brass was 34.52 and iron 15.87. The ratio of these
coefficients is 1.825. If we assume that he was using Rmers
temperature scale these coefficients are equal to 23.33/ K
for brass and 10.72/ K for steel. The accepted average
modern value for brass is 20.5 10
-6
/ K, and for mild steel
11.5 10
-6
/ K.. The ratio of these modern values is 1.783.
His determination of the coefficients of thermal
expansion given in the 1730 document indicate that he was
most probably using Sheffield brass for the rods of his
Pendulum Rod Material Properties
CTE Hardness Sp.Ht Density Th.Cond
Free Machining Brass 35.5 %Zn 3.1%Pb 20.5 BNH 80 0.38 8.49 109
CZ120 Leaded Brass 33.5 %Zn 20% Pb 20.3 BNH 78 0.38 8.5 109
Yellow Brass 34 %Zn 20.3 BNH 58 0.38 8.47 110
Cartridge Brass 30 %Zn 19.9 BNH 54 0.38 8.53 114
Low Brass 20 %Zn 19.1 BNH 57 0.38 8.67 134
Red Brass 15 %Zn 18.7 BNH 56 0.38 8.75 146
Mild Steel 0.27-0.34 %C 11.5 BNH 120 0.486 7.87 64.9
Cementation Steel 0.74-0.88 %C 11.1 BNH 140 0.49 7.85 64.9
Iron 12 BNH 110 0.45 7.87 86.5
Ratio CTE
Iron 12.6:13.5 11.9
Sheffield Brass 12.6:23 20.3
London Brass 12.6:20.5 18.1
Holland Brass 12.6:19 16.8
Copper & Silver 12.6:17.6 15.5
116
pendulum. Calamine was mined in the Halkyn mountains
(Rhosesmor) in North Wales from 1720. The ore from this
source has a higher lead content than that from Somerset.
Harrisons Sheffield brass if made from this source of
calamine and Anglesey copper would have excellent machining
properties, thermal expansion and hardness as he describes.
But a Pendulum moving in a Cycloid to Regulate a Clock true, which is
not in Vacuo, must not always keep its same Length, unless the Air
gave always the same resistance; which agrees both with reason and
experiment. Viz. the Pendulum must be rather shorter in Warm than
in Cold Weather, which is contrary to the operation of the Wire.
John Harrison 1730 Manuscript
There is a hint of the influence of Robert Hooke
in this comment in 1730 of the motion of pendula in
a vacuum. Hooke managed Robert Boyles vacuum
pump for many years and demonstrated many
Ole Christensen Rmer (1644-
1710), Danish astronomer who
made the first quantitative
measurements of the speed of light
in 1676 using Cassinis method of
timing the eclipses of Jupiters
moon, Io, from different places on
earth. He invented the meridian or
transit circle and the alt-azimuth
mount for telescopes. He
introduced standard weights and
measures throughout Denmark and developed the first
temperature scale for an alcohol filled thermometer in 1701.
Zero was defined as the freezing point of saturated brine and the
boiling point of water as 60. Newton had previously created a
scale based on the expansion of linseed oil.
Daniel G. Fahrenheit (1686-1736) introduced his modified
Rmer scale in 1724 having the freezing point of water as 32
and the boiling point 212.
Anders Celsius (1701-1744) developed his temperature scale in
1742.
117
experiments with it to the Royal Society in Gresham
College. Jeremy Thackers pamphlet of 1714 has an
illustration of his proposed sea clock which is enclosed in
an evacuated flask. The illustration also shows his
suggestion for timing using a distant chimney piece as
described by Harrison in his 1730 manuscript.
This is probably the first reference to air
viscosity affecting a pendulum. It was further developed
by Bloxam. (See Bloxam, J.M., Memoirs Roy. Astr. Soc.,
22, 1853.) The brass rods of the pendulum in the
portrait and the 1728 regulator are thicker than the steel
ones. This can also be discerned in the drawing in the
1730 manuscript even though he used shading to
indicate the different materials. He had quantified
coefficients of expansion by his own investigations before
1730 and had probably also realised that materials had
their own distinctive specific heat capacity.
But there is no Wire, of any Metal whatever, whereof to make
a Pendulum, but what is continually altering its Length according to
the degrees of Heat or Cold; and this I discoverd above 2 years ago,
before I ever heard a Word of, or read any about it. But a little after I
had discoverd it I was shewd the same in a Book, which had its been
sooner would have been some service to me.
John Harrison 1730 Manuscript
Harrison particularly notes that the length of the
wire continually alters with temperature i.e. he considered
it to be in a dynamic rather than a steady state. He
published his comparative measurements of the
expansion of metals in the 1730 manuscript. He
remarks that Sheffield brass is harder than the other
supplies. The addition of lead to brass makes it harder.
Although Harrison draws the brass and steel rods
having different diameters in his 1730 manuscript he
does not mention this in his script. He has presented the
concept of thermal diffusibility. Thermal diffusibility, ",
is a measure of how quickly a material can absorb or lose
heat from its surroundings. It is the ratio of thermal
conductivity to the product of density and heat capacity.
118
p
k
C
a =
r
where k = thermal conductivity and = volumetric
p
C r
heat capacity
Fouriers Law is an empirical law based on
observation. It states that the heat flow through a
Q
t
D
D
homogenous solid is directly proportional to the area, A,
of the section at right angles to the direction of heat
flow, and to the temperature difference along the path of
the heat flow, .
T
x
D
D
The rate at which heat can pass through a
material is determined by its thermal conductivity.
Thermal conductivity, k, is defined through the
equation:
J. B. Joseph Fourier (1768-1830), campaigned with Napoleon in
Egypt and became Secretary of the Academie des Sciences. A French
mathematician who exerted a strong influence on mathematical
physics through his Theorie analytique de la chaleur published in
1822. He showed how the conduction of heat in solid bodies may
be analysed in terms of infinite mathematical series now called by
his name, the Fourier series. He proposed that any function could
be written as an infinite sum of the trigonometric functions cosine
and sine. Fouriers achievement is better understood if we see it as
twofold: treating first the formulation of the physical problem as
boundary-value problems in linear partial differential equations,
which extended rational mechanics to fields outside those defined
in Newtons Principia; and second, the powerful mathematical
tools he invented for the solution of the equations, which yielded
a long series of descendants and raised problems in mathematical
analysis that motivated much of the leading work in that field for
the rest of the century and beyond
119
Q T
kA
t x
D D
=-
D D
where

is the rate at which heat flows across the area
Q
t
D
D
A in Watts (joules per second) and

is the change in
T
x
D
D
temperature over the distance x in degrees Kelvin per
square metre. Thermal conductivity, k is a property of
the material.
When heat flows from one material to another,
the temperature of the contact layer of the cooler
material increases. From the contact layer thermal energy
has to spread throughout the colder material by
conduction of convection. The rate at which the
temperature in the contact area increases depends upon
the specific heat and volume of the material.
The increase in temperature in each of the
different rods is an exponential growth that can be
described by the equation :
t
T Ce
a
D =
where T is the change in temperature, C a constant, ",
the exponential decay constant and t the time interval.
The time constant, ", is the reciprocal of the exponential
decay constant, J, ( ). For the different rods to
1
t=
a
expand at the same rate we need to have the same time
constant for each.
The time constant is proportional to the specific
heat and mass and inversely proportional to the thermal
conductivity and the area of heat exchange.
sp.ht density volume
tau
thermal conductivity area

which can be reduced to when comparing rods of unit

length to
120
2
c. . r l c. .r
2 rl.k 2k
r p r
t = =
p
For 2.00 mm. diam. mild steel having a specific heat of
0.486 kJ per kg. per K, density 7870 kg. per cubic
metre, and thermal conductivity of 64.9 Watts per metre
per K

0.486 7870 1.00
29.46
2 64.9

t = =

For 3.175 mm. diam. brass having a specific heat of

0.380 kJ per kg. per EK, density 8500 kg. per cu. metre
and thermal conductivity of 109 Watts per metre per E

0.380 8500 1.587
23.51
2 109

t = =

The disparity between the time constants of these

rods can be lessened by a matt black finish on the steel
rod and a highly polished finish on the brass rod. The
length of the short, lower, outer brass rods is probably
chosen because of the heat sink properties of brass and
also to maintain a sense of proportion.
Changes in the temperature of the rods occurs by
conduction and convection from the surrounding air, and
by radiation. Radiation is minimised by situating the
clock in a shaded place. An outer case minimises the
transfer of heat by convection and conduction from the
surrounding air.
Joseph Black , (1728-1799), Professor of Medicine,
Glasgow, discoverer of carbon dioxide, defined latent
heat and investigated and quantified specific heats with
the aid of James Watt and described them to the
University Philosophical Club in 1762. James Watt,
(17361819), mathematical-instrument maker had spent
a year in London before taking up his post at the
University with Black.
121
The suspension spring of the pendulum is
supported by radial cheeks. Harrisons 1730 drawing of
his temperature compensated pendulum shows cycloidal
arcs attached to the suspension and he writes
Let s be a thin Brass spring suspending a Natural Pendulum in
a Cycloid, and B the Ball of the Pendulum;
John Harrison 1730 Manuscript
The suspension block of the RAS regulator has
two adjusting screws. The machined cheeks have a radial
profile. Furthermore the cheeks can only be adjusted by
re-machining in the lathe. Harrison most probably used
thin rolled bronze ribbons attached to the pendulum
which were adjusted by the screws.
The pendulum bob is drawn as 3 inches
diameter in the 1730 manuscript. The bob in the 1728
regulator is 3 inches diameter and the bob painted in the
King portrait is estimated to be 3 inches diameter also.
The proportions of the parts in the portrait have been
used to determine the dimensions of the pendulum rods
with a range of temperature correction similar to
Harrisons ideal.
A pendulum of the form shown in the portrait was
made by Thomas Earnshaw for the Armagh Observatory.
The clock was specified by Nevil Maskelyne in 1789 and
installed at Armagh in 1794. The pendulum was found to
be overcompensated and was replaced with a mercury
compensated pendulum in 1830. A second pendulum, of
the same design was made by Earnshaw. It is now fitted to
the Buchanan clock at Armagh. The movement of this
clock is identical to those made by Matthew Dutton in the
late 1790s in London. Matthew Dutton, partner of
Thomas Mudge had been apprenticed to George Graham.
The advantage of the gridiron pendulum compared
to Grahams mercury compensated pendulum is its rapid
response to changes in temperature. It is probable that the
excellent performance of the Shortt regulator and similar
regulators that have their pendulums in an evacuated
chamber is mostly due to the insulation provided by the
122
vacuum preventing changes in temperature. The
disadvantage is the increased weight of the rod compared
to the bob, which makes its performance more compound
and the poor rigidity of the rods. The rigidity of the rod is
improved in the later form of the gridiron.
The pendulum in the Guildhall regulator is very
different from that described in the 1730 document. It
does not have a brass top-piece and it is fitted with a tin-
whistle adjuster. Calculations have been made assuming
that the top-piece is not original but the adjuster is, to
determine dimensions for a pendulum as shown in the
King portrait having the characteristic described in the
1730 document.
John Shelton (1690-1778), a previous employee of George
Graham, made five regulators for use in the observation of the
transit of Venus in 1761 and 1767. Three of these clocks were
carried by Captain Cook on his voyages. Nevil Maskelyne used one
of these clocks for his observation of the transit of Venus in 1761
in St. Helena, for checking the performance of H4 on its trial to
Barbados and for his experiments on gravity at Schiehallion. The
pendulum in the Royal Societys Shelton Regulator RS 35 is most
probably from another of the series but original. It has the same
form as the pendulum in the King portrait, but lacks the
adjustment feature. The negative temperature compensation (2.8)
is similar to Harrisons recommendation. Shelton described
Grahams dead beat escapement for the Gentlemans Magazine in
1754, three years after Grahams death.
123
Calculat ion of Thermal Compensat ion and Lengt h
of Rods for Gridiron Pendulums
20.5 Coefficient of Linear Thermal Expansion Brass 10
-6
/EC
11.5 Steel 10
-6
/EC
1730 Manuscript
Modern Units Harrison Units
3.26 158.8 mm. 4.6 Top Piece Brass
9.13 794.1 mm. 12.6 1st Steel Descender
-16.28 794.1 mm. -23.0 1st Brass Ascender
9.13 794.1 mm. 12.6 2nd Steel Descender
-16.28 794.1 mm. -23.0 2nd Brass Ascender
9.13 794.1 mm. 12.6 3rd Steel Descender
0.51 44.1 mm. 0.7 Bottom Piece Steel
-1.40 -2.9 Compensat ion
Guildhall Regulator with Plain Steel Top Piece and Adjuster
Max Mid Min
2.13 185.2 mm. 2.13 185.2 mm. 2.13 185.2 mm. Top Piece Steel
8.52 740.7 mm. 8.52 740.7 mm. 8.52 740.7 mm. 1st Steel Descender
-15.18 740.7 mm. -15.18 740.7 mm. -15.18 740.7 mm. 1st Brass Ascender
8.52 740.7 mm. 8.52 740.7 mm. 8.52 740.7 mm. 2nd Steel Descender
-15.18 740.7 mm. -15.18 740.7 mm. -15.18 740.7 mm. 2nd Brass Ascender
7.10 617.7 mm. 7.10 617.7 mm. 7.10 617.7 mm. 3rd Steel Descender
0.00 0.0 mm. 1.15 56.1 mm. 2.30 112.2 mm. 3rd Brass Descender
1.41 123.0 mm. 0.77 66.9 mm. 0.12 10.8 mm. 3rd Steel Descender
1.52 74.1 mm. 1.52 74.1 mm. 1.52 74.1 mm. Bottom Piece Brass
-1.17 -0.66 -0.16 Compensat ion
Guildhall Regulator with Compensated Top Piece and Adjuster
Max Mid Min
1.78 86.6 mm. 1.78 86.6 mm. 1.78 86.6 mm. Top Piece Brass
-1.02 98.6 mm. -1.02 98.6 mm. -1.02 98.6 mm. Compensator
8.52 740.7 mm. 8.52 740.7 mm. 8.52 740.7 mm. 1st Steel Descender
-15.18 740.7 mm. -15.18 740.7 mm. -15.18 740.7 mm. 1st Brass Ascender
8.52 740.7 mm. 8.52 740.7 mm. 8.52 740.7 mm. 2nd Steel Descender
-15.18 740.7 mm. -15.18 740.7 mm. -15.18 740.7 mm. 2nd Brass Ascender
7.10 617.7 mm. 7.10 617.7 mm. 7.10 617.7 mm. 3rd Steel Descender
0.00 0.0 mm. 1.15 56.1 mm. 2.30 112.2 mm. 3rd Brass Descender
1.41 123.0 mm. 0.77 66.9 mm. 0.12 10.8 mm. 4th Steel Descender
1.52 74.1 mm. 1.52 74.1 mm. 1.52 74.1 mm. Bottom Piece Brass
-2.54 -2.04 -1.53 Compensat ion
Portrait Pendulum
Max Mid Min
1.83 158.9 mm. 1.83 158.9 mm. 1.83 158.9 mm. Top Piece Steel
9.13 794.3 mm. 9.13 794.3 mm. 9.13 794.3 mm. 1st Steel Descender
-16.28 794.3 mm. -16.28 794.3 mm. -16.28 794.3 mm. 1st Brass Ascender
9.13 794.3 mm. 9.13 794.3 mm. 9.13 794.3 mm. 2nd Steel Descender
-16.28 794.3 mm. -16.28 794.3 mm. -16.28 794.3 mm. 2nd Brass Ascender
8.77 763.0 mm. 8.20 713.0 mm. 7.62 663.0 mm. 3rd Steel Descender
0.64 31.3 mm. 1.67 81.3 mm. 2.69 131.3 mm. 3rd Brass Descender
0.62 30.0 mm. 0.62 30.0 mm. 0.62 30.0 mm. Bottom Piece Brass
0.69 60.0 mm. 0.69 60.0 mm. 0.69 60.0 mm. Bottom Screw Steel
-1.75 -1.30 -0.85 Compensat ion
158.8 mm
794.1 mm 740.8 mm
X
5X
142.4 mm
185.2 mm
Y
4Y
76.20 mm
37.86 mm
95.25 mm
78.23 mm
0 10 30 50 70 90 10
Pendula Compared
Pendulum Suspension
The earliest clocks with verge escapement had the
escapement directly coupled to the pendulum. Later
development of the flexible suspension spring led to the
escapement forces being conveyed to the pendulum through a
crutch as described and illustrated by Christiaan Huygens.
Harrison indicates that he is aware of the interaction between
pendulum, crutch and escapement in his drawing of the
geometry of his escapement. All his regulators have knife
edge suspension for the crutch.
Let A represent two bits of Glass, whereon each end of the
Axis of the Pendulum being a Sharp Edge of Brass moves in a little
notch in the Glass, which renders that Friction insensible, as also the
Wearing (for tis but the Axis of the Pallats, which communicate force
to the Pendulum; for the Pendulum is itself otherwise suspended, viz.
by a thin Brass Spring).
John Harrison, 1730 Document
The pendulum suspension fitted to the 1728
regulators is attached to the wooden frame of the clock but he
notes that it is essential to fix the clock case firmly to a brick
wall.
The Pendulums being suspended by the Clocks, and Clocks by the Cases
(as all commonly are) might not be deceivd by the suspension of the
Pendulum, I made two holes in each Clock-Case Back, through which
into the seams of the Bricks I drove two strong Irons, without touching
or depending on the Case at all, to which I screwd the plate by which
the Pendulum and Cycloid are suspended.
John Harrison 1730 Document
The suspension spring fitted to the 1726-30 clocks
have a metal ribbon suspended from a slotted carrier on a
backplate. It is then firmly held between two pivoted and
linked curved cheeks, one sprung and the other having a
graduated adjustment screw, A. The cheeks have attached
pieces with two further adjustment screws, B and C. The
maximum angle between the cheeks is 18. The pivoted parts
have stops to fix their position. Screwing A downwards
reduces the length of the suspension spring by lowering the
126
point of contact. The length of the lever and the fine pitch
of the screw facilitate its accuracy. This arrangement enables
fine adjustment of the position of the cheeks and thereby fine
adjustment of the pendulum length without stopping its
motion. Even though Harrison understood how to rate a
clock or watch he would nevertheless wish to regulate his
accurate clocks to keep as nearly possible normal time.
Rawlings stated this problem succinctly :-
If you try to get a very fine regulation by a rating nut at the bottom of
the pendulum, you will soon find that clocks are kittle-kattle, and that
merely stopping the pendulum and starting it again may cause its rate
to change by two or three seconds a week, even though you do not touch
the regulator.
A.L.Rawlings, The Science of Clocks and Watches, 1948
Harrison further developed his pendulum suspension
for the RAS regulator. He made adjustment in the lateral
axis much easier and because he constructed the suspension
separate from the clock he fitted registration pillars to the
backplate.
For, as to this weighty Article, it is to be understood, viz. from the
Construction of my Pallats, (as intimated above) that the Force of my
Pendulum Wheel is so disposed of, as that, whether the said Force at
any Time becomes somewhat greater or less in itself, or that the Air at
any time gives somewhat more or less Resistance to the Pendulum, or
both, it is the same as here above required, [but indeed this must be with
supposing, as in the Note, or as taking for granted, that the Pendulum
must be rather shorter when warm than when cold] and so by means of
something at the Top, but not according to Mr. Huygenss
A
B C
127
Demonstration of the Cycloid, [for that would not suit the Matter at
all] it is brought to such a Degree of Perfection, as not to gain or lose
so much as one Second in a Month; the Suspension of the Pendulum [a
Thing highly Material in the Matter] being from the Brick-Wall, as
having nothing to do with, or as having no Dependence from the Clock,
its Case or Wainscot, for if not so, the true Goodness of what I have
shewn could not be had.
John Harrison, Concerning Such Mechanism, 1775
The backplate of the pendulum suspension of the
RAS regulator is made to be bolted to a wall. It has two
tapered arms by which the regulator is connected to the
suspension and held in place with screwed coques. These
enable the clock to be parted from the suspension for
adjustment and returned exactly to its prior position.
The pendulum is supported by a split steel projection
from the backplate. The central split is wide enough to
accommodate a thin suspension spring and rolled brass
cheeks. Two plates with radiused faces are screwed to the
backplate These plates abutting the pendulum have provision
for lateral adjustment and there are two calibrated capstan
adjustment screws which can act on the extended cheeks
attached to the pendulum. These screws can alter the shape of
the curved cheeks thereby enabling very fine adjustment of the
pendulum length without stopping its motion.
128
Suspension Spring
A pendulum has three primary modes of oscillation; it
can have left to right (coronal) motion, front to back
(anterior-posterior) motion and rotation about the long axis
of the rod. The Coriolis force that causes the tendency to
oscillate in the anterior-posterior direction was demonstrated
by Foucault with his pendulum having a 26 kg. bob on a 67
metre wire. The pendulum rotates through 360 (2B
radians) in a sidereal day at the poles but 360 sin NE (2B sin
N radians) in a sidereal day at a latitude N. The pendulum
rotates approximately 284 per day in London. A pendulum
restrained to oscillate in the coronal direction by its
suspension spring will therefore experience small changes in
gravitational force having a maximum every 15 hours. There
is also a diurnal and semi-diurnal variation in gravity due to
the gravitational forces of the sun and moon. These are the
same forces that create tides and can be harmonically
analysed.
Rotation along the long axis of the rod is most
commonly caused by an irregularly shaped bob. The
Guildhall pendulum bob has a plano-convex lenticular shape.
The bob illustrated in Kings portrait is also plano-convex and
has a wide rectangular connection to its rod to minimise
rotation.
Two secondary modes of oscillation occur; the
pendulum rod vibrating sideways like a string with a node at
the middle of the rod, and the horizontal vibration of the top
of the rod at the bottom of the suspension spring. All modes
of oscillation except the left to right motion of the bob have
to be suppressed to eliminate undesirable errors.
Harrison achieves minimal error in his ultimate
regulator by using a ribbon spring firmly held between radial
adjustable cheeks.
I have said that Mr. Huygenss Demonstration of the Cycloid
can be of no Service in the Affair, viz. no farther than as it it was to be
129
supposed in Vacuo, &tc. if that may said to be of Service; and whenas
[whether in Vacuo or not] was it to be applied [accordingly as demon-
strated] to a very thin Spring at the Top of the Pendulum, [for thin to
the purpose it must be] it would only occasion the said spring to neck,
or break off at the top of the Cycloid, and that in a short Time; nor can
the Arch or Arches, as described in the long Pendulum Way, [as
supposing at the most 15 Degrees] be but hardly said to want it, [viz.
as in the Manner demonstrated] therefore such a Matter as must be in
lieu of what is called a Cycloid, must be chiefly to preserve the Spring
[or Suspension of the Pendulum] from its ever breaking, and the which
Spring, as being to be very thin, ( but may be short) must be best to be
made of Gold, properly allayed with Copper, and to be well hammered
before it be brought to its Thinness, [as being then more elastic, than as
or when allayed with Silver]. Now the Nature of such a Matter, or
Cycloid to the Purpose, [and as consequently withal for preserving the
Spring] must be as in some Measure reverse to what is demonstrated by
Mr. Huygens, &tc. that is, it must be so as to occasion little Vibrations
of the Pendulum, viz. all such as are less (and unregarded) than so as
to let, or such as will let the Pallats interchange, to be still sooner per-
formed, than what they would be as otherwise be without it; and at such
an Arch describing, as whereby just to let the Pallats interchange, or
rather as a little bigger, the Length of the Pendulum to be so [viz. as by
or from its adjusting] as then to swing Seconds, and also, as when in
its fetching farther [ as from the Nature of such a Cycloid as must be,
and as when together upon such other Foundation as above described]
the same; for as thence, from the Continuation of the circular
Curvature of the Cheeks, [viz. of this artificial Cycloid] that Matter,
as here in Hand, is to be ascertained, but of the Radius, in each to the
Purpose. viz. as subsequent to the Action of the Pallats I shall not
here speak, nor can Cambridge and Oxford Education have any Thing
to do with either that or the Action of the said Pallats, [viz. so as to
define such a Matter or Matters to Exactness, had such Particulars as
them ever before been thought of]; but however, as each cheek, with
regard to the Property I have shewn of the Pallats, or as a Tenor to
their Result, must be the Arch of a Circle, [viz. for so far as will, in
this material Point, be suitably wanted or used, i.e. besides an Overplus
of the same for its more truly making] it can be done to a mathematical
Truth, whenas the other (as according to Mr. Huygens) could not, was
it so to be wanted.
John Harrison, Concerning Such Mechanism, 1775
When Robert Hooke experimented with pendulums
he suspended his bob on a long flexible ribbon or cord. In his
130
drawing of a pendulum in the 1730 manuscript Harrison
shows it suspended by a thin brass spring occupying one fifth
of the length of the rod.
Let s be a thin Brass spring suspending a Natural Pendulum
in a Cycloid, and B the Ball of the Pendulum;
John Harrison 1730 Document
When making his wooden clocks he avoided using
steel because of the corrosive effect of oak on iron. When he
introduced the tin-whistle adjuster into his gridiron rod he
was able to use a steel suspension spring. Youngs modulus
of elasticity is greater for steel than brass. This meant that
together with the different tensile strength the width and
thickness of the suspension spring could have a smaller area.
Because of its ductility Harrison considered that the
work hardened alloy of gold and copper (9 ct. red gold)is the
most appropriate material for the suspension spring of his
ultimate regulator. He also realised that the flexible
suspension spring must be short.
The length of the suspension spring governs the
length and radius of its circular arc produced by the swing of
the pendulum. The proportion of circular arc to linear part
will depend upon the weight of the pendulum bob and the
stiffness of the spring. As the pendulum swings to its
extremity the centre of rotation of the centre of gravity is
lowered. The design of the supporting cheeks with their
graduated adjusting screws facilitates controlling the dynamic
morphography of the spring. The curved cheeks prevent point
contact with the spring limiting stresses which would
ultimately cause its breakage. Harrison considered that the
circular arc created in this way approximated closely to the
involute of a cycloid and referred to it as an artificial cycloid.
Harrison has chosen to pivot the pallet frame and
crutch arbor on knife edges. His pendulum is suspended on a
firmly supported thin ribbon. Because the spring ribbon
bends the centre of rotation of an equivalent simple
pendulum alters according to its amplitude. But the
momentum of the compound grid-iron pendulum also exerts
131
a side thrust on the spring which pushes the centre of
rotation laterally. He would have compared the interaction
of the crutch and pendulum to the behaviour and transfer of
forces of a clapper in a swinging bell.
So now to the Purpose it is to be understood, that from the
Force or Draught of the Pendulum-Wheel, as being by the Pallats
properly disposed of, and as wherein with a proper recoiling of the same,
that the Cycloid may be so, as that when the said Wheel may have [or
as when occasioning the Wheel to have] somewhat a greater Force, I say
the Vibrations of the Pendulum, as thence becoming bigger, may still be
performed in the same Time, and as also the same when the Air gives
or may give somewhat a lesser Resistance to the Pendulum, so that its
Vibrations become somewhat greater without any more Force from the
Wheel, and this, because for the most part, that from the same Cause
which will lessen the Resistance of the Air, the Pendulum is to become
in itself, as mathematically speaking, shorter; and indeed, from the
whole of what I have shewn, a great Exactness in the measuring of
Time is to be had; for withal, it is to be considered, that the Draught or
Force of the Pendulum-Wheel, in my Sort of Clockwork, will alter but
very little, neither the Vibration of the Pendulum, as from thence, or
any other Cause, much at any Time, or in any Sort of Weather.
The curvature of the suspension spring depends upon
many variables, viz. the length, width, thickness and material
of the spring, the weight and moment of inertia around its
centre of gravity of the pendulum, and the point of contact
and the magnitude of the force applied by the crutch. Further
factors affecting the performance of the spring a changes in
temperature altering the characteristics of the material and air
resistance. Harrison gives a qualitative description in
Concerning Such Mechanism . A comprehensive quantitative
analysis based on observation was given by K.James , Precision
Pendulum Clocks - Circular Error and the Suspension Spring
Antiquarian Horology, Vol. 8, pp. 868-883.
The tensile strength, F, is the maxima of the
stress/strain curve. The first part of this curve up to the yield
point is linear as the material obeys Hookes Law. Below the
yield point deformation is recoverable, the material returning
to its pre-load state (elastic deformation). Beyond the
maxima point necking occurs; the curve dips because the
132
cross-sectional area is reduced. The cyclical flexing of the
suspension produces fatigue. This begins as surface cracking
leading to precipitous failure. The ductility of the gold and
the surface work hardening prescribed by Harrison improve
the properties of the suspension. A cross sectional area
should be chosen to give the desired flexibility whilst
maintaining elastic deformation. The midpoint force imparted
by a lead pendulum bob 76.2 mm. diameter and mass 1.035
kgm. is 10.15 Newtons. This is 10.15 Mpa when exerted on
an area of 1 square millimeter and has been marked on the
following graph.
If the upper section of the pendulum rod is made of
steel then the most suitable material for the crutch rod is also
steel. The axis of rotation of the crutch and centre of rotation
of the pendulum are the nearly concentric. The suspension
block of the RAS regulator allows adjustment of both axes.
The rod shown in Cottinghams photographs is almost
certainly made of steel and had probably rusted badly when
Gould came to restore the regulator. Gould replaced it with a
heavy adjustable brass fixture which totally altered the
dynamics of the escapement.
The frequency of oscillation of the pendulum is
dependent upon its length. If we wish to alter the rate of the
Steel Tensile Strength = 300 Mpa
Brass = 200 Mpa s
9 ct Gold = 102 MPa s
L/L
MPa
10.15 MPa
Y
o
u
n
g

s

M
o
d
u
l
u
s

E

=

2
0
0

G
P
a
E

=

1
2
5

G
P
a
E

=

7
9

G
P
a
Stress
Strain
133
clock we alter the length of the rod. To alter the rate of a one
second pendulum by one second in thirty days we have to alter
its length by 0.00077 mm. It is impossible to make this
adjustment by stopping the pendulums motion and turning
the nut below the pendulum bob. Harrison introduced his
method of adjusting the length at the suspension spring at the
top of the rod in his 1728 regulators. Screwing A
downwards reduces the length of the suspension spring by
lowering the point of contact. The length of the lever and the
fine pitch of the screw facilitate its accuracy.
The adjustment screws have a pitch of 56 t.p.i. and
have 20 graduations. It is estimated that each graduation
represents a movement of the screw of 0.0226 mm. which
will give less than 1 second per month alteration.
But as here to return to the Clock, it must be indeed, as at
first to get such a Clock to be really adjusted or truly adjusted . . . . .
There being to be concerned in that Proceeding four different
things, and wherein two of which (as touching the Point) do as it were
pretty much conspire to, or in the same Purpose, viz. the Composers of
the Pallats to relative Rest, and the correspondent Curvature thereto of
the Cycloidal Cheeks. I say, these two Things may only as almost be
taken as one, viz. in their joint Effects, for so far as belongs to this
Matter, but not quite so; and the other two Particulars in the Case, are
the adjusting [viz. by proper Provision] the Wires, or the redoubling
of the Wires for Heat and Cold in the Pendulum, and the Screw for
fast and slow in general at the Bottom of the Pendulum.
John Harrison, Concerning Such Mechanism, 1775
134
Ageing of the Pendulum
Yet from several observations, I still endeavourd to make
further Corrections in this Motion; and in the 3 last Years have
brought a Clock to go nearer the truth than can be well imagind,
considering the vast Number of seconds of Time there is in a Month, in
which space of time it does not vary above one second, and that mostly
the way I expect.
John Harrison, 1730 Manuscript
As noted previously a very small increase in the length
of a pendulum produces the alteration in period responsible
for one second in a hundred days. The pendulum has a heavy
bob suspended by a rod which should have a small mass.
Harrisons gridiron pendulum has five effective rods with
three being extended and two compressed by the weight of the
bob. There over a long period the effective length of the
pendulum will increase and the rate slow.
Ageing will affect:-
a. Suspension
b. Pendulum rods
c. Bob
Because the suspension is so short changes will be
imperceptible but the bob is supported in its centre and is
likely to sag over a long period. If antimony is added to the
lead as in type-metal and bullets it imparts strength, hardness
and stability to the alloy. The convex brass front face of the
bob if fused to the lead alloy further improves its stability.
Cottingham records that the RAS regulator had a five
inch diameter bob on a deal rod when he restored the clock in
1909. This pendulum would most probably have been fitted
by Frodsham in 1856-60 when the clock was in his workshop.
It does not accord with any of the comments in Harrisons
writings.
It must be observed, however, that the recoil, such as it is, takes place at
the extremities of the vibration, when the momentum of the pendulum
is almost exhausted, and is therefore the most unfavourable part of the
vibration to occur in; notwithstanding which disadvantage the
135
escapement has performed as well as any other that had preceded,
perhaps as well as any that has followed it, as the regulator to which it
was applied has seldom if ever been excelled in steadiness and accuracy
of performance for fourteen years of succession. The imperfection
arising from the recoil taking place at the extremities of the vibrations
seems to have been compensated by the immense momentum of the
pendulum that this escapement admitted of with a comparatively small
maintaining power, while the absence of friction and of oil gave it the
advantage of permanently uniform action. Of course the pendulum
was of the grid-iron form, that compensated the effects of varying
temperature, and allowing the escapement to have its merit appreciated.
The report of the performance of Harrisons regulator with the
escapement in question was, that it did not vary one second per day
from any preceding or following day for the space of fourteen years, and
that the aggregate of the variations from true mean time did not amount
to half a minute.
Rees Cyclopaedia, 1820
136
Revd. Abraham Rees (1743-1825) was
Minister at the Unitarian Chapel in Old
Jewry, London for the last forty years of
his life. In 1778-1788 he edited a
revised edition of Chambers Cyclopaedia
which had been first published in 1728.
He followed this by publishing The New
Cyclopaedia, or, Universal Dictionary of the Arts
and Sciences in parts between January 1802
and August 1820, running to 39 volumes
of text, 5 volumes of plates, and an atlas.
It contains around 39 million words, and
more than 500 of the articles are of
monograph length. It was written by
about 100 contributors, most of whom
were nonconformists. They were
specialists in their fields, covering the arts
and humanities, agriculture, science,
technology, and medicine. Its engraved
plates particularly those of plants and animals
are of exceptional quality.
Rees's Cyclopaedia is important today for the
information it contains, particularly, about the
science and technology of the period. He was a
Fellow of the Royal Society during the period
of intense development of the marine
chronometer. The entry describing Harrisons
escapement is copied as an indication of the
opinion of his contemporary clockmakers.
His portrait was painted by James Lonsdale,
(1777-1839) a prolific painter who also
painted the portrait of Sir William Congreve,
the instigator of government manufacture of
gunpowder and inventor of a rolling ball clock.
Harrison's clock escapement. The escapements which we have hitherto described require oil on the
pallets, to diminish the friction of the parts that rub against one another in action; but the
escapement we are now going to describe as the invention of the ingenious Harrison requires no oil,
on account of his construction being free from rubbing at the time of the pallets action.
Fig. 5. of Plate XXXIII. exhibits the shape of Harrison's
escapement G D O is a portion of the swing wheel, moving round
M as a centre: A is the verge or arbor of the crutch that
communicates with the rod of the pendulum; to this are attached
the pallets, each consisting of a long and a short lever with a joint
at the place of union ; the short lever A B is fast to the arbor or axis
of motion, and carries the longer lever B C which has a claw or hook
a C, and a slender spring is so applied as to keep these two levers
nearly at right angles to each other; at right angles to A B is also the
other short lever A E fast likewise to the axis of motion, and bearing
the longer lever E F, as A B bears the lever B C; the joints are at B
and e, and a similar small spring holds the levers A F and E F in like
manner at right angles to each other, when not otherwise acted on.
The lever E F has not only a claw, like the other long lever, but a
heel to it, or projecting part f. The action of this contrivance may
be thus the heel of the detent E F, in which situation it opposes the vibration through the medium
of the two levers E F and E A, which are attached to the crutch of the pendulum. The vibration being
finished the pendulum returns towards the left, and the wheel is at liberty to advance again; the
pressure of the tooth G on the claw of the detent now assists the vibration through the arc of
escapement or arc of action and causes the pal let C to approach the wheel, till the tooth following
D falls on it and experiences a recoil in its turn. This recoil sets the detent F at liberty from the
pressure of the tooth G, which now flies back by means of its slender spring to its rectangular
situation as it regards the small lever A E: presently this vibration terminates and the pendulum
returns, while the tooth following D aids its vibration by pulling forwards the pallet C, when the
process of making two successive vibrations is gone through, and the same operation is repeated.
The reader will have observed in this account that the tooth and pallet, or tooth and detent,
do not slide on one another, but give each other a direct push or pull, the business of the small
springs being to remove them from contact by one sudden leap, where rubbing is altogether avoided,
which contrivance supercedes the necessity of having oil applied to the pallets, and is certainly an
ingenious contrivance, as well as useful in practice.
It must be observed however, that the recoil, such as it is, takes place at the extremities of
the vibration, when the momentum of the pendulum is almost exhausted and is therefore the most
unfavourable part of the vibration to occur in; notwithstanding which disadvantage the escapement
has performed as well as any other that had preceded, perhaps as well as any that has followed it, as
the regulator to which it was applied has seldom if ever been excelled in steadiness and accuracy of
performance for fourteen years of uninterrupted succession. The imperfection arising from the recoil
taking place at the extremities of the vibrations seems to have been compensated by the immense
momentum of the pendulum that this escapement admitted of with a comparatively small maintaining
power, while the absence of friction and of oil gave it the advantage of permanently uniform action.
Of course the pendulum was of the grid-iron form, that compensated the effects of varying
temperature, and allowed the escapement to have its merit appreciated. The report of the performance
of Harrison's regulator with the escapement in question was, that it did not vary one second per day
from any preceding or following day for the space of fourteen years, and that the aggregate of the
variations from true mean time did not amount to half a minute.
Cyclopaedia, 1820, ed Abraham Rees
RAS Regulator Escapement
As first, or rather as here at the first, [viz. as without the
taking of any Notice of the great or chief Matter, viz. of what pertains
to different Vibrations, or rather, as more properly speaking, of what
Advantage pertains to, or accrues from the Largeness of a Vibration]
the bare length of a Pendulum can be no otherwise rightly considered or
esteemed, but only to what it bears, or may [as according to the
common Application] bear in Proportion to the Length of the Pallats,
and as together with such improper Powers or Circumstances thereunto
belonging, or may, as farther thereunto belong; i.e. in other Words, [and
as still in the first Place] to the equivalent Distance from its Center of
Motion, to where the Pallats, according to their Construction, and may,
or will as in the common Way touch, or are applied to the Wheel.
And now, as granting a large Vibration to be a necessary, it
farther remains to be understood, that a Pendulum cannot still truly or
strictly regulate a Clock, viz. to any continuing Perfection, nor could
any Thing, of or as in lieu of what is called a Cycloid, occasion it so
to do; unless, as in the first Place, the Force from the Wheel whereby its
Motion is to be maintained against the Airs Resistance, be the very
same, or can prove the very same as with respect to it in the whole, as
that its Vibrations, or [as here to speak more closely to my Construction
of the Pallets] any two of its Vibrations as next in Succession, may be
performed exactly, or, as it were mathematically speaking, in the same
Time as if at the same mathematical Length it went, or might be
supposed it would go by itself, [viz. as without any Force from the
wheels of a Clock] in Vacuo.
John Harrison, Concerning such Mechanisms, 1775
The escapement in the RAS regulator is the
culmination of Harrisons experimentation and is a further
development of those fitted to the 1726-1728 regulators. It
is the most important factor enabling the accurate
timekeeping that he claimed and study of this escapement
reveals his perceptive understanding of mechanics.
He begins Concerning Such Mechanism with a long
statement about the interaction of the escapement and the
pendulum. He considers that the length of the crutch and
diameter of the escapement wheel as a proportion of the
length of the pendulum can affect its natural period - so the
which in Effect is no better, or can in Power [as even in this Point the
139
Length] be no better, than as long Pendulum rendered a short one. The
pendulum must regulate the clock by being maintained at a
constant amplitude by a constant force properly applied.
I from being a Ringer, or taking a hint therefrom had made
use of an Artificial Cycloid (but had no name for it) some years before
I had so much as heard of Mr. Huygens name
John Harrison, An Explanation of My Watch, 1763
When he made the Brocklesby Hall clock in 1720 he
fitted a modified anchor escapement (the entry pallet has a
plane face and the exit pallet has a curved one). This
escapement proved to be unreliable, most probably because he
desired a large pendulum amplitude (8E). In 1722 it was
altered by replacing the sliding action anchor pallets with
static hinged pallets. The faces of these hinged pallets both
pointed the same way, opposite to the rotation of the
escapement wheel. As the escape wheel pushes the pallet, the
hinge moves away from the escapement wheel thereby making
the other pallet approach the wheel. When it contacts the
wheel, the continuing movement of the pendulum
(supplementary arc) produces recoil and contact between the
wheel and the first pallet is broken. The pallets are slightly
tail heavy so that gravity moves them away from the wheel
when they are not in contact with a tooth. The first pallet
therefore moves out of the path of the escape wheel and the
second pallet is then able to impulse the pendulum by
continuing forward movement of the escapement wheel.
Composers, short arms fixed to the anchor arms ensure that
the pallets move towards the escapement wheel. The small
movement of the pallets about their hinges involves little
friction. This escapement has no sliding action, no drop and
minimum recoil.
Harrison continued to develop this escapement for
his pendulum clocks by making it asymmetrical. He hinged
both pallet arms and both composers on a common axis which
was offset to one side of the centre-line on a pivoted pallet
frame. This resulted in having one pallet pushing and the
other pulling. Because the length of the pallets and their
angles of action differed, they applied differing torque from
the escapement wheel to the crutch, which was compensated
140
by the weight of the pallet frame.
Let A represent two bits of Glass, whereon each end of the
Axis of the Pendulum being a Sharp Edge of Brass moves in a little
notch in the Glass, which renders that Friction insensible, as also the
Wearing (for tis but the Axis of the Pallats, which communicates force
to the Pendulum; for the Pendulum it self is otherwise suspended, viz
by a thin Brass Spring). Let the Arch CD, and the least Vibration the
Pendulum must make to let the Wheel move forward, be similar;
therefore the Arch a.a. is described with the same Vibration by the
Center whereon the Pallats have their relative Motion, and from the
extremities of the said Arch or Vibration a.a. and with the length of
each Pallat may be described the portions of Circles ee.ff. which must
cut the Limb of the Wheel so, that there be half the space of one of the
Teeth between; because one Vibration lets but the space of half a Tooth
go. Now if these Pallats were fast at their relative Center a so that they
had no relative Motion, their Friction at the Wheel would be equal to
that of the common way, if the Arch described by Vibration was as
large, and maintaind with the same draught of the Wheels; but that
cod not be, for twoud require a deal more. But without, that, as their
length or Radius is at least 60 Times the semidiameter of the Pin, on
which they relative move, I may safely say the friction is 40 Times less
than the other, and the wearing is still less, especially in regard to the
Motion being alterd thereby; for the Wheel pressing against the End of
the Wood (or the Pallats being made lengthway of the Wood) there will
scarce be any impression made in 40 or 50 Years for the Wheel has no
drop, and its draught is but very little, in comparison to what it would
require to have with the other Pallats. But suppose it does make a little
impression; since the extremity or Edges of the Teeth has not to do, as
the common way has, i.e. to slide out of the little hole thats worn, and
then over it; for here the tooth takes the Pallat along with it, and never
departs from the same relative place, until the other Pallat take hold,
and then it departs from it perpendicular to the surface of the Pallat;
therefore no sliding, and consequently no alteration of the Motion on
this account. But again, if we suppose the Pallats to alter in length the
quantity of a little impression and a little wearing at their Center of
relative Motion; yet this, as they are applyd like Tangents to the Wheel,
can make no sensible difference in the impulse at the Pendulum in
many Years. Now by experience I cannot find, that the Pendulum
ever alters in its Vibration above 1/30 part of the whole Vibration,
and part of what it alters must be owing to the Air, and the other part
we cannot well attribute to difference in Friction in such Pallats as
treated oi above; For certainly when the Friction increases in other
141
Joynts of the Clock, there will less force come to the Pallats; and
consequently, as they are not Oyld, nor slide at the Wheel, the Friction
must there rather decrease. So the less force the Wheel impresses on the
Pallats or Pendulum, with less the Pendulum can draw it back, as
Nature requires.
John Harrison, 1730 Manuscript
Harrisons development of an escapement controlling
the pendulum is the result of his experience as a bell-ringer.
A bell hung for full circle ringing is effectively a double
compound pendulum controlled by rope and wheel. When
raising a bell from the resting position to the fully raised
position the arc of this compound pendulum is gradually
increased and determined by the sound of the clapper hitting
the bell. The forces, incremental and decremental, can only be
applied at the beginning of the raise by a rope acting on one
side of the swing of the wheel. It is only when the bell is
swinging through an arc of over 240E that the rope can be
used to apply forces to the other side of the wheel. Harrison
required a large pendulum amplitude to solve the problems
whereas Graham strove to use a small amplitude where
circular error is minimal.
For it may be notified, as just here hinted, that the Actions of
each Pallat are not equally the same upon the Pendulum, but not so, as
to be easily perceived to be otherwise, viz. As in the looking at the Clock,
or Seconds in Motion, although in the most highly material
Circumstance, or Construction of the Pallats, pretty much different,
but still, not so to be taken [viz. As with respect to the essential Point
here in Hand] as to be any the least worse for the same, but as that the
Action of one with that of the other, are quite right, for or to the
Purpose.
John Harrison, Concerning Such Mechanisms, 1775
Harrisons drawing of this escapement in his 1730
manuscript is further developed in a manuscript drawing in
the Guildhall Library (MS 3972/3 p.5). Neither of these
drawings have the same proportions or dimensions as the
escapement in the RAS Regulator. Careful analysis of these
two drawings of his asymmetrical escapement indicate his
intentions. It is important to understand the features of his
escapement demonstrated in these drawings before attempting
how he devised them. Construction of a similar drawing
142
from their design parameters helps to confirm that the
escapement in the RAS Regulator is as Harrison intended.
Now this highly material Matter is from the Wheels acting
[by my Contrivance of the Pallats] more weakly upon the Pendulum
in each and every one of its Descents, viz. from the Extremity of each
Vibration, [and there, as at the first, the weakest of all] to the Bottom
or Middle of each, and then [as still continuing] more and more
strongly upon the Pendulum in each of its Ascents, and the strongest of
all just before the interchanging of the Pallats, and the which
interchanging not being, or to be far from the Extremity of each
Vibration, and in which little, or as it were overplus Part, a necessary
recoiling of the Wheel to be, viz. as not only for the interchanging of the
Pallats, the which, as according to my Construction of the same is
required, but as withal to have in some Measure to do with the Effect
of a Cycloid, but not to be the whole concern in that Matter; the
Pendulum withal requiring to be [viz. as from my Contrivance of its
Combination of Brass and Steel Wires] rather, as mathematically
speaking, shorter when warm than when cold. But as here, to come a
little nearer in this material Point, let, as I order the Matter, the Force
[from the Wheel] upon the Pendulum, as just before the interchanging
of the Pallats, to be as by or from them the said Pallats supposed or
taken as 3, then, as just after their interchanging, [and the Force to
contrary Direction], it must be about as 2, that is, it must be so
ordered , [as may hereafter be observed by the drawing] viz. as that it
be so by the taking, or supposing for the Purpose, a Mean betwixt the
Actions of each Pallat, and withal, as farther to the Purpose, that, as
in the little recoiling of the Wheel, to become less and less to the
extremity of each Vibration, but as whence, or as still on Course, the
greater at any Time the whole Vibration may be, or more Efficacy the
same small Force [ and still as it were in the little recoiling] must from
its Quantity or Duration prove, and that in such small Measure as
required; the resters of the Pallats [viz. their Composers to relative
Rest] having for withal for the better a little to do in the Matter, and
the which is still from, or as with Respect to the Length or Shortness of
the Time [though always to be esteemed a but little] that is to be
occupied in the recoiling; it being to be understood, that was the Force
of the said Wheel, or Pendulum Wheel, to be disposed of uniformly
upon the Pendulum, throughout the whole of each Vibration, that then
the bigger that Force, as with some recoiling of the Wheel, the more it
must tend towards occasioning a Greater Vibration, or somewhat
greater Vibration of the Pendulum to be the sooner performed; but that
is not the Case in what is, as just here above shewn, but, to the Purpose
143
, is otherwise to be taken or considered; as that the bigger the Force
towards the latter End of any Vibration, viz. as in Comparison to
what it may be at the fore End or Beginning, and as in supposing
without a Recoil of the Wheel, the slower the Vibration must be, or the
longer Time it must take up, as mathematically speaking, wherein to be
performed, but as with a proper recoiling, and artificial Cycloid, rightly
adapted, not so, viz. as when in such Largeness of Vibration as must
to this Purpose, as well as to other Purposes, be as the most necessary
herein required, and as when, as must unavoidably be the Case, in the
moving by the Draught of a Wheel, or of the Wheels of a Clock in the
Medium of Air; neither, as by any Means, does a Suitable Matter of
this, viz. of the Airs Resistance, want to be avoided, as many have
foolishly imagined, but is of real or great Use, as the which I shall shew
presently; and so, as I may make bold to say, there has not as heretofore
appeared in the World, or to the Public, any Thing towards a strict, or
rather, as along with other Things of mine, to, as I may say, the most
strict or exact Mensuration of Time.
John Harrison, Concerning Such Mechanisms, 1775
Description of the RAS Escapement
144
Cottingham included a photograph of the pallet
frame, pallets and crutch in his 1909 description of the
regulator. An enlarged copy is useful to show the shape of
the crutch and its adjustment on the pallet frame arbor even
though it is shown on the wrong side of the frame.
The pallet frame is a fabricated brass assembly. Two
pan head screws have been added to fix the arbor in position.
Originally the arbor was pinned so that no adjustment of the
arbor knife edge was possible. The remaining hole for
pinning corresponds to the vertical as in Harrisons drawing
and confirms the intended angle between the pallet frame and
the knife edge axis.
The pallet arms were most probably made of
snakewood. This is an extremely fine grained resilient
timber. Harrison played the viol. Barak Norman (1651-
1724), the finest English viol maker worked at the Bass Viol
in St. Pauls Ally. Baroque bows were made of snakewood.
The characteristic mottling of this wood can be seen on the
pallet arms. Gould had to replace or repair the left pallet arm
and in so doing he reversed the position of the lead
counterweight button.
The pallet frame, without the pallet arms, composers
and crutch is balanced about its arbor knife edges. Lead
buttons counterbalance the pallet frame and the pallet arms.
The lead button, weighing approximately 2.75 grms. is added
to the pallet frame counterbalance part of the weight of the
pallet arms and composers.
The composers are fabricated from sheet brass silver
soldered. They are pivoted on the same pin as the pallet
arms and have three arms. The longest arm in both instances
has a pad by which it can press on the pallet arms. The other
two arms fix the position of the pad in relation to the pallet
frame. The inferior middle arm of the right composer has
broken off and been replaced with a bronze arm fixed with a
screw. The inferior, posterior arm of the left composer is
shorter than it should be so that this composer is not fixed in
position and can therefore not function correctly. In the
1728 wooden regulators the composers are moved into
contact with an escapement wheel tooth by a crossbar in the
145
pallet frame; they can move independently of the pallet frame.
In the RAS regulator the composers are in a fixed position
relative to the pallet frame. This enables the contact point on
the pallet face to be adjusted very accurately. The angle of
incidence of the entry pallet face facilitates adjustment so that
both pallets can be in contact with an escapement wheel tooth
at both extremities of pendulum action.
the rester of the Pallats, [viz. their Composers to relative
rest] having withal for the better a little to do in the Matter, and the
which is still from, or as with Respect to the Length or Shortness of the
Time
John Harrison, Concerning Such Mechanisms, 1775
The crutch shown in Cottinghams photographs
would have been a 142.42 mm. long bent steel rod with brass
fittings at either end (pendulum length divided by seven).
Steel has a lower co-efficient of expansion than brass and the
central rod of both forms of his gridiron pendulum are made
of steel. At the upper end a brass ring fixes it to the pallet
frame arbor with an adjusting screw. At the lower end a
horizontal pin rests in a brass fitting in the pendulum rod.
The escapement wheel has 120 teeth and a low mass which is
concentrated on the periphery so that its moment of inertia is
small. There are four slender quadroidal pointers attached to
the escapement arbor to further help in balancing.
Analysis of Harrisons Drawings
Harrisons drawings have to be examined carefully in
conjunction with the remaining parts of his RAS regulator
escapement to understand the design features of this final
version of his escapement.
1730 Manuscript Drawing
There are no dimensions on this drawing. A copy has
been enlarged so that the diameter of the escapement wheel is
76.2 mm.(3 inches) and it is shown reversed to make it
comparable with the MS 3972/3 drawing. The escapement
wheel has sixty teeth and the pallets span 11/12 teeth. The
distance between the top of the escapement wheel and the
146
pallet frame arbor is one third of the radius of the escapement
wheel. Harrison has not drawn a pallet frame/pallet arbor
circle, but he has drawn an arc passing through the pallet
pivot points of radius equal to the distance between the pallet
circle centre and his chosen lower exit pallet escapement wheel
tooth. He then drew a line from this exit pallet tooth
through the intersection of the escapement circle and the
centre-line. Torque circles have been overlaid on Harrisons
drawing, showing that the torque of each pallet is the same at
the mid swing point of the pendulum but the maximum and
minimum torque exerted by the escapement wheel is different
for each pallet.
MS 3972/3 Manuscript Drawing
Manuscript MS 3972/3 is important because it
demonstrates the development in four stages. In this drawing
he is explaining the properties and actions of his improved
escapement. When these are comprehended it is possible to
draw and dimension in a similar way.
Harrisons Escapement Geometry Drawing MS 3972/3
147
The marks of the divider points on the paper give an
indication of the important reference points and the division
marks of the circles were most probably pounced through a
stencil with graphite. Circles of radius 2 and 3" radius
were drawn and 120 division marks drawn between them. A
third circle of radius 71.21 mm. was drawn. This is one
fourteenth part of a seconds pendulum measuring 996.95
mm. (39 inches).
The lower drawing, marked A in the annotated copy,
has a pallet pivot circle radius 1/3rd of the 3 inch escapement
wheel radius with its centre-line passing through a mid-tooth
position. The pallets are acting on the 3 inch radius circle.
The line of the lower exit tooth passes through the intercept
of the apex of the escapement circle and the centre-line and is
extended to intercept the pallet pivot circle. This intercept
is marked as the lower pallet frame pivot position. The upper
pallet pivot position is marked at the same distance from its
respective tooth. The four pallet arm lines are drawn and the
torque circles drawn as tangents to them. The pendulum
amplitude lines are drawn and arcs through the four tooth
positions.
The second drawing, B, has the radius of the pallet
pivot circle 1/3rd of the 71.21 mm escapement wheel circle
and the pallets act on this circle. The centre-line is in a mid-
tooth position. The pallet lines, torque circles, pendulum
amplitude lines and arcs are drawn similarly to drawing A.
There are a further three small circles drawn from the pallet
pivot circle centre. A perpendicular line has been drawn from
the extended line of the entry pallet in the pendulum extreme
left position to the centre of the pallet frame circle. This line
has been calibrated with units of the pitch chord. The
148
distance from the innermost circle to the torque circle of the
entry pallet in the extreme up position is three units on this
line and the distance to the torque circle of the exit pallet in
the extreme up position is two units. The inner circles
represent the forces exerted by the crutch and the pallet frame
mass.
But as here, to come a little nearer in this material Point, let, as I order
the Matter, the Force [from the Wheel] upon the Pendulum, as just
before the interchanging of the Pallats, to be as by or from them the said
Pallats supposed or taken as 3, then, as just after their interchanging,
[and the Force to contrary Direction], it must be about as 2, that is,
it must be so ordered , [as may hereafter be observed by the drawing]
viz. as that it be so by the taking, or supposing for the Purpose, a Mean
betwixt the Actions of each Pallat,
John Harrison, Concerning Such Mechanisms, 1775
The drawings C and D have a pallet pivot circle radius
1/6th of the 3 inch escapement circle. The centre line is 1/4
before and 1/4 behind the mid tooth position in C and D
respectively to show the effect on the angle of inclination of
the pallet frame by altering the length of the exit pallet arm.
The entry pallet lines are not drawn because they do not fulfil
the condition of being the same length in both positions.
This drawing demonstrates that the escapement is
asymmetrical with the pallet arms pivoted off-centre, of such
length that they can both be in contact with escapement
wheel teeth at equal amplitudes of the pendulum. Torque
circles have been drawn and calibrated in units of the pitch of
the escapement wheel. Escapement wheel teeth should
ideally be equidistant to the midline at each extreme
amplitude.
Escapement Drawing for RAS Regulator
If we draw the escapement of the RAS Regulator in
the same way as Harrison in his 1730 manuscript and 1775
pamphlet we can correlate the dimensions of the escapement
in its present state. We know that at least one of the pallets
is not original and therefore need to confirm that the present
measurements are similar to what Harrison intended. The
pallet frame and its arbor are the only parts of the
escapement that can be considered absolutely original. We
149
have an accurate measurement of 23.09 mm. for the pallet
arbor circle radius of the RAS Regulator and can derive the
pallet arm lengths spanning any number of teeth. The
distance between the pallet frame arbor and the escapement
wheel arbor is 79.47 mm. according to the previously
described geometrical layout of the plates based on Goulds
measurements. Gould measured the outside radius of the
escapement wheel as 2.66 inches (67.56 mm) giving a
diameter 5.32 inches (135.13 mm.) but the escapement wheel
in the RAS regulator at present has a radius of 71.44 mm.
As first, the bare Length of a Pendulum can be no otherwise
rightly considered or esteemed, but as only to what it bears, or may bear
in Proportion to the Length of the Pallats, i.e. in other Words to the
equivalent Distance from its Centre of Motion, to where the Pallats,
according to their Construction, as in the common Way touch, or are
applied to the Wheel; the Pendulum, as with respect to the length of the
Pallats, and as here in the first Place to be notified, being no more than
as about 14 to 1, so the which in Effect is no better, or can in Power
be no better, than as a Long Pendulum rendered a short one.
John Harrison, Concerning Such Mechanisms, 1775
Harrison was a proportional geometrician using
Newtonian mechanics. The present escapement wheel
diameter does not accord with the measurement recorded in
Goulds notebook of the measurement (67.56 mm.)or
proportionally to the measurement of the distance on the
pallet frame between the pallet arbor and the pallet frame
arbor. Careful analysis of Cottinghams escapement
photograph confirms Goulds measurement. Harrison
considered the length of a one second pendulum to be 39
inches (996.95 mm.) as taught by Saunderson. This
pendulum length divided by 14 gives 71.21 mm. and divided
by 14 gives 68.76 mm. It is considered that Gould replaced
the escapement wheel when he replaced the entry pallet with
one of a different shape and replaced the crutch. Gould
designed and fitted a Berthoud style stop mechanism acting
on the escapement wheel.
The escapement has been drawn having proportional
relationships to the distance between the escapement wheel
axis and the pallet frame knife edge in the RAS regulator, that
is, an escapement wheel of 68.12 mm. , pallet arbor circle
diameter 22.71 mm. (68.12/3) and distance of pallet frame
arbor above apex of escapement wheel of 11.36 mm.
150
(68.12/6). The distance between the escapement axis and the
pallet frame axis is 79.48 mm. (R+R/6, 68.12 + 11.36).
The escapement wheel has 120 teeth and the centre line is
equidistant between two teeth. The pallet lengths, torque
circles and pendulum arc are drawn. The pallets span 22/23
teeth, with arm lengths of 51.72 mm. and 33.94 mm. The
torque circles show that the torque applied at the mid point
of the pendulums swing by each pallet is equal. The
pendulum arc is 9.33E. The pallet lengths as suggested by
these drawings have been confirmed by calculation.
as I order the Matter, the Force [from the Wheel] upon the
Pendulum, as just before the interchanging of the Pallats, to be as by or
from them the said Pallats supposed or taken as 3, then, as just after
their interchanging, [and the Force to contrary Direction,] it must be
about as 2,
John Harrison, Concerning Such Mechanisms, 1775
The force acting on the pendulum consists of the
force from the escapement wheel as Harrison describes and in
addition that due to the masses of the pallet frame and the
crutch. Ideal masses for the pallet frame and crutch have been
added for three escapement dimensions. It is found that an
equal force exerted by both pallets onto the pendulum with
the pallet frame raised is most easily achieved using an
escapement wheel having a radius of 68.12 mm. with pallets
spanning 23 teeth.
Action of the Escapement.
Because of its asymmetry, when the entry pallet lifts
the pallet frame the distance between the pallets is diminished
(scissors action) and the pallet arbor is moved to the left.
Conversely when the exit pallet has lowered the pallet arbor
the distance between the pallets is increased and the pallet
arbor moved to the right. This means that the supplementary
arc at the interchange of the pallets with the pallet frame in
both positions is limited and recoil minimal. The
momentum of the pendulum at the extremity of its arc is
expended in producing recoil.
Referring to diagram A, clockwise movement of the
escapement wheel, assisted by the pendulum, lifts the pallet
frame. This results in the pallet arm arbor shifting gradually
to the right with respect to the escapement wheel centre.
This is more clearly seen as a large shift to the right of the
151
perpendicular positions of the pallet arbor. As the pallet
frame tilts the right composer places the right pallet in
position to meet the approaching escapement wheel tooth and
the radial movement causes the right pallet and a tooth to
make contact. As the pendulum moves to its maximal
position on the left with nearly zero velocity the pallet frame
has been lifted and the exit pallet composer has lowered the
exit pallet onto an escapement wheel tooth. The escapement
is now locked with both pallets in contact with the escapement
wheel. Further movement of the escapement wheel in a
forward direction increases this locking of the pallets. This
locking is released by the momentum of the pendulum forcing
retrograde movement of the scape wheel allowing the entry
pallet to move away from the escapement wheel. With the exit
pallet in contact with a tooth further clockwise movement of
the escapement wheel pulls the pallet frame allowing the heavy
side to fall with the added assistance of gravity and thereby
impulse the pendulum.
In diagram B, the pallet frame has descended and the
pendulum is on the right. Both pallets are again in contact
with the scape wheel. Further forward motion of the
escapement wheel is prevented by the entry pallet. The
pendulum can continue in an accessory arc and the weight of
the pallet frame will produce slight retrograde movement of
the escapement wheel. Because the pallet arms are
counterbalanced with lead buttons as soon as the pressure
maintaining contact is released the exit pallet will lift away
from the escapement wheel tooth because a composer is not
holding it in position.
For it may be notified, as just here hinted, that the Actions of
each Pallat are not equally the same upon the Pendulum, but not so, as
to be easily perceived to be otherwise.
John Harrison, Concerning Such Mechanisms, 1775
It is difficult to determine whether Harrison was
referring to qualitative or quantitative differences in the
action of the pallets in his statement above.
But farther, as this retrogradation of the Pendulum Wheel
is much less than is required to be in other Clocks, and the larger the
Vibration of the Pendulum is, the less proportion this small
retrogradation bears to the Vibration, and the Contrivance.
John Harrison, Concerning Such Mechanisms, 1775
The torque applied by the escapement to the exit
152
pallet is arranged to be less so that the fall of the pallet frame
is dominated by gravity.
The impulse given to the pendulum by the exit pallet
and gravity acting on the left side of the pallet frame may
mean that it reaches zero velocity further to the right than the
position where both pallets are in contact with the escapement
wheel. It is free to do this unhindered because it is the
increasing divergence of the pallets that causes the exit pallet
to leave the escapement wheel; the momentum of the
pendulum in addition to the torque derived from the
escapement is required to lift the left side of the pallet frame.
The slight recoil occurring when the pendulum has
moved to the left controls the amplitude of the pendulum arc
closely and limits any change caused by ambient conditions.
The torque transferred by the pallet arms is controlled by the
remontoire. It is arranged to be insufficient to raise the pallet
arm without the aid of the momentum of the pendulum but
allows the recoil. A significant proportion of the impulse
given by the exit pallet is the constant force of gravity acting
on the pallet frame. The efficiency of these actions is realised
by the angle of action of the pallet arms. Mudge, Cummings,
Airy and Cottingham studied Harrisons work and each
designed a gravity escapement. By accurately limiting the left
amplitude of the pendulum, Harrison was able to reduce the
error caused by changes in amplitude produced by variation in
air viscosity.
and the which Interchanging not being, or to be far from the
extremity of each Vibration, and in which little, or as it were overplus
Part, a necessary recoiling of the Wheel to be, viz. as not only for the
interchanging of the Pallats, the which, as according to my
Construction of the same is required, but as withal to have in some
Measure to do with the Effect of a Cycloid, but not to be the whole
Concern in that Matter.
John Harrison, Concerning Such Mechanisms, 1775
The development of the escapements in Harrisons
clocks from his earliest of 1713 until his RAS regulator
demonstrates his logical reasoning. The changes he made to
the escapement of the Brocklesby Hall clock was the most
important stage in the development. His pamphlets and
manuscripts give no hints to his reasoning behind the
progression from the Brocklesby Park modification to an
asymmetrical contrivance in the 1728 clocks and RAS
153
regulator.
The asymmetrical escapement functions as described
above with an asymmetrical crutch. There is no written
evidence that he did do this and there are no extant artifacts
giving any absolute indication. The pendulum drawing in the
1730 manuscript has a plain rod and the Cottingham
photographs show a plain crutch pin. The Guildhall
pendulum is doubtful evidence because the upper temperature
compensating rods are pinned differently than the lower and
the temperature compensation does not conform to
Harrisons requirements in his 1730 manuscript. The pendula
fitted to the other 1728 clocks have been fitted with modern
copies of the Guildhall pendulum. The pendulum fitted at
present to the RAS regulator was made when Gould restored
it. Confirmation of this hypothesis has been made by trials of
a copy of the regulator escapement, which has been shown to
be extremely reliable, efficient and accurate.
There is a description and drawing of a Harrison
escapement in Reess Cyclopaedia which perhaps
supplements the above and maybe a record of the antagonism
preventing its further use by contemporary and subsequent
makers of regulators.
It must be observed however, that the recoil, such as it is, takes
place at the extremities of the vibration, when the momentum of the
pendulum is almost exhausted and is therefore the most unfavourable
part of the vibration to occur in; notwithstanding which disadvantage
the escapement has performed as well as any other that had preceded it,
perhaps as well as any that has followed it, as the regulator to which it
was applied has seldom if ever been excelled in steadiness and accuracy
of performance for fourteen years of uninterrupted succession. The
imperfections arising from the recoil taking place at the extremities of
the vibrations seems to have been compensated by the immense
momentum of the pendulum that this escapement admitted of with a
comparatively small maintaining power, whilst the absence of friction
and of oil gave it the advantage of permanently uniform action.
Of course the pendulum was of the grid-iron pattern, that
compensated the effects of varying temperature, and allowed the
escapement to have its merit appreciated. The report of the performance
of Harrisons regulator with the escapement in question was, that it did
not vary one second per day from any preceding or following day for the
space of fourteen years, and that the aggregate of the variations from
true mean time did not amount to half a minute.
Cyclopaedia, 1820, ed. Abraham Rees
154
9
.
6
9

9
.
6
9

9
.
6
9

9
.
6
9

E
s
c
a
p
e
m
e
n
t
w
h
e
e
l
l
i
f
t
i
n
g
p
a
l
l
e
t
a
r
m
a
s
s
i
s
t
e
d
b
y
m
o
m
e
n
t
u
m
o
f
p
e
n
d
u
l
u
m
E
n
t
r
y
p
a
l
l
e
t
r
e
l
e
a
s
e
d
.
P
e
n
d
u
l
u
m
i
m
p
u
l
s
e
d
b
y
f
a
l
l
i
n
g
p
a
l
l
e
t
a
r
m
,
u
n
d
e
r
i
n
f
l
u
e
n
c
e
o
f
g
r
a
v
i
t
y
a
n
d
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
.
B
o
t
h
p
a
l
l
e
t
s
i
n
c
o
n
t
a
c
t
w
i
t
h
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
.
P
e
n
d
u
l
u
m
c
o
n
t
i
n
u
e
s
i
n
a
c
c
e
s
s
o
r
y
a
r
c
.
R
e
v
e
r
s
e
d
i
r
e
c
t
i
o
n
o
f
p
e
n
d
u
l
u
m
u
n
l
o
c
k
s
p
a
l
l
e
t
s
r
e
l
e
a
s
i
n
g
e
x
i
t
p
a
l
l
e
t
A
B
C
D
B
o
t
h
p
a
l
l
e
t
s
i
n
c
o
n
t
a
c
t
w
i
t
h
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
d
u
e
t
o
'
s
c
i
s
s
o
r
s
'
a
c
t
i
o
n
.
R
e
m
a
i
n
i
n
g
m
o
m
e
n
t
u
m
o
f
p
e
n
d
u
l
u
m
p
r
o
d
u
c
e
s
s
l
i
g
h
t
r
e
c
o
i
l
r
e
l
e
a
s
i
n
g
e
n
t
r
y
p
a
l
l
e
t
.
9.33
77.17 mm
9.33
0.97 mm
74.20 mm
R68.12 mm
3.57 mm
Harrison's Asymmetrical Escapement Action
Pendulum bob on extreme left. Pallet pivot in highest position.
Pendulum bob on extreme right. Pallet pivot in lowest position.
Note particularly differing angular movement of the pallets on release.
Determining Pallet Arm Lengths
Let A represent two bits of Glass, whereon each end of the
Axis of the Pendulum being a Sharp Edge of Brass moves in a little
notch in the Glass, which renders that Friction insensible, as also the
Wearing (for tis but the Axis of the Pallats, which communicates force
to the Pendulum; for the Pendulum it self is otherwise suspended, viz
by a thin Brass Spring). Let the Arch CD, and the least Vibration the
Pendulum must make to let the Wheel move forward, be similar;
therefore the Arch a.a. is described with the same Vibration by the
Center whereon the Pallats have their relative Motion, and from the
extremities of the said Arch or Vibration a.a. and with the length of
each Pallat may be described the portions of Circles ee.ff. which must
cut the Limb of the Wheel so, that there be half the space of one of the
Teeth between; because one Vibration lets but the space of half a Tooth
go.
John Harrison, 1730 Document
Harrison determined the length of the pallet arms
geometrically so that at each extremity of the pendulums
amplitude both of the pallets are in contact with escapement
teeth. Unfortunately he does not describe his method in detail
so that we have to find a method of drawing which will fulfil the
conditions that he defined. If the geometrical method is correct
then it can be confirmed by algebraic analysis. A determination
for the RAS regulator can then be compared to the dimensions
of the pallets in their present form as we know that at least one
pallet arm is not original. Also careful analysis of the results
will give further indication of Harrisons design intentions. My
drawing for the RAS regulator is based on measurements, made
separately by Gould and Betts, of the pallet frame circle, the
diameter of the escapement wheel and the distance between the
escapement wheel arbor and pallet frame arbor.
Geometrical Construction
1. Draw a circle of radius equal to escapement wheel and
mark centre, 0. In his 1730 document Harrison drew
a circle 3 inches radius. In the Guildhall manuscript
and in the RAS regulator the radius is a proportion of
the length of the pendulum (either 1/14 or 1/14,5
157
respectively).
2. Divide circumference of circle into double number of
escapement wheel teeth divisions.
To divide the circle into sixty teeth it is first
necessary to divide the circumference by ten and then
divide by six commencing at each of the ten division.
Each of these sixty divisions can then be divided by two
or four to give circles of one hundred and twenty or two
hundred and forty divisions.
Draw a horizontal line through centre, O. Mark where
this line cuts the circumference of the circle M and N.
Erect a perpendicular through O, cutting the
circumference at A by drawing equidistant arcs from M
and N. Bisect MO at P and with centre P and radius
AP, describe an arc to cut NO at Q. OQ equals the side
AF of the regular decagon within the circle.
158
The circumference is divided into a regular hexagon by
chords equal to its radius. By starting the division of
the circumference at each point of the decagon the circle is
divided into sixty division. Three sets of division are shown in
the diagram starting at 1, 2, and 3.
Each of the sixty divisions can be bisected once or twice.
The perpendicular line is at the half tooth position and
it is necessary to divide the circle with twice as many
divisions as escapement wheel teeth
3. Mark the centre of the pallet frame circle. In the 1730
document this is 1/3 of the radius of the escapement
circle above the apex of the circle, or 4R/3 above the
centre of the escapement circle. In the Guildhall
document and the RAS regulator it is 1/6 of the radius
of the escapement circle above the apex of the circle, or
7R/6 above the centre of the escapement circle.
4. Draw the pallet frame circle. For the Guildhall
document and the RAS regulator this has a radius one
third of the escapement wheel. For the 1730 document
the pallet frame circle passes through the exit pallet
159
tooth position with the pendulum in extreme right
amplitude.
5. The line representing the length of the exit pallet arm
for all the three escapements passes through the apex
intersection of the escapement circle and the midline.
For the 1730 MS drawing and the Guildhall document
draw the line through this point to the chosen position
of the exit pallet and extend the line to the left to cut
the pallet frame circle. For the RAS regulator draw a
line from apex point of the escapement wheel to the
tangent point on the pallet frame circle of a tangent to
this circle which passed through the escapement wheel
centre and extend this line to the right to cut the
escapement wheel circle.
6. Draw arcs of length determined in para. 5 from each exit
pallet position to cut the pallet frame circle. Draw arcs
from these points to determine where they intersect.
7. Draw a tangent to the escapement circle from the lower
entry pallet point to determine the entry escapement
wheel teeth. Draw lines from the points on the pallet
frame circle to the relevant escapement wheel teeth.
These lines should be the same length. Draw arcs from
the points on the pallet frame circle from the entry
pallet escapement wheel teeth to determine their
intersect point.
8. Draw extended lines from the extreme points on the
pallet circle through its centre to determine the angular
movement of the crutch.
9. Draw circles tangential to the extended pallet lines to
complete the construction.
Drawing RAS Escapement
To fit distance from escapement wheel axis and top of frame of
79.47 mm. and to be compatible with Goulds measurement
radius of escapement wheel of 2.66" 67.56 mm.
1. Draw perpendicular line and circle 68.12 mm. radius
(R)centred on line. Divide circle into 120 divisions
equidistant to centre line.
160
2. Draw circle 22.71 mm, radius (R/3) centred on vertical
line 79.47 mm. (R+R/6) above centre of primary
circle.
3. Draw tangent from circumference of secondary circle to
centre of primary circle and mark tangent point A.
4. Draw line from tangent point through apex of primary
circle to cut circumference at B.
5. Draw a tangent on primary circle to tangent point A
marked on secondary circle and mark the tangent point
C. The angular distance between this tangent point C
and the intersection of the circumference B is 68.34,
that is nearly 23 divisions. Line AB is 51.99 mm. and
line AC is 34.05 mm.
6. Using dividers find the distance between the tangent
point on the secondary circle and the circumference of
primary circle such that points on the circumference of
the secondary circle struck from the 8 and 8 tooth
positions on the right are also equidistant from the 14
and 14 tooth positions on the left. These distances
are 51.72 mm. and 33.94 mm. respectively.
7. Draw arcs from these points from their respective tooth
position as Harrison has. A line joining the
intersections of these arcs passes through the centre of
the secondary circle.
8. Draw the torque circles. The torque circles show that
the torque applied at the mid point of the pendulums
swing by each pallet is equal. The pendulum arc is
9.33E.
B
A
C
68.34
R22.71 mm
R68.12 mm
51.99 mm
34.05 mm
79.47 mm
161
Algebraic Calculation
The Fixed Parameters are the escapement wheel radius (R) and
number of escapement wheel teeth (N).
Conditions
1. Both pallets must be in contact with an escapement
wheel tooth at both extreme amplitudes.
2. The line from the exit pallet L, passes through the apex
of the escapement wheel when the pendulum is at the
extreme amplitude on the right and the pallet pivot is in
its lowest position.
3. A half tooth position of the escapement wheel is on the
centre line.
4. The lower entry escapement wheel tooth position u,v is
the tangent point on the escapement circle of entry
pallet line, l.
5. The pallet circle radius and distance between the centres
of the escapement circle are different for the three
escapements.
The calculations can be made with a spreadsheet
(Lotus 1-2-3, Microsoft Excel or Corel Quatro Pro) or
MathCad. They are illustrated on a spreadsheet as this
software is more generally available and understood. The exact
length of the pallet arms, L and l, are calculated by iteration
and then used to obtain the other parameters.
162
2
2
1
2
1
2
(
a
a
)
(
b
b
)
2
r

u
=
2
2
t
P
y
z
x
t
t
a
r
c
t
a
n
x
u
=

=
+
=
2
2
2
L
z
r
a
r
c
c
o
s
2
L
z

=
P
A
l
g
e
b
r
a
i
c

C
a
l
c
u
l
a
t
i
o
n

o
f

P
a
l
l
e
t

L
e
n
g
t
h
s
.
R
7
R
r
P
3
6
=
=
2
2
B
a
b
=
+
2
2
2
2
l
(
a
u
)
(
v
b
)
L
(
x
a
)
(
y
b
)
=

a
L
s
i
n
(
)
x
b
L
c
o
s
(
)
y
u

+
=

+
2
2
d
u
(
P
v
)
b
v
a
r
c
t
a
n
a
u
P
v
a
r
c
t
a
n
u
T
d
s
i
n
(
)
A
z
s
i
n
k
k

=
+

= =

=
u
R
s
i
n
(
)
v
R
c
o
s
(
)
|
o
|
o
=
+
=
+
2
2
2
2
2
2
L
B
R
a
r
c
c
o
s
2
L
B
l
B
R
a
r
c
c
o
s
2
l
B

o
+

=
+

=
B

b
x
R
s
i
n
y
R
c
o
s
o
o
= =
R
R
d
a
E
s
c
a
p
e
m
e
n
t

C
a
l
c
u
l
a
t
i
o
n

D
i
a
g
r
a
m

A
.
t
c
w
r
x
q
-
g
g
b
-
v
u
,
v
a
,
b
r
s
d
L
0
,
P
T
q
z
t x
,
y
a
-
u
k
l
R
0
,
0

The sequence of calculations is numbered on the first
column of the sample spreadsheets.
1. Enter the escapement wheel radius, R, and the
number of escapement teeth, N.
2. Calculate or enter radius of pallet wheel, r. (R/3 for
Guildhall MS and RAS regulator). Calculate or enter
the distance between the centres of escapement wheel
circle and pallet frame circle (P). (4R/3 for 1730
document, 7R/6 for the Guildhall MS and RAS
regulator)
3. Calculate position of exit pallet, x,y
For the RAS regulator the position of the pallet arm
pivot a,b is calculated. The gradient of the exit pallet
line passing from this point through the apex of the
escapement is determined to enable the exit pallet
position x,y to be calculated.
Length of tangential line, B
2 2
B P -r . =
Angle between tangential line and centre-line, (.
r
arcsin
P
g =
Cartesian co-ordinates, a,b
a Bcos
b Bsin
= g
= g
Equation of exit pallet line

y mx c
b R
y .x R
a
= +
-
= +
Equation of circle

2 2 2
x y R + =
The position where the exit pallet line crosses the
escapement wheel circle is found by solving the
simultaneous equations for the line and circle.

2
-b- R
2 2
x .x R R
a


+ + =

164
( )
-b- R 2
2 2
x -mx R R where - m
a
+ + = =

2 2 2 2 2
x m x - 2mRx R R + + =

( )
2 2
1 m x 2mRx 0 + - =
Using quadratic formula
2
-b b - 4ac
x
2a

=

( )
2mR 2mR
x
2
2 1 m
+
=
+

( )
4mR 2mR
x 0 or
2 2
1 m 2 1 m
= =
+ +
y - mx R = +
For the 1730 document and Guildhall MS the co-
ordinates of the exit pallet, x,y are determined for the number
of teeth spanned by the pallet to the right of the centreline.
The gradient of the exit pallet line passing from this point
through the apex of the escapement is determined to enable
the pallet arm pivot position a,b to be calculated.
Enter number of teeth between exit pallet and centreline, N
rt
,
to determine angle, ,

360
.N
rt
N
=
Calculate x,y

x Rsin
y Rcos

=
=
x Rsin
y Rcos

=
=
For the 1730 document the radius of the pallet frame circle
can then be calculated.
( )
2 2
r P y x = +
Calculate the position of the pallet arm pivot, a,b for the
1730 document and the Guildhall MS.
Determine equation of exit pallet line

y-R
b .a R
x
= +
Equation of pallet frame circle
2 2 2
a b r + =
165
Solve simultaneous equations to find intercept points
( )
( )
2
0
y R
2
2 2
a ga p r where g
x
and p P R
2 2 2 2 2
a g a 2gpa p r
2 2 2
1 g a 2gpa p r =

+ + = =
=
+ + =
+ +
By quadratic formula

( )
( )( )
( )
2
2 2 2
2gp 2gp 4 p r 1 g
a
2
2 1 g
+
=
+
y-R
b .a R
x
= +
4. Calculate estimated length of exit pallet arm, L.

( ) ( )
2 2
L x-a y b = +
5. Determine the maximum number of teeth spanned by
the pallets and the the number of teeth on the right
hand side of the centreline.
The three angles , and are calculated from their
respective right angled triangles.
x
arcsin
R
r
arcsin
P
R
arcsin
B

=
=
=
Maximum number of teeth spanned by pallets, n,
( )N
n
360
+ +
=
Number of teeth on right of centreline, m,
.N
m
360

=
6. The positions of the pallets, x,y and u,v at the
extreme amplitudes of the pendulum can be calculated
166
360
(m 0.5)
1
N
360
. m
2
N

=
=

360
(n 0.5)
1 1
N
360
.n
2 2
N


+ =
+ =
x R. sin
y R. cos
u R. sin( )
v R.cos( )

=
=
= +
= +
7. Calculate the positions of the pallet pivot, a,b for the
estimated length, L,
2 2
z x (P y)
(P y)
a tan
x
2 2 2
L z r
arccos
2Lz
a x Lcos( )
b Lsin( )

= +

=
+
=
=
=
8. Calculate the lengths of the short pallet arm at each
extreme amplitude and find difference
2 2
l (a u) (v b)
l l l
1 2
= +
=
9. Alter length of long pallet arm, L, by a multiple of l
and recalculate until the error is acceptable.
L L 1.4 l
Corr
=
10. Calculate angular movement of pallet frame, ,
2 2
(a a ) (b b )
1 2 1 2
sin( )
2 2r
2 2
(a a ) (b b )
1 2 1 2
2. arcsin
2r

+
=
+
=
11. Calculate the radius of the entry and exit pallet
torque circles, T and A.
167
For the entry pallets,
2 2
d u (P v)
b v
arctan
abs(a u)

= +

P v
arctan
abs(u)
T dsin( )

=
=
For the exit pallets
2 2
z x (P y) = +
2 2 2
L x r
arccos
2Lz
A zsin

+
=
=
12. Determine torque units, f, and the offset, o.
T A
f
5
3(T A)
o T
5
+
=
+
=
13. Calculate pallet arm angles, and .

2 2 2 2
L a b R
arccos
2 2
2L a b
2 2 2 2
l a b R
arccos
2 2
2l a b

+ +
=
+
+ +
=
+
14. Determine lateral shift of pallet frame pivot, s, and
difference in angle between pivot arms at their upper
and lower positions, .

s a a
1 2
( ) ( )
1 1 2 2

=
= + +
The Royal Astronomical Regulator has had a stormy
history since it left Harrisons hands as an unfinished
pendulum clock, although it was almost certainly used by him
as an experimental tool in the conception of his coach clock,
H4, which fulfilled the accuracy conditions required by
Queen Annes 1714 Act of Longitude. Gould restored the
clock in 1927 but in so doing made major changes to it,
168
including fitting a new pallet arm of altered design.
At present the clock has an escapement wheel having
120 teeth and radius of 71.44 mm., a heavy brass English
style adjustable crutch, pallet arms of doubtful provenance
spanning 22/23 teeth, repaired composers and altered pallet
frame. The only parts that we can be certain were Harrisons
work are the pallet frame, composers, pallet frame arbor and
the knife edge fittings in the main frame of the clock.
If we wish to reconstruct as Harrison intended it is
necessary to understand how the escapement operated whilst
complying with the known critical dimensions in the RAS
regulator and other relevant data. We therefore have to
consider:-
Goulds measurement of the radius of the
escapement wheel of 2.66 inches (67.56
mm.) written on page 138 of his notebook
Pallet frame arbor axis to pallet arbor axis
measurement of 23.09 mm.
Escapement arbor axis to to top of main
plates measurement of 79.47 mm.
The ratio of the radius of the escapement
wheel to the length of the pendulum should
be 14 to 1 giving a radius of 68.76 mm.
(996.95/14.5) as described on page 4 of
Concerning Such Mechanism
Gould fitted stopping gear made by
R.J.Hopgood which acts on the escapement
wheel. The present escapement wheel was
most probably made with this stopping gear
accounting for the difference in radius
between it and Goulds measured radius.
Goulds replacement of pinning of pallet
frame to knife edge arbor with two pan-head
screws enabling adjustment of the inclination
of pallet frame.
The dimensions of the pallet arms at present
assuming that replacements were made to the
original dimensions.
Cottinghams 1909 photograph of pallet
frame and crutch.
169
1730 Drawing Figure 2 Reversed.tcw
The drawing in Harrison's 1730 document has been reproduced with an escapement wheel diameter of 3 inches
and reversed so that the escapement wheel is shown rotating clockwise as in his MS 3972/3 drawing. Torque
circles have been added. Note the care which Harrison has taken in drawing the angle of the contact face of the
escapement wheel.
R
9
.
5
0

m
m
9
.
5
4

R
1
4
.
1
9

m
m
R
3
.
9
7

m
m
R
3
8
.
1
0

m
m
R24.71 mm
R
2
0
.
4
5

m
m
5
0
.
8
0

m
m
21
5
7

R
3
2
.2
0
m
m
0 10 20 30 40 50
Scale 1:1
Determining Pallet Arm Lengths for 60 Tooth Escapement Wheel ( 1730 Ms)
Numbers in first column are the paragraph numbers of the preceding description. These calculations derive the pallet lengths, angle of escaping arc, radius of
circles and analysis for an escapement where the both pallets touch the escapement wheel at the extreme amplitudes, wheel and the distance between the ce
of the pallet frame and escapement wheel is 4R/3. The pallet frame radius passes throught the exit pallet position.
F E D C B A
60 1 Enter N, number of escapement wheel teeth N 1
38.10 2 Enter R. escapement wheel Radius R
50.80 4 Distance between centres +4*R/3 P
3.5 3 Enter m, maximum number of teeth on right of midline m
21 5 +360*P/N "
13.65 6 +R*@sin("*@pi/180) x 3
35.57 7 +R*@cos("*@pi/180) y
20.45 8 Pallet circle radius @sqrt((+P-y)*(P-y)+x*x) r 2
-0.19 9 Gradient of line (y-R)/x g
12.70 10 +P-R p
-18.21 11 (+2*p*g-@sqrt(+4*p*p*g*g-4*(1+g*g)*(p*p-r*r)))/(2*(1+g*g)) a
41.47 12 +a*m+R b
32.40 13 @sqrt((a-x)*(a-x)+(b-y)*(b-y)) L 4
45.29 14 @sqrt(a*a+b*b) A
23.70 15 @abs(@atan(a/b)*180/@pi) $
32.74 16 @acos(R/A)*180/@pi (
9.41 17 ($+g)calc*N/360 ncalc 5
9.5
19
@if(+ncalc-@int(ncal)<0.5,@int(ncal)+0.5,@int(ncalc+0.5)) nint

18.00
20
+360/N*(m-0.5) a1
21.00
21
+360*m/N a2
54.00
22
+360/N*(n-0.5) b1+d1
57.00
23
+360/N*n b2+d2
13.65 11.77
24
+R*@sin((+"*@pi)/180) x 6
35.57 36.24
25
+R*@cos((+"*@pi)/180) y
-31.95 -30.82
16
+R*@sin((+$*@pi)/180) u
20.75 22.39
27
+R*@cos(($*@pi)/180) v
15.23 14.56
28
P-y t
20.45 18.73
29
@sqrt((x*x)+(t*t)) z
32.20 32.19 32.40 30 Estimated L A13 +B 42 +D42 L
48.12 51.05 31 @atan(t/x)*180/@pi 2
38.09 36.54 38.12 36.57 37.62 35.97 32 @acos(((L*L)+(z*z)-(r*r))/(2*L*z))*180/@pi (
10.03 14.51 10.01 14.48 10.50 15.08 33 2-( 2-(
-18.05 -19.40 -18.04 -19.39 -18.21 -19.51 34 +x-(L*@cos((2-()*@pi/180)) a 7
41.18 44.30 41.16 44.29 41.47 44.66 35 (L*@sin((2-()*@pi/180))+y b
-13.90 -11.43 -13.91 -11.43 -13.75 -11.31 36 a-u a-u
20.43 21.91 20.41 21.89 20.72 22.27 37 b-v b-v
24.71 24.71 24.70 24.70 24.87 24.98 38 @sqrt(((a-u)*(a-u))+((b-v)*(b-v)) l 8
0.0002 -0.005 0.11 39 l1-l2 l1-l2
1.35 1.35 1.31 40 a1-a2 a1-a2
3.13 3.12 3.19 41 b1-b2 b1-b2
32.20 32.20 32.19 42 @if(@abs(l1-l2)<=0.00001,L,L-1.4*(l 1-l2))) Iterate 9
9.54 43 @asin(@sqrt((a1-a2)*(a1-a2))+(b1-b2)*(b1-b2))/(2*r))*(360/@pi) M 10
43.86 41.92 44 @sqrt((u*u)+((P-v)*(P-v))) d
55.76 62.45 45 @atan(b-v/@abs(a-u))*180/@pi 6
43.24 42.66 46 @atan((P-v)/@abs(u))*180/@pi 8
9.51 14.19 47 +d*@sin((6-8)*@pi/180) T 11
12.49 11.00 48 +z*@sin((*@pi/180) A
-0.03 4.74 49 (T+A)/5 T-3(T+A)/5 with pendulum at right f1 12
-1.61 4.70 50 (T+A)/5 T-3(T+A)/5 with pendulum at left f2
56.30 51.85 51 @acos(L
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi D 13
57.91 51.19 52 @acos(l
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi F
4.45 1.35 53 a1-a2 (D1+F1)-(D2+F2) s : 14
15.00
3
9
.0
0

8
2
.
9
7

m
m
9
.
9
2

R
2
3
.
7
1

m
m
R
7
6
.
2
0

m
m
R
7
1
.
2
1

m
m
R
6
3
.
5
0

m
m
R29.59 mm
R
4
0
.
8
2

m
m
Harrison Escapement Drawing
Harrison MS Escapement.tcw/P
Determining Pallet Arm Lengths for 120 Tooth Escapement Wheel (MS 3972/3)
Numbers in first column are the paragraph numbers of the preceding description. These calculations derive the pallet lengths, angle of escaping arc, radius of torque
circles and analysis for an escapement where the both pallets touch the escapement wheel at the extreme amplitudes, the pallet frame radius is 1/6 radius of escapement
wheel and the distance between the centres of the pallet frame and escapement wheel is 7R/6.
F E D C B A
120 1 Enter N, number of escapement wheel teeth N 1
71.21 2 Enter R. escapement wheel Radius R
83.08 3 Distance between centres +7*R/6 $A$2*7/6 P
5 4 Enter m, maximum number of teeth on right of midline m
15 5 +360*P/N "
18.43 6 +R*@sin("*@pi/180) x 3
68.78 7 +R*@cos("*@pi/180) y
23.74 8 Pallet circle radius, R/3 +$A$2/3 r 2
-0.13 9 Gradient of line (y-R)/x g
11.87 10 +P-R p
-21.97 11 (+2*p*g-@sqrt(+4*p*p*g*g-4*(1+g*g)*(p*p-r*r)))/(2*(1+g*g)) a
74.10 12 +a*m+R b
40.75 13 @sqrt((a-x)*(a-x)+(b-y)*(b-y)) L 4
77.29 14 @sqrt(a*a+b*b) A
16.52 15 @abs(@atan(a/b)*180/@pi) $
22.88 16 @acos(R/A)*180/@pi (
13.13 17 ($+g)calc *N/360 ncalc 5
13.0 @if(+ncalc-@int(ncal)<0.5,@int(ncal)+0.5,@int(ncalc+0.5)) nint
13.50
20
+360/N*(m-0.5) a1
15.00
21
+360*m/N a2
37.50
22
+360/N*(n-0.5) b1+d1
39.00
23
+360/N*n b2+d2
18.43 16.62
24
+R*@sin((+"*@pi)/180) x 6
68.78 69.24
25
+R*@cos((+"*@pi)/180) y
-44.81 -43.35
16
+R*@sin((+$*@pi)/180) u
55.34 56.49
27
+R*@cos(($*@pi)/180) v
14.29 13.84
28
P-y t
23.32 21.63
29
@sqrt((x*x)+(t*t)) z
40.87 40.88 40.75 30 Estimated L A13 +B42 +D42 L
37.80 39.77 31 @atan(t/x)*180/@pi 2
30.01 27.04 29.98 27.00 30.30 27.39 32 @acos(((L*L)+(z*z)-(r*r))/(2*L*z))*180/@pi (
7.79 12.73 7.82 12.77 7.50 12.38 33 2-( 2-(
-22.06 -23.24 -22.07 -23.25 -21.97 -23.18 34 +x-(L*@cos((2-()*@pi/180)) a 7
74.32 78.25 74.34 78.28 74.10 77.98 35 (L*@sin((2-()*@pi/180))+y b
-22.75 -20.11 -22.74 -20.10 -22.84 -20.17 36 a-u a-u
18.98 21.76 19.00 21.78 18.76 21.48 37 b-v b-v
29.63 29.63 29.64 29.64 29.56 29.47 38 @sqrt(((a-u)*(a-u))+((b-v)*(b-v)) l 8
-0.0032 0.006 -0.09 39 l1-l2 l1-l2
1.179 1.176 1.21 40 a1-a2 a1-a2
3.930 3.936 3.88 41 b1-b2 b1-b2
40.87 40.87 40.88 42 @if(@abs(l1-l2)<=0.00001,L,L-1.4*(l1-l2))) Iterate 9
9.92 43 @asin(@sqrt((a1-a2)*(a1-a2))+(b1-b2)*(b1-b2))/(2*r))*(360/@pi) M 10
52.70 50.85 44 @sqrt((u*u)+((P-v)*(P-v))) d
39.84 47.25 45 @atan(b-v/@abs(a-u))*180/@pi 6
31.76 31.52 46 @atan((P-v)/@abs(u))*180/@pi 8
7.41 13.79 47 +d*@sin((6-8)*@pi/180) T 11
11.77 9.95 48 +z*@sin((*@pi/180) A
1.07 4.24 49 (T+A)/5 T-3(T+A)/5 with pendulum at right f1 12
-1.26 4.34 50 (T+A)/5 T-3(T+A)/5 with pendulum at left f2
65.68 60.73 51 @acos(L
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi D 13
66.70 59.29 52 @acos(l
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi F
4.96 1.18 53 a1-a2 (D1+F1)-(D2+F2) s : 14

Escapement Drawing RAS Regulator
9.33
R68.12 mm
R51.72 mm
R33.94 mm
R22.71 mm
79.47 mm
The tangent line from the escapement arbor to the pallet arbor circle, perpendicular to pallet frame arbor and line
from intersection to apex of escapementcircumference are drawn in blue. Note that the dimensions and geometry of
the RAS regulator have the both pallets in contact with the escapement wheel at the escaping arcs whilst a midtooth
position is in the perpendicular position.
Determining Pallet Arm Lengths for 120 Tooth Escapement Wheel ( RAS Regulator)
Spanning 23 teeth 68.12 mm. Radius
Numbers in first column are the paragraph numbers of the preceding description. These calculations derive the pallet lengths, angle of escaping arc, radius of torque
circles and analysis for an escapement where the both pallets touch the escapement wheel at the extreme amplitudes, the pallet frame radius is 1/6 radius of escapement
wheel and the distance between the centres of the pallet frame and escapement wheel is 7R/6.
F E D C B A
120 1 Enter N, number of escapement wheel teeth N 1
68.12 2 Enter R, radius of escapement wheel R
22.71 3 Pallet circle radius, R/3 +$A$3/3 r 2
79.47 4 Distance between centres R+r/6 +$A$2+$A$2/6 P
76.16 5 %(P
2
-r
2
) @sqrt(+A4*A4-A3*A3) B
16.60 6 @asin(R/B) @asin(A2/A5)*180/@pi *
21.76
7
B*sin d +A5*@sin(A6*@pi/180) aest
72.98
8
B*cos d +A5*@cos(A6*@pi/180) best
-0.22
9
Gradient of line b-R/-a (+A8-A2)/-A7 g
29.01
10
-mR/1+m
2
(-2*A9*A2)/(1+A9*A9) xest 3
61.63
11
mg + R +A9*A10+A2 yest
52.03
12
((x+a)
2
+(b-y)
2)
@sqrt((A10+A7)*(A10+A7)+(A8-A11)*(A8-A11)) Lest 4
34.06
13
(B
2
-R
2
) @sqrt($A$5*$A$5-$A$2*$A$2) lest
25.21
14
arcsin x/R @asin(A10/A2)*180/@pi aest
26.57
15
arccos R/B @acos(A2/A5)*180/@pi best
68.38
16
a+*+b +A6+A14+A15 a+*+b
23.00
17
Estimated number of teeth spanned @int((+A16*A1/360)+0.5) nest 5
0.40
18
a*N/360-@int(a*N/360)
8.50
19
No of teeth on right @if(A18<0.5,@int(A14*A2/360)+0.5,@int(A14*A2/360+0.5)) mest
24.0
20
+360/N*(m-0.5) a1
25.5
21
+360*m/N a2
42.0
22
+360/N*(n-0.5) b1+d1
43.5
23
+360/N*n b2+d2
29.33 27.71
24
+R*@sin((+"*@pi)/180) x 6
61.48 62.23
25
+R*@cos((+"*@pi)/180) y
-46.89 -45.58
26
+R*@sin((+$*@pi)/180) u
49.41 50.62
27
+R*@cos(($*@pi)/180) v
17.99 17.24
28
P-y t
34.40 32.63 30 @sqrt((x*x)+(t*t)) z
51.72 51.72 52.03
29
Estimated L A13 +B42 +D42 L
31.53 31.89 31 @atan(t/x)*180/@pi 2
20.04 17.20 20.04 17.21 19.48 16.48 32 @acos(((L*L)+(z*z)-(r*r))/(2*L*z))*180/@pi (
11.48 14.69 11.48 14.69 12.04 15.42 33 2-( 2-(
-21.36 -22.33 -21.36 -22.33 -21.56 -22.45 34 +x-(L*@cos((2-()*@pi/180)) a 7
71.78 75.34 71.78 75.34 72.33 76.06 35 (L*@sin((2-()*@pi/180))+y b
-25.52 -23.25 -25.53 -23.25 -25.33 -23.13 36 a-u a-u
22.37 24.72 22.37 24.72 22.92 25.44 37 b-v b-v
33.94 33.94 33.94 33.94 34.17 34.38 38 @sqrt(((a-u)*(a-u))+((b-v)*(b-v)) l 8
-0.00 -0.00 0.22 39 l1-l2 l1-l2
0.96 0.96 0.89 40 a1-a2 a1-a2
3.57 3.56 3.72 41 b1-b2 b1-b2
51.72 51.72 51.72 42 @if(@abs(l1-l2)<=0.00001,L,L-1.4*(l1-l2))) Iterate 9
9.33 43 @asin(@sqrt((a1-a2)*(a1-a2))+(b1-b2)*(b1-b2))/(2*r))*(360/@pi) M 10
55.70 53.94 44 @sqrt((u*u)+((P-v)*(P-v))) d
41.23 46.76 45 @atan(b-v/@abs(a-u))*180/@pi 6
32.66 32.33 46 @atan((P-v)/@abs(u))*180/@pi 8
8.30 13.44 47 +d*@sin((6-8)*@pi/180) T 11
11.79 9.65 48 +z*@sin((*@pi/180) A
0.40 4.35 49 (T+A)/5 T-3(T+A)/5 with pendulum at right f1 12
-1.08 4.29 50 (T+A)/5 T-3(T+A)/5 with pendulum at left f2
61.94 58.80 51 @acos(L
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi D 13
63.40 59.34 52 @acos(l
2
+a
2
+b
2
-R
2
/(2L%(a
2
+b
2
))*180/@pi F
3.14 0.96 53 a1-a2 (D1+F1)-(D2+F2) s : 14
Forces Applied to Pendulum
9.33
Pendulum Amplitude
A
B
C
A+B+C
A
B
C
A+B
The total force exerted by the entry and exit pallet should be equal when the pendulum is
at its maximum amplitude to the left and the mean force should be as equal and as near to
9.97 as possible.
Left
0
0
Right Right
T
A
a
t
a
t
B
-B
T-B
A+B
a+B
t-B
T A 2B
C
2
+
=
C
-C
C
a+B
t-B
T-B-C
A+B+C
T-B-C = A+B+C
a+B+C
t-B-C
t-B-C
a+B+C
T t
B
2
+
=
-
Forces Compared
Pendulum Amplitude
A+B+C
A
B
C
A+B
Left
8.37
10.97
13.57
-11.90
-9.74
0.82
-0.82
9.19
12.75
-11.08 -10.56
1.10
-1.10
10.29
-9.98
11.65
-11.66
-11.90
-10.82
8.37
-11.08
9.21
10.98
-10.82
1.10
Right Right
10.97
-9.98
-10.81
68.76 mm. Radius 22/23 Teeth Spanned
10.29
0
7.41
10.60
13.79
-11.77
-9.95
1.01
-1.01
8.42
12.78
-10.76 -10.96
0.91
-0.91
9.33
-9.85
11.86
-11.87
-11.77
-10.86
7.41
-10.76
8.42
10.60
-10.86
0.91
Right Right
10.60
-9.85
-10.86
MS 71.21 mm. Radius 17/18 Teeth Spanned
Left
9.33
Left
8.30
10.87
13.44
-11.79
-9.65
0.83
-0.83
9.13
12.62
-10.96 -10.48
1.07
-1.07
10.20
-9.89
11.55
-11.55
-11.79
-10.72
8.30
-10.96
9.13
10.87
-10.72
1.07
Right Right
10.97
-9.89
-10.72
RAS 68.12 mm. Radius 22/23 Teeth Spanned
10.20
0
9.51
11.85
14.19
-12.49 -11.00
0.85
10.36
13.34
-11.64
-11.85
0.75
11.11
-10.89
12.59
-12.60
-11.75
Right Right
1730 MS 11/12Teeth Spanned
Left
Pendulum Amplitude
A+B+C
A
B
C
A+B
-12.49
9.51
-0.85
-11.64
10.36
11.85
-11.75
0.75
-0.75
-10.89
11.11
11.85
-11.75
The pallet frame is the only part of the escapement
that can be considered absolutely original. We have accurate
measurements of 23.09 mm. for the pallet arbor circle radius
and 79.47 mm. for the distance between the axis of the
escapement wheel and the top of the main frame. The
pendulum length of 996.95 mm. divided by 14 gives 71.21
mm. and divided by 14 gives 68.76 mm.
Results for the 1728 wooden regulators, the
Guildhall drawing, for 68.76 mm. radius escapement wheels
having pallets spanning 17/18 and 22/23 teeth, and for a
68.12 mm. radius escapement wheel are tabulated giving
moments as radius of torque circles in mm.

1730 Guildhall MS
22/23 RAS
No. of Teeth
60 120 120 120
Esc. Radius R
38.1 71.21 68.76 68.12
r
20.07 23.70 22.92 22.71
P
50.8 83.06 80.22 79.47
Teeth Spanned
11/12 17/18 22/23 22/23
Pallet length L
32.2 40.87 52.21 51.72
l
24.71 29.43 34.26 33.94
Angle of Action
9.54 9.92 9.33 9.33
Moments T
14.19 13.79 13.57 13.44
A
11.00 9.95 9.74 9.65
t
12.49 11.77 11.90 11.79
a
9.51 7.41 8.37 8.30
But as here, to come a little nearer in this material Point, let, as I order
the Matter, the Force [from the Wheel] upon the Pendulum, as just
before the interchanging of the Pallats, to be as by or from them the said
Pallats supposed or taken as 3, then, as just after their interchanging,
178
[and the Force to contrary Direction], it must be about as 2, that is,
it must be so ordered , [as may hereafter be observed by the drawing]
viz. as that it be so by the taking, or supposing for the Purpose, a Mean
betwixt the Actions of each Pallat,
John Harrison, Concerning Such Mechanism, 1775
Careful study of Harrisons writing shows it to be as precise as his
draughtsmanship. In the above quotation he has added from the Wheel in
parenthesis when considering the forcing of the pendulum. The total forces
acting on the pendulum are the force from the escapement wheel plus the
forces from the masses of the pallet frame and crutch. Harrison indicated
these added forces by drawing the three inner torque circles on his diagram.
They have been graphically plotted for the four suggested solutions.
A strong proof indeed, that the Force or Draught of the Pendulum-
Wheel of my Clock, and as in the first Place to be understood, with its
right duly adapted Proportion, and that as partly for its Number of
Teeth, corresponding to its Revolution of 4 Minutes, but as thence in
chief, or as indispensably so to be acquired thereby, [viz. from which the
said Proportion] such a Qualification as whence, by my Construction
of the Pallats, the said Force so to be disposed of, as to give to the
Pendulum no more Irregularity in the Maintenance of its Motion in
Air, than as if it went, or could go by itself, and that for a long time in
Vacuo;
John Harrison, Concerning Such Mechanism, 1775
Harrison describes in Concerning Such Mechanism how he compared the
rate of his driven pendulum to its undriven rate. He was thus describing
escapement error which he could control and minimise by altering the effective
mass of the pallet frame and crutch to produce the square wave impulse
illustrated in the diagram showing the addition of the three forces. This ideal
solution can be most easily attained within the dimensions of the RAS regulator
by having an escapement wheel 68.76 mm. radius with pallets spanning 22/23
teeth.
The pendulum amplitude is similar in the three escapements (9.33E,
9.54E, and 9.92E). This is maintained in the Guildhall MS and the RAS
regulator by reducing the radius of the pallet arm circle.
The distribution of torque is most even in the RAS regulator and the
lateral shift and difference in pallet arm angles are significantly reduced.
Now, as touching the Matter, or first Matter in mine, viz. the length of
179
the Pendulum, as with respect to the Distance from its Center of
Motion, to where the Force of the Wheel is applied, is about in
Proportion as 100 to 1, and as without any such wedge-like thrusting
as spoken of above , and as withal no different Clamminess of Oil,
there not being, from the Construction and Material of my Pallats, any
Oil required, but as on the other Hand, and as implying, at a small
Distance from the Center of Motion of the Pendulum, and that besides
such other Virtue of Virtues as below, the Friction at the said Pallats
to be so far diminished by the Contrivance, as not to come to the 100
ed
Part of of what is in the common Way.
John Harrison, Concerning Such Mechanism, 1775
In conclusion, it is considered that John Harrisons pallets most likely
spanned 22-23 escapement teeth in his ultimate regulator and that the force
exerted by the escapement wheel should be taken as 3, then as just after their
interchanging it must be about as 2' in order to impulse the pendulum with a uniform
force having a moment 1/100 part of the length of the pendulum.
Cottinghams 1909 photograph, the elegant construction of the diagram by
Euclidean geometry and the dimensions at present in the RAS regulator are
evidence that the pallet lengths intended by Harrison are 52.21 mm. and 34.26
mm. spanning 22/23 teeth acting on escapement wheel 68.12 mm radius.
Escapement Error
A strong Proof indeed, that the Force or Draught of the
Pendulum Wheel of my Clock, and as in the first Place to be
understood, with its right duly adapted Proportion, and that partly for
its Number of Teeth, corresponding to its Revolution of 4 Minutes, but
as thence in chief, or as indispensably so to be acquired thereby, [viz.
from which the said Proportion] such a qualification as whence, by my
Construction of the Pallats, the said Force so to be disposed of, as to
give to the Pendulum no more Irregularity in the Maintenance of its
Motion in Air, than as if it went, or could go by itself, so as to be
observed to what Truth it had continued to measure Time, viz. as by
or from its going by itself, and that for a long time in Vacuo; and
therefore I may ask, if such a Matter be not highly worthy
Encouragement, what other Sort of Ingenuity or Discovery in the
World must be so? my Longitude Time-Keeper, own Sister to this,
excepted.
John Harrison, Concerning Such Mechanism, 1775
The effect of the escapement on the period of the pendulum whilst
180
driving it is quantified as escapement error and in the ideal clock is minimal.
Harrison draws attention to the problem in his writings and by drawing the
inner three circles in drawing B of the MS. 3972/3.
From the quotation above from Concerning Such Mechanism it is clear that
Harrison considered the concept of his asymmetrical pendulum escapement as
important as his use of a high velocity balance in a marine clock. He describes
how he considers advances in clock mechanics can only be made by comparing
the performance of two similar clocks when innovations are being made.
Escapement error can only be determined by comparing the performance of a
free pendulum to its performance when driven by an escapement.
Now this highly material Matter is from the Wheels acting [by my
Contrivance of the Pallats more weakly upon the Pendulum in each or
every one of its Descents, viz, from the Extremity of each Vibration, and
there, as at the first the weakest of all] to the Bottom or Middle of each,
and then [as still continuing] more and more strongly upon the
Pendulum in each of its Ascents, and the strongest of all just before the
interchanging of the Pallats, and the which Interchanging not being, or
to be far from the Extremity of each Vibration, and in which little, or as
it were overplus Part, a necessary recoiling of the Wheel to be, viz. as not
only for the interchanging of the Pallats, the which, as according to my
Construction of the same is required but as withal to have in some
Measure to do with the Effect of a Cycloid, but not to be the whole
Concern in that Matter ; the Pendulum withal requiring to be [viz, as
from my Contrivance of its Combination of Brass and Steel Wires]
rather, as mathematically speaking shorter when warm than when cold.
But as here, to come a little nearer in this material Point, let, as I order
the Matter, the Force [from the Wheel] upon the Pendulum, as just
before the interchanging of the Pallats, to be as by or from them the said
Pallats supposed or taken as 3, then, just after their inter changing, [and
the Force to contrary Direction it must but be about as 2, that is, it must
be so ordered, [as may hereafter be observed by the Drawing viz, as that
it be so by the taking, or supposing for the Purpose a Mean betwixt the
Actions of each Pallat and withal, as farther to the Purpose, that, as in
the little recoiling of the Wheel, to become less and less to the Extremity
of each Vibration but as whence, or as still on Course the greater at any
Time the whole Vibration may be, of more Efficacy the Same small Force
[and still as it were in the little recoiling] must from its Quantity or
Duration prove, and that in such small Measure as required; the resters
of the Pallats [viz, their Composers to relative Rest] having withal for
the better a little to do in the Matter, and the which is still from, or as
with Respect to the Length or Shortness of the Time [though always to be
181
esteemed as but little] that is to be occupied in the recoiling; it being to be
understood that was the Force of the said Wheel, or Pendulum-Wheel, to
be disposed of uniformly upon the Pendulum, throughout the whole of
each Vibration, that then the bigger that Force, and as with some recoiling
of the Wheel, the more it must tend towards occasioning a greater
Vibration, or somewhat a greater Vibration of the Pendulum to be the
sooner performed ; but that is not the Case in what is) as just here above
shewn, but, to the Purpose, is otherwise to be taken or considered as that
the bigger the Force towards the latter End of any Vibration, viz, as in
Comparison to what it may be at the fore End or Beginning, and as in
supposing without a Recoil of the Wheel, the slower the Vibration must
be, or the longer the Time it must take up, as mathematically speaking
wherein to be performed, but as with a proper recoiling, and artificial
Cycloid, rightly adapted, not so, viz, as when in such Largeness of
Vibration as must to this Purpose, as well as to other Purposes be as the
most necessary herein required, and as when, as must unavoidably be the
Case, in the moving by the Draught of a Wheel, or of the Wheels a Clock
in the Medium of Air; neither, as by any Means does a suitable Matter
of this, viz, of the Airs Resistance want to be avoided, as many have
foolishly imagined but is of real or great Use as the which I shall shew
presently and so as I may make bold to say there has not as heretofore
appeared in the World or to the Public, any Thing towards a strict or
rather, as along with other Things of mine, to, as I may say the most
strict or exact Mensuration of Time ;
John Harrison, Concerning such Mechanism, 1775
Harrison wishes to arrange his escapement so that it
being to be understood that was the Force of the said Wheel, or Pendulum-Wheel, to be
disposed of uniformly upon the Pendulum, throughout the whole of each Vibration,
His drawing of the escapement geometry indicate that the combined
forces of the escapement wheel and the weight of the pallet frame and
crutch give a uniform force throughout the pendulums motion.
Adjustment of the counterweight on the pallet frame and the crutch
pin will equalise the force exerted on the pendulum by each pallet.
If the amplitude of the pendulum becomes equal then the
force applied is a symmetrical square wave. This can be considered as
a fundamental sine oscillation with odd numbered harmonics. Using
Rolles theorem we can show that the mean value of this fundamental
is at the point of maximum velocity of the pendulum and the two can
be considered as being in phase.
182
If the balancing weight on the crutch is altered the force acting
on the pendulum becomes asymmetrical in shape. This can be
considered as a sine wave with odd and even harmonics. The harmonics
of this distribution are out of phase with the fundamental. The mean
value of this fundamental is no longer in phase with the pendulum.
There will therefore be escapement error. Harrison has designed an
escapement where the escapement error can be corrected by altering
the mass at the periphery of the crutch and the counterweight of the
pallet frame.
Gould replaced at least one of the pallet arms in the RAS
regulator and it is not known whether they had been replaced
previously. There is therefore doubt about whether the number of
teeth spanned by the present pallets is as Harrison intended. The
Guildhall escapement drawing MS 3972/3 does not have all the same
dimensions as the RAS regulator. Drawings and calculations have
been made for the dimensions of the RAS regulator to determine the
forces exerted when spanning 18 and 23 teeth. This shows that it is
significantly easier to organise uniform force with pallets spanning 23
teeth as found in the RAS regulator at present.
Sir George B. Airys (1801-92) classic paper On the Distur-
bances of Pendulums and Balances, and on the Theory of Escapements
(Cambridge Philosophical Transactions, 1827, Vol.3 part 1, pp.
105-128) analyses isochronous pendulums. His theory and equations
are particularly applicable to Harrisons pendulum and escapement.
Michel Rolle (1652-1719), French mathematician, elected to the
Acadmie Royal des Sciences in 1685. An early critic of calculus
who introduced the notation of the nth root radical sign, , in
n
x
his Traite dalgebre, 1690.
183
Having lately had occasion to investigate the disturbance produced in the
motion of a pendulum by a small external force, and having found by a
very general investigation a result of great simplicity, I perceived that the
usual theorems for the alteration in the time and extent of vibration
produced by the difference between cycloidal and circular arcs, by the
resistance of the air, by the friction at the point of suspension, etc. could be
made to depend upon it; and that these alterations could in fact be found
with greater facility form this general theorem than from independent and
unconnected investigations.
G.B.Airy, On the Disturbance of Pendulums, 1827
George B. Airy, was elected Lucasian Professor at Cambridge in 1826.
Became director of the Royal Observatory at Greenwich, with the associated
title of Astronomer Royal, in 1835 when he was 34, and held the post until his
retirement in 1881 at the age of 80.
At Greenwich he designed and installed the famous transit circle now named
after him, used for timing the passage of stars across the meridian. It is the
position of the Airy transit circle that defines the Greenwich meridian and
since the Washington Conference of 1884 that meridian has been the basis of
the worlds timekeeping and navigation. Hence Airy can be said to have
brought the worlds clocks into step.
In 1825 Airy designed the first eyeglasses to correct for astigmatism, from
which he suffered. His 1827 paper on the disturbance of pendulums read to
the Cambridge Philosophical Society was widely accepted; he specified that the
Westminster clock when completed should keep time accurate to one second.
This accuracy would not be achieved by the dead beat escapement proposed by
Vulliamy and the need to achieve Airys specification led to Lord Grimthorpes
development of his double three legged gravity escapement forcing a 2 second
zinc-iron temperature compensated pendulum. He also published important
works on the theory of tides, particularly on the propagation in canals and
estuaries. He contrived a simpler method of clearing lunar distances which
only requires the use of five figure logarithms.
184
As Mr Graham*s scaping is frequently mentioned by them, it will not here be
improper to give a just description of it, illustrated with a figure taken exactly from a real
scaping, executed and communicated by Mr John Shelton, who was chiefly employed for this
purpose by the late Mr Graham.
Describe a circle whose diameter is that of the intended
swing wheel, and through its center draw a perpendicular, or
vertical line, prolonged upwards. Then, if the number of teeth in
the swing wheel be* thirty, as in clocks vibrating seconds, set off
on the circle, on either side the vertical point (from an exact line
of chords) an arch of 69 degrees, the double whereof, 138
degrees, is the exact space taken up by eleven teeth and one half,
on the same circumference. From the center of the circle to the
points of 69 degrees draw radii, and on their extremities erect
perpendiculars, whose intersection in the vertical line, will be the
center of motion, of the anchor represented in the figure, the
circle passing through the extreme points of the teeth of the wheel, shewn also in the figure.
From the centre (sic) of motion, through the points of 69 degrees, draw a circular arch, with
that part of each of the pallet of the anchor, which receives the last escaped tooth, and keeps the
second ha*nd from recoiling, must coincide, as the figure shews; lastly, the inclined planes of
the pallets must make an angle of about 60 degrees, with lines drawn* from the center of the
wheel, to their obtuse terminations. Such is the construction of this scaping, which acts in the
following manner. The tooth c for example, having just escaped from the pallet a, the opposite
pallet instantly receives the full shock of the tooth b, on its circular arc; and the vibration
proceeding, this pallet enters deep between the teeth, but not so far as to touch the bottom, the
swing wheel and second hand remaining motionless, till by the succeeding vibration, the tooth
b is brought to the edge of the inclined plane of its pallet, at which instant it begins to act,
pushing away the pallet till it scapes at its lower point, when immediately another tooth striking
on the circular part of the other pallet, rides at rest upon it, till the inclined plane begins to
present itself, and then following the slope of the pallet, pushes it away, and at last scapes, as
did the first tooth c; and so on.

This, I believe, is the first of all contrivances of dead scapings; being put in execution by Mr
Graham, before the year 1720, in a clock he made for the reverend Mr John Whiteside, then
keeper of the Ashmole museum at Oxford, I do not know that it has been any where particularly
described, except in M. Marononi*s book of astronomical instruments, and in M. Thiout*s
treatise on clock making, but imperfectly in both. It has the property of acting with a very small
moving force, a matter of great consequence in many respects, and is little liable to wear, tho*
the friction is considerable; but this sort of pendulums, vibrating but a very small arc, and the
ball, or regulating power, being very heavy, cannot undergo any sensible variation, in measuring
time, on account of the friction, which is invariably the same.
from The Gentlemans Magazine 1754
185
John Shelton made a series of at least five regulators for the observations of the
transit of Venus in 1761 and 1769.
Captain James Cook took one of
these regulators for his observations
of the transit of Venus on his first
voyage, and two of them on his
subsequent two voyages. Nevil
Maskelyne used two of these
regulators for his observation of the
transit of Venus on St. Helena in
1761, for determining the latitude
and longitude of Barbados for the
trial of H4 and for his investigation
at Schiehallion measuring the
deflection of a plumb line by the mountain mass. The regulators were used for
many other investigations into the force of gravity, particularly with Katers
compound pendulum.
Henry Kater (1777-1835) described his method in Philosophical Transactions of the
Royal Society, 1818, vol. 11. The detached pendulum was pivoted on a knife edge in
front of the regulator pendulum. The time when the pendula are synchronised can
be observed through a telescope to eliminate parallax. The number of vibrations
between successive synchronous beats are used to determine the frequency of Katers
pendulum and thereby measure the force of gravity.
Friedrich W. Bessel (1784-1845) improved Katers method by eliminating
atmospheric corrections. He described the corrections in Untersuchungen ber die
Lnge des einfachen Secundenpendels in 1826.
The Shelton regulators have in many ways been altered since they were first made but one
original pendulum, suspended on a knife edge, still exists. It displays many features of the
Harrison pendulum seen in the King portrait.
Chapter 7
Power Supply
Bearings and Friction Wheels
Friction is the force between two surfaces in contact.
The first known definition of friction was given by Leonardo
da Vinci. Da Vinci observed that "the friction made by the same
weight will be of equal resistance although the contact may be of different
breadths and lengths" and that "friction produces double the amount of
effort if the weight be doubled."
Da Vinci's ideas about friction remained buried in his
notebooks for some 200 years, until a French engineer,
Guillaume Amontons (1663-1705), reformulated them.
Then, in 1781, C. A. Coulomb (1736-1806) verified
Amontons' observations in the laboratory and stated them in
the form of two laws of friction.
1. Friction is independent of the area of the solids.
2. Friction is proportional to the load between the
two surfaces.
Coulomb concluded that the nature of friction might
depend on the interlocking of surface irregularities, and the
"drag" we feel would be due to the work required to lift the
load over the irregularities. F. Philip Bowden and David
Tabor (1986) gave a physical explanation for the laws of
friction. They determined that the true area of contact is a
very small percentage of the apparent contact area. The true
contact area is formed by the asperities. As the normal force
increases, more asperities come into contact and the average
area of each asperity contact grows. The frictional force was
shown to be dependent on the true contact area.
Static friction is the force between two objects that
are not moving. Kinetic friction occurs when two objects are
moving relative to each other and rub together; it is
significantly less than static friction. Rolling friction is the
frictional force associated with rotational movement of a
187
wheel or other circular object along a surface and is much less
that kinetic friction. It must not be confused with the
sliding friction occurring between a shaft (arbor or gudgeon)
rotating in a plain bearing.
The less rubbing there be of the Axle, the better it is for this
Effect; upon which account, Steel Axes and Bell-Metal Sockets, are
much better than Wood, clamped or shod with Iron; and Gudgeons of
hardened Steel, running in Bell-Metal Sockets, yet much better, if there
be Provision made to keep out Dust and Dirt, and constantly to supply
and feed them with Oil, to keep them from eating one another; but the
best Way of all, is to make the Gudgeons run on large Truckles, which
wholly prevents gnawing rubbing and fretting.
Robert Hooke, Paper to Royal Society, 1685
The pendulums oscillation is maintained by the
escapement. By virtue of its design Harrisons escapement
has no drop and hence no impact forces to propagate
vibrations through the supporting structure to assist in
unlocking static friction in the gear train. In a clock the
torque from the falling weight is
delivered to the escapement
through a train of gears acting in
the opposite direction from most
transmissions. The motion is
driven from slow speed (falling
weight) to the higher speed of the
scape wheel. Harrison used
friction wheels and rollers in the
lantern pinions to significantly
reduce friction throughout this
train. Henry Sully (1680-
1728), who had been apprenticed
to Graham, used friction wheels in the escapement of a
proposed marine timekeeper that he presented to the
Acadmie des Sciences in 1724.
Let GH represent two Rolls, on which one end or Pevet cf the
Pendulum-Wheel Axis moves, (the other end also haveing the same) but
one of em at each end bears most of the Weight thats there; so the
Friction is abated almost as the Diameter of the Pin to that of that
Roll.
John Harrison 1730 Manuscript
188
The train of gears driving the hour wheel and calendar
wheel are acting from high speed to low speed; the bearings
are wooden collets on brass pins and the pinions are fixed.
The centre, third, remontoire and escapement wheel
arbors run on pairs of brass and wood friction wheels. The
remontoire and escapement wheel arbors are co-axial having
their own friction wheels. The brass friction wheels are
mounted on wooden collets running on brass pins. The grain
and colour of the wood used for the collets in the RAS
regulator suggest that it is lignum vitae.
The angle subtended by the pins of the wheels and
the arbor that they support is a right angle, the bisector of
which is parallel to the tangent of the junction of the driving
pinion and wheel. Harrison drew these details in his 1730
document. The arbors are held against the friction wheels by
the driving torque. These bearings are therefore only suitable
where the torque is in one direction, i.e. they will only
support recoil if their arbors are horizontal. The depthing of
the wheels can be controlled easily by altering the
circumference of the friction wheels. This is a consequence
of their geometry.
Many of the pairs of friction wheels have different
diameters so that the combination of bearing surfaces is
continually altering.
The pallet frame arbor bearings are 35E knife edges
resting in v-grooves in glass. They can accommodate
horizontal and vertical thrust.
Let A represent two bits of Glass, whereon each end of the
Axis of the Pendulum being a Sharp Edge of Brass moves in a little
notch in the Glass, which renders that Friction insensible, as also the
Wearing (for tis but the Axis of the Pallats, which communicates force
to the Pendulum; for the Pendulum it self is otherwise suspended, viz
by a thin Brass Spring).
John Harrison 1730 Manuscript
Harrison would almost certainly have had experience
of moving heavy bells with boards and rollers. Here the
rollers move along the board at the same rate that they move
189
along the ground so that there is rolling friction but no
sliding friction. When rollers are used between two
concentric circular surfaces there is a small amount of sliding
friction to accommodate the difference in circumference.
This friction is appreciably less than in a plain bearing. The
Madrid manuscript of Da Vincis drawings has details of a
roller bearing. Roller bearings can be designed for high load
bearing and work equally in both directions.
The great wheel arbor has two roller bearings on the
front and rear plates. The bearing on the front plate has a
larger diameter than that on the rear. They were both
originally fitted with cover plates to keep out dust and dirt.
They are designed to work without lubrication. The rollers
have the same diameter as the inner bearing surface. They are
constructed of a copper and tin alloy which appears by its
colour to have a larger proportion of tin than bell metal
(probably speculum metal defined as an alloy of copper and
tin containing a high proportion of tin - maybe up to 45%).
Gould noted that in the large front bearing the brass pins are
fixed into the rollers and turn in the brass ring.
The rear bearing has a smaller diameter than the
front. When the torque of the driving weight is applied to
Leonardos Drawing
190
this arbor it exerts an axial force holding it to the lignum
vitae thrust bearing attached to the rear bearing. This thrust
bearing also accommodates the increased axial thrust when
rewinding.
Driving Weight and Pulley
The efficiency of the the roller bearings and friction
wheels reduces the power losses in the train. The remontoire
controls the torque available to the escapement; this torque is
adjusted by the properties of the helical springs and levers on
the periphery of the remontoire wheel. The power necessary
to re-tension the remontoire springs and to move the going
and motion work is adjusted so that the fly is able to control
the movement. The fly fitted by Harrison as seen in
Cottinghams photographs is smaller than Goulds
Roller Bearing Metal
The metal used for the roller
bearings is an alloy of 60% tin
and 40% copper. It is made in
the workshop by dissolving
copper scrap (copper pipe
offcuts) in molten tin in an
iron crucible. Boric acid is
used as a flux. The
constituent metals are weighed
before melting to attain the
correct proportions. This alloy when cooled consists of large
rhomboidal copper crystals in a tin matrix. This structure is readily
seen on the surface of the metal if it cracked open before complete
cooling. Whilst cooling the alloy develops a characteristic pasty
constituency similar to plumbers wiping solder. Before pouring ensure
the molten metal is covered with flux and that it is stirred to release
dissolved gases. When machined the copper-tin crystals break
away from the matrix producing a black powdery swarf. See D.
Tabor The Friction Properties of Some White Metal Bearing Alloys, the
Role of the Matrix and the Hard Particles J.Appl.Phys. (1945) v.16,
pp.325-337
191
replacement. Gould had to overpower the clock because of
his large, heavy brass crutch which unbalanced the
escapement. Very little driving weight is necessary when the
clock is set up properly. The small forces together with the
substantial frame ensure the great stability necessary to attain
his quoted accuracy.
Cottinghams photographs do not show a driving
weight or pulley; the cord attached to the drum is bundled
roughly between the plates. The lower left pillar has a deep
nick to secure the standing part of the driving cord,
suggesting that a pulley is required. This feature could have
been fitted by Harrison or the subsequent restorers.
Gould drew a pulley with friction wheels in two places
in his notebook. He communicated with Cottingham whilst
he was restoring the regulator and used Cottinghams
photograph in his 1927 account in the Horological Journal.
There is no evidence that Harrisons pulley has survived, but a
pulley with friction wheels would accord with his principles.
The pulley drawn in Goulds notebook is not in Harrisons
style but an interpretation is suggested.
The nick in the pillar for the standing end is neither
midway between the plates or opposite the mid position of
the drum. As the cord unwinds off the drum the moving part
of the cord twists the pulley as it approaches the front plate.
When the geometry is examined it is seen that the maximum
movement is during
the last two thirds of
the unwinding when
the distance between
the the drum and
pulley is greatest. The
distances between the
standing part and the
moving part at fully
wound, fully unwound
and minimum have
been averaged. The
average is the same as
fully wound, 84.32
mm. It is suggested
4
4
.
4
5

m
m
82.60 mm
84.32 mm
85.95 mm
192
that this should be the diameter of the pulley. This detail
leads one to believe that Harrison cut the nick and intended
the clock to have a two part driving cord. The nick is cut on
the top of the pillar.
Every part of this clock when closely examined
exhibits Harrisons intense understanding of mechanics.
There is no part that has not been designed for optimal
efficiency.
Wheels and Maintaining Power
The Pendulum Wheel is best made of Brass, but all the other
Wheels rather better of Wood than of Brass, But as I do not here
undertake to treat in full, I need not mention what Wood or how I
193
order it. Now my Pinnions are not made as the ordinary ones; for
some parts thereof have when in use a relative Motion, which said parts
must be made of Wood, viz. Rolls, and move upon small Brass pins,
which reach betwixt two Brass plates, in which the Pins are fixed, on
which I find by experience these Rolls of Wood move so freely, as never
to need any Oyl; But from the nature of the Contrivance the Friction
will be abated in such proportion, as the Diameter of the Pin to that of
the Roll. And chiefly in the application of the Pendulum to the Wheels,
or its Wheel (which is the head place) there 'tis abated no less than 40
times; And also the Pevets or ends of the Pendulum Wheel Axis moving
upon Rolls of great Diameter in proportion to the said Pevets do greatly
abate the Friction: And the Wearing is also abated in a greater
proportion.
John Harrison, 1730 Document
Harrison used oak for the frames and wheels of his
pendulum clocks prior to making the RAS regulator. The
coefficient of thermal expansion of well seasoned oak (0.0049
mm./m./EC) is nearly a quarter (0.259) of that of brass
(0.0189 mm./m./EC). Dimensional changes caused by
alteration in humidity can be lessened by sealing the wood with
linseed oil or varnish. When making his wheels he followed
standard millwrights practice of having the grain of the wood
radially disposed in the teeth. He did this by laminating
sectors of teeth, most probably made of apple or pear wood,
into a radial grove in an oak plate. (See James Newlands, The
Carpenters Assistant, 1860).
When he introduced lantern pinions throughout his
1728 regulators he arranged for the pinion rollers to make
contact throughout their action on the radial sides of the teeth.
The radiused ends of the teeth were to enable clearance. This
meant that any slight change in the dimensions of the frame
altering the depthing has insignificant effect.
Harrison 1730 drawing shows
that the chordal pitches rather than the
circular pitches of the wheel and pinion
are equal. He made the roller diameter
93% of the pitch. Two adjacent teeth
are in contact with rollers when the
rollers are equally placed about the line
of centres at which point the drive is
transferred from one tooth to the
other.
Let AB be a portion of a Wheel, CDE the Pinnion which it
194
turns, let a a a etc. be Rolls, which move relatively on small pins at
their centers to submit to the Teeth of the Wheel as they reach nearer to,
or draw farther from, the Center of the Pinnion; by which there is no
Friction at the Teeth, but only at the Center of the Rolls, which must
be less than if it were at their Circumferences, as the Circumference of
a Pin to that of a Roll.
John Harrison, 1730 Document
When examined in detail,
Harrisons drawing of the wheel and
lantern pinion show that the force is
transmitted from the wheel before the
line of centres. His drawing shows the
point of first contact with the wheel
moving clockwise and shows clearly the
movement of lantern pinion roller down
the wheel tooth.
The radial movement of contact of the roller on the
tooth as described by Harrison above means that there is a
cyclical variation in the transmitted torque. This is
demonstrated for wheel ratios in the 1728 and RAS regulators.

360
2N
360
2n

=
=
For 1728 Regulator 8/60 wheel ratio
If r = 1 then
1. sin22.5
R 7.312
sin3
= =
OE = 7.312 cos 3 = 7.302
C O
E
R
r
P/2
a b
a
b
195
CE = 1.cos 22.5 = 0.9239
OC = OE + CE = 8.2259
CE 0.9239
Minimum Torque 0.1265
OE 7.302
1
Maximum Torque 0.1384
OC 1
Mean Torque 0.1325
= = =
= =

=
For RAS Regulator 12/90 wheel ratio
If r = 1 then
1. sin15
R 7.4161
sin2
= =
OE = 7.4161 cos 2 = 7.4116
CE = 1.cos 15 = 0.9659
OC = OE + CE = 8.3775
CE 0.9659
Minimum Torque 0.1303
OE 7.4116
1
Maximum Torque 0.1355
OC 1
Mean Torque 0.1329
= = =
= =

=
Intermediate values have been calculated and are
graphed as percentage differences from the mean values. The
calculations have been made for a theoretical lantern pinion
where the diameter of the rollers is zero. The rollers having
93% pitch diameter skews the curve and slightly alters the
values because the point of contact of a tooth against the roller
is a tangent.
The variation in torque becomes accumulative in a train
of gears. It is therefore necessary to minimise the number of
wheels in the train and as demonstrated above maximise the
number of rollers on each pinion and teeth on each wheel.
196
0
+
1
+
2
+
3
+
4
+
5
+
6
S
e
q
u
e
n
c
e

o
f

F
i
g

1

d
r
a
w
i
n
g
T
h
e
1
7
3
0
d
r
a
w
i
n
g
s
h
o
w
s
c
o
n
t
a
c
t
b
e
t
w
e
e
n
r
o
o
l
e
r
a
n
d
t
o
o
t
h
b
e
f
o
r
e
t
h
e
l
i
n
e
o
f
c
e
n
t
r
e
s
j
u
s
t
a
f
t
e
r
c
o
n
t
a
c
t
h
a
s
b
e
e
n
m
a
d
e
(
0
.
5

)
.
T
h
e
s
e
q
u
e
n
c
e
o
f
d
i
a
g
r
a
m
s
s
h
o
w
s
t
h
a
t
t
h
e
p
r
o
p
o
r
t
i
o
n
s
o
f
h
i
s
d
r
a
w
i
n
g
r
e
s
u
l
t
i
n
m
i
n
i
m
u
m
c
l
e
a
r
a
n
c
e
b
e
t
w
e
e
n
r
o
l
l
e
r
a
n
d
n
o
n
-
e
f
f
e
c
t
i
v
e
t
o
o
t
h
.
T
h
e
d
e
t
a
i
l
d
r
a
w
i
n
g
s
h
o
w
s
t
h
e
e
f
f
e
c
t
i
v
e
p
i
t
c
h
d
i
a
m
e
t
e
r
i
n
r
e
d
.
T
h
e
r
o
l
l
e
r
c
e
n
t
r
e
s
a
r
e
e
q
u
i
d
i
s
t
a
n
t
o
n
t
h
e
c
h
o
r
d
a
l
d
i
a
m
e
t
e
r
d
r
a
w
n
i
n
g
r
e
e
n
.
T
h
e
r
o
l
l
e
r
s
m
u
s
t
b
e
c
o
n
s
i
d
e
r
e
d
a
s
a
n
e
x
t
e
n
s
i
o
n
o
f
t
h
e
t
o
o
t
h
,
n
o
t
p
a
r
t
o
f
t
h
e
p
i
n
i
o
n
!
T
h
e
r
e
f
o
r
e
t
h
e
p
o
i
n
t
o
f
a
p
p
l
i
c
a
t
i
o
n
a
n
d
r
e
c
e
p
t
i
o
n
o
f
t
h
e
t
o
r
q
u
e
i
s
t
h
e
p
i
v
o
t
o
f
t
h
e
r
o
l
l
e
r
,
n
o
t
t
h
e
p
o
i
n
t
o
f
c
o
n
t
a
c
t
o
f
t
h
e
r
o
l
l
e
r
a
n
d
t
o
o
t
h
.
f
i
g

1

E
n
l
a
r
g
e
d
.
t
c
w
/
3
Harrison improved the RAS regulator in this manner
and also introduced the remontoire to smooth the power
applied to the escapement. The wheel ratios of the great
wheel and the third wheel to their pinions are not integers
thereby increasing the number of tooth combinations.
It is important that the regulator is not accidentally
stopped, particularly when the remontoire is unwound. In the
pendulum clocks prior to the RAS regulator he fitted a
conventional bolt and shutter weighted maintaining power
acting on pins in the second wheel of the train. A spring
maintaining power was introduced in his first sea clock
mounted on the great wheel arbor. It is mounted on the arbor
of the centre wheel of the RAS regulator. It is shown in
Kings portrait as a drawing under his left hand. The
maintaining power on H1 has vernier adjustment; that fitted
to H2, H3 and the RAS regulator is simplified. The
maintaining power spring of the RAS regulator can be
tensioned by anti-clockwise motion of the winding arbor.
This is helpful when restarting the fully unwound clock.
A pamphlet was published in 1714 by Jeremy
Thacker, of Beverley, describing a timekeeper which he was
preparing for use at sea. The machine was suspended in
gimbals and was driven by springs with an auxiliary spring
maintaining power. The movement was enclosed in an
198
evacuated glass flask. The most important feature was the
replacement of a pendulum with a balance wheel having a
spiral spring. The pamphlet is remarkably perceptive for an
unknown clockmaker. It is now considered that the pamphlet
could have been written by Dr John Arbuthnot as a satirical
comment on the response to the 1714 Queen Anne
Longitude Act. Queen Anne died three days after the Act was
passed by Parliament. There are indications in Harrisons
writings that he was aware of the contents of this pamphlet.
In the RAS regulator and the third sea clock the
centre wheel pinion is loose upon its arbor, being mounted on
a barrel containing and fixed to a coiled spring of three and
half turns. The other end of the spring is attached to a catch
plate fixed to the arbor, which operates between the extended
pivots of the pinion rollers. The centre wheel is fixed to the
arbor. The ratchet wheel is attached to the barrel. It has two,
gravity operated, half tooth pawls which are fixed to the rear
plate. This arrangement is necessary because the centre arbor
is kept in contact with the associated friction wheels by the
force of the driving weight. If this force is relieved the arbor
falls away from its friction wheels.
Diagram from Jeremy Thackers Pamphlet
199
The great wheel exerts is torque on the centre wheel
pinion which tensions the spring until the catch plate comes
up against its stop. Continuing torque drives the clock
carrying the ratchet wheel past its pawls. Whilst the clock is
being wound the ratchet wheel prevents retrograde movement
of the centre wheel and the maintaining spring tension
continues to drive the clock. The spring is tensioned by the
catch plate to be slightly weaker than the driving weight. The
shape of the catchplate allows 22 rotation giving
approximately seven minutes of maintaining power.
Harrison lowered the position of the great wheel
arbor by 4.46 mm. from the position determined by his
layout concept. This enabled him to make maximum use of
the space by fitting a 78 tooth wheel, place the winding hole
below the minute outer circle and have the same tooth pitch
as the centre wheel. He increase the width of the plates by
the same amount which facilitated the fitting of the friction
wheel for the third wheel.
Dr John Arbuthnot (1667-1735) F.R.S.,
London polymath, physician and satirist.
Published the first book in English on
probability, On the Laws of chance in 1692.
This was a translation of Christiaan
Huygens De ratiociniis in ludo aleae. He
became physician extraordinary to Queen
Anne in 1705 and with Halley and
Newton forced John Flamstead to publish
his Historia coelestius in 1712. He inspired
Jonathan Swifts Gullivers Travels and created John Bull in his
satirical pamphlets written to assist the Tory party. When Queen
Anne died he lost all of his royal appointments including physician
of Chelsea Hospital.
200
Remontoire
...and it is certain that my next Clock, when finished,
properly set up, and duly adjusted, will come to the Truth nearer still.
John Harrison, Concerning Such Mechanisms, 1775
The most significant improvements Harrison made to
his RAS regulator when compared to his 1728 clocks is his
maintaining power and remontoire. He fitted a remontoire to
his third sea clock and to his Longitude watch, H4. A spiral
spring is fitted to drive the escapement arbor from the
remontoire wheel on the RAS regulator. The remontoire
wheel has the same fittings on its periphery as the remontoire
wheel in H3. In H3 the remontoire wheel drives the
escapement wheel by helical springs acting on a cam. In the
regulator there is sufficient space available for a similar
system which must have been intended as this arrangement
provides a superior performance.
The Guildhall Library has a drawing showing
Harrisons geometrical construction of this structure. The
dimensions of given in this drawing match the remontoire
wheel in the RAS regulator. Harrison uses a spiral spring in
the remontoire of his watch, H4, because of the constraints
of size. This suggests that he intended to use a helical spring
201
and cam in his ultimate astronomical regulator, and perhaps
used the RAS regulator as an experimental tool in developing
the form used in H4.
Harrisons drawing shows his geometrical and
analytical construction drawn in metal point. These lines are
extremely fine and accurately drawn. The lines of the cam and
pitch circle have been emboldened with black ink. Beside the
drawing there are handwritten annotations.
Springs No 6 Plate Mark 9 with 3 Pricks having 10 holes in it; each
had 14 turns besides what brings the end to the centre, but had better
had 15, and the curve a little higher at the end a; and about as long at
that end as here described, viz. about inch longer. The part at the
other end of the curve where the Hook hangs had better been shorter, and
also the Hook but the curve itself i.e. taken off it, the out end, not the
curve more thereby longer.
The balance springs, No 3 in a place marked with a cross, having six
holes in it.
3 ounce would draw each little spring 3 inch, including the space
the spring itself took and the curve was adjusted to draw 1 ounce at
3
16
the inner end, or bottom of the wheel teeth. Avoirdupois weight.
John Harrison Collected Drawings 3972/3 Page 13
The remontoire fitted with helical springs and cams enables the
torque applied to the escapement wheel to be accurately
controlled and smoothed. Any irregularities in the power train
become irrelevant because it is only tensioning the helical
spring intermittently.
202
The escapement wheel and remontoire arbors are co-
axial and each have their own friction wheels. The only
connection between them is
through the driving springs. The
detent mechanism is divided
into two functional parts,
locking the main train and
resetting of the fork on the
escapement wheel pins. A
stop pin is carried by an arm
attached to the fly arbor.
This pin is locked against a
detent carried by a weighted
arm attached to its own arbor.
The arbor is turned by
another arm which is pushed
down by one of the pins
attached to the escape wheel.
The pin pushes on a small fork pivoted on the end of the arm.
As the fork is pushed, it pivots away from a fixed stop attached
to the arm. When the fork is pushed far enough, the detent has
moved out of the way of the stop pin on the fly and the train is
free to turn.
The remontoire wheel turns tensioning the helical
springs being controlled by the fly. The detent arm carries a
203
roller on its end which is picked up by a cam on the remontoire
arbor whilst it is moving. This lifts the fork clear of the pins
and then drops it back onto the next pin as the train locks.
When the fork is lifted away from the pin it falls back onto its
stop. It is much easier to correctly tension the helical springs
than the spiral spring. Gould cut a window in a plate to enable
him to see the escapement wheels pins and fork. He also
reshaped the fork.
During the rewind phase the escapement is powered
directly by the drive weight or maintaining power. The rate of
the clock can therefore be affected by increasing the driving
weight. This is described in Goulds notebooks. When he
increased the driving weight he also had to enlarge the
remontoire fly to maintain control.
With a fly, sized according to Cottingham photographs,
204
governing the remontoire, the driving weight should be adjusted
to provide enough torque to drive the going train, motion work,
to rewind the remontoire springs allowing enough extra torque
for the escapement to impulse the pendulum. This can only be
done when the torsional force of the helical springs required to
correctly drive the escapement have been determined.
The remontoire fitted with helical springs provides
controllable and even torque to the escapement wheel. Gould
altered the operating cam by rounding its leading edge. This
was unnecessary because the roller has been lifted by the
latching arm at the beginning of the rewind. Gould fitted a
heavy crutch to the regulator which altered its dynamics so that
the torque applied to the escapement wheel had to be markedly
increased. In order to control this extra torque he had to
increase the size of the fly. The copper shaped remontoire fly
was replaced with a flat brass fly of irregular shape necessary to
enable clearance. If a crutch is fitted as it Cottinghams
photographs the fly having its original size will control the
necessary torque. The time taken for rewinding the remontoire
spring is reduced as much as possible whilst allowing safe
operation of the fly detent.
The manuscript drawing of the remontoire cam is
fascinating as it shows Harrisons design methods most clearly.
At first sight the drawing appears symmetrical but the analysis
lines are different for each side. The drawings are made so that
it is possible to confirm them by calculation. The geometrical
derivation of curves for the cams which have the optimal shape
is perhaps the most remarkable aspect in the design of this
clock. Concerning Such Mechanism has calculations for his method
of tempering the musical scale using five figure logarithms to
the base 10; he may have calculated after deriving his design by
drawing. The drawing displays his skilful draughtsmanship and
knowledge of trigonometry.
His geometrical method has been followed,
superimposed upon a tracing of his drawing and then algebraic
calculations made to analyse and ascertain the efficiency of his
design.
205
Geometrical Design
1. Draw horizontal base line. Draw line AB 18.43E (arctan
1/3 ) crossing horizontal line.
2. Draw a perpendicular CD to line AB at the intersection
of line AB with horizontal line.
3. Bisect the angle BOC by drawing line QR 26.57E
(arctan ). Draw the perpendicular EF.
4. Draw action circle 13/8 diameter, 11/16" (17.463 mm.)
radius centred on intersection of axes.
5. Mark intersection of circle with line EF G.
6. Draw circle radius 1 13/32"(1.40625, 35.72 mm.) from
G and mark intersection with QR, H.
7. Mark the points I and J on the circle 45E from line GH
to produce a 90E segment. This is Harrison base arc.
8. Draw extended line through G and H. The centre of
the adjusted cam arc is on this line.
9. Draw arc 1 4/12" (1.667", 42.342 mm.) radius centred
on extended line GH passing through points I and J.
The arc of this circle between I and J is the outline of
the Harrisons working part of the cam.
10. Draw peripheral circle 4" radius
11. Divide the peripheral circle into 32 parts.
12. Draw tangents from a 45 sector of the arc to the
divisions on the peripheral circle.
13. Mark equal 45E sectors on either side of line EF.
14. Draw arc 1 1/4" radius from G to mark the inner profile
of the cam.
50.00 mm
50.00 mm
18.43
A
B
C
D
E
F
G
H
I
J
R0.69 in
Q
R
50.00 mm
26.57
45.00
R42.33 mm
R35.72 mm
206
15. Draw circle 3/8" radius for outer margin of collet.
16. Draw arc 13/16" centred on vertical axis and passing
through origin to join inner and outer profiles of the cam.
17. Harrison marked the outer end of the spring 3 3/4"
from the centre at 45Efrom the horizontal axis, together
with the maximum and minimum position of the
adjustment arm. The position of the fittings on the
periphery of the remontoire wheel are also shown thus
enabling the determination of their dimensions.

The drawing has two series of calibrations.
a. Five concentric circles centred on O 1/8 apart
beginning with 1" radius.
b. Non-linear scale of 1/5" chords of a 2 3/8 circle
tangential to axis CD, centred 1 5/8 from axis
AB, projected onto axis CD.
In this drawing Harrison has corrected the cam profile
to improve the linearity of its performance. The drawing is
remarkable in the choice of axes. The perpendicular line to EF
which is the base line for the drawing of the cam has a gradient
of 0.5 and the line 45E to it has a gradient of 0,333.
G
G+
H
P
E
F
52.76
45.00
207
The cam profiles have been calculated and the results
plotted on a graph drawn to demonstrate their properties. For a
cam to provide completely uniform torque the graph of spring
extension and resultant torque would be a line whose gradient is
dependent upon the force applied by the spring. Harrisons cam
does not have this absolutely ideal relationship but the graphs
show that he was successful in obtaining an adequate
compromise allowing sufficient curvature on the cam to enable
it to function properly.
Algebraic Method
1. Determine gradient, m
b
, of line EF, the baseline.
1 1
m 2.00
tan tan26.56
= = =

2. Determine point G at intersection of line EF and action
circle 17.46 mm. radius using simultaneous equations.
2 2 2
x y R + =
y = mx
( )
2
2 2
0 x mx R + =
( )
2 2 2
2
2
1 0
1
m x R
R
x
m
+ =
=
+
2
2
17.463
x 7.8095
1 ( 2)
= =
+
y mx 2 7.8095 15.619 = = =
3. Determine point H at intersection of baseline and
35.72 mm. radius circle centred on G.
2 2
2
h
2
2 2
h g g
h
2 2
h
2 2
(x-7.809) +(y+15.618) =35.7188
1.25x 7.809
x =27.870
y =0.5x =1
(x +(y =R
3.93
y
5
=mx
15.618 3
x ) y )
5.7188 +

+ =
4. Determine the equation of the line GH. The position
of G is 7.809, -15.619, H, 27.870, 13.935
Determine gradient:-
208
h g
h g
29.554
=1.473
20.061
y -y
m = =
x - x
Determine equation:-
c = y - mx
= 1.473 x 27.870 - 13.935 = 27.118
y =1.473x- 27.118
5. Determine equations for lines from centre to a 45E
sector of the cam periphery.
y = mx
2 = 4.07E, 9.695E, 15.32E, 20.945E, 26.57E,
32.195E, 37.82E, 43.445E, 49.07E
m = tan 2
m = 0.0712, 0.1708, 0.2739, 0.3828, 0.500,
0.6296, 0.7762, 0.9471, 1.153
6. Determine torque circle radius for line 3 where y =
0.500x by finding position of intercept of circle and
line using simultaneous equations.
2 2 2
g g
(x-x ) +(y-y ) =R
y = mx
2 2
2 2 2 2 2 2
g g g g
2 2 2 2 2
g g g g
2 2
2 2
2
x - 27.809x +7.809 +0.500 x
+20.50015.619x+15.619 =35.7
x - 2x .x+x +m x +2
19
1.25x +0
my .x+y =R
(1+m )x +2(my - x )
.0034x- 970.894 =0
x =27.8
x+(
72
y
x +y -R )=0
=13.939

2 2
Torquecircle radius 27.872 13.939 = +
= 31.163
7. Find intersection point, P, of tangent line and 100 mm.
radius peripheral circle by solving simultaneous
equations
2 2 2
+ = x y R
2
= + y m x c
( )
2
2 2
2
0 x m x c R + + =
( )
2 2 2 2
2 2
1 2 0 m x m cx c R + + + =
The gradient of the tangential line is derived from the
gradient of the line from H.
209
2
1
m =-
m
c
2
= y - m
2
x
Solve using quadratic equation.
2
2 2 2
2
c =13.939 0.6787 27.872 32.855
x +( 0.6787x+32.855) =100
1.461x 44.603x 8920.43 0
x =-64.357
y =(-0.6787 64.357) 32.855
76.539
+ =

=
+
=
8. Find distance to peripheral circle

2
2 2 2
2 1
2 2
2 1
d ( 64.357 27.872) (76.539 1
d =(y
3.93
- y ) +(x
9)
- x
d =111.46
)
7
= +
9. Determine position of I and J by calculating relative co-
ordinates to G.
Angle of line GH = atan 1.4733 = 55.833E
Angle of line GI = 100.833E, m = -5.226
Angle of line GJ = 10.833E, m = 0.191
For I

2
r
a
2
a
r
2
2
35.719
=6.713
1+5.226
7.809 - 6.713=1.096
6.713 5.226 = 35.082
y =35.082 - 1
R
x = =
1
5.619
+m
=1
x
9.463
=
y =
10. Determine the centre on the line GH of the cam arc with an
altered radius. (R=42.418)

2 2 2
i i
G
2
H
1.4733
15.619 1.4733 7.8095 27.125
3.1706x 139.4
(x x ) (y y ) R
y mx c
m
c y mx
x y
67x 372.3256 0
2.8550, 22.9182
=
+
+ =
= +
=
=
=
=
= =
10. Repeat calculations for arcs which pass through points I and
J having incremental radii and centres on line GH.
11. Plot results of torque circle radii and distances to peripheral
circle to compare the characteristics of varying cam circle
radii.
The tracing of Harrisons drawing has the cams
210
coloured as in the graph with other calculated arcs in background
indicating the suitability of Harrison choice.
211
A
B
C
D
R
e
m
o
n
t
o
i
r
e

L
a
t
c
h
A
t
t
h
e
b
e
g
i
n
n
i
n
g
o
f
t
h
e
3
0
s
e
c
.
c
y
c
l
e
t
h
e
c
o
u
n
t
e
r
b
a
l
a
n
c
e
d
a
r
m
h
a
s
d
r
o
p
p
e
d
o
n
t
o
p
i
n
a
n
d
m
o
v
e
m
e
n
t
o
f
r
e
m
o
n
t
o
i
r
e
w
h
e
e
l
r
e
s
t
r
a
i
n
e
d
b
y
c
a
m
a
n
d
s
t
o
p
.
T
h
e
r
e
m
o
n
t
o
i
r
e
s
p
r
i
n
g
s
a
r
e
f
u
l
l
y
t
e
n
s
i
o
n
e
d
a
n
d
t
h
e
f
l
y
l
o
c
k
e
d
.
A
r
m
l
i
f
t
e
d
s
l
i
g
h
t
l
y
o
f
f
s
t
o
p
b
y
e
n
g
a
g
e
m
e
n
t
o
f
f
o
r
k
o
n
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
p
i
n
A
r
m
p
u
s
h
e
d
d
o
w
n
b
y
p
i
n
o
n
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
r
e
a
d
y
t
o
u
n
l
o
c
k
f
l
y
a
n
d
r
e
m
o
n
t
o
i
r
e
w
h
e
e
l
.
R
e
m
o
n
t
o
i
r
e
w
h
e
e
l
u
n
l
o
c
k
e
d
a
n
d
m
o
t
i
o
n
r
e
t
e
n
s
i
o
n
s
s
p
r
i
n
g
s
,
t
h
e
c
a
m
p
i
c
k
s
u
p
t
h
e
r
o
l
l
e
r
a
n
d
m
o
v
e
s
a
r
m
s
o
t
h
a
t
t
h
e
f
o
r
k
c
l
e
a
r
s
t
h
e
p
i
n
o
n
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
.
T
h
e
r
o
l
l
e
r
d
r
o
p
s
i
n
t
o
c
a
m
t
o
s
t
o
p
m
o
v
e
m
e
n
t
o
f
t
h
e
r
e
m
o
n
t
o
i
r
e
w
h
e
e
l
a
n
d
f
l
y
.
a
t
t
h
e
s
a
m
e
t
i
m
e
m
o
v
e
s
f
o
r
k
u
p
t
o
t
h
e
n
e
x
t
p
i
n
o
n
t
h
e
e
s
c
a
p
e
m
e
n
t
w
h
e
e
l
.
T
h
e
a
r
m
h
a
s
f
a
l
l
e
n
o
n
t
o
i
t
s
s
t
o
p
o
n
t
h
e
r
e
a
r
p
l
a
t
e
T
h
e
s
e
d
i
a
g
r
a
m
s
d
o
n
o
t
s
h
o
w
t
h
e
h
e
l
i
c
a
l
s
p
r
i
n
g
s
o
r
t
h
e
i
r
c
a
m
s
o
n
t
h
e
e
s
c
a
p
e
m
e
n
t
a
r
b
o
r
.
203.20 mm
8"
R17.460 mm
R22.62 mm
R42.300 mm
R35.880 mm
R76.200 mm
R92.440 mm
0 10 20 30 40 -10
C
A
B
D
G
E
H
F
Tracing of Harrison's drawing of the remontoire cams. This drawing almost certainly refers to the
RAS regulator, as the fixing points for the helical springs have exactly the same dimensions.
Construction lines and torque circles have been drawn for two different profiles. No indication is
given of which one Harrison preferred.
Scale 1:1


































































S
p
r
i
n
g

L
e
n
g
t
h
s

a
n
d

T
o
r
q
u
e

C
i
r
c
l
e

R
a
d
i
i

f
o
r

C
a
m

A
r
c
s
2
6
.
5
6
5
1
1
E
n
t
e
r

a
n
g
le

o
f

b
a
s
e
lin
e

t
o

h
o
r
iz
o
n
t
a
l
1
0
0
.
0
0
2
E
n
t
e
r

R
a
d
iu
s

o
f

P
e
r
ip
h
e
r
y
1
7
.
4
6
2
5
3
E
n
t
e
r

R
a
d
iu
s

a
c
t
io
n

c
ir
c
le
-
2
.
0
0
4

-
1
/
@
T
A
N
(
D
4
*
@
P
I
/
1
8
0
)
G
r
a
d
ie
n
t

o
f

lin
e

E
F







































































1
7
.
8
0
9
5
5

@
S
Q
R
T
(
D
6
*
D
6
/
(
1
+
D
7
*
D
7
)
)
x
P
o
s
it
io
n

o
f

G

a
t

in
t
e
r
c
e
p
t

o
f

a
c
t
io
n

c
ir
c
le

a
n
d

lin
e

E
F




2
-
1
5
.
6
1
9
6

(
+
E
3
7
*
E
4
2
)
+
E
3
8
y
3
5
.
7
1
8
8
7
E
n
t
e
r

R
a
d
iu
s

o
f

b
a
s
e

c
a
m

a
r
c

0
.
5
0
8
-
1
/
D
7
G
r
a
d
ie
n
t

o
f

lin
e

Q
R






































































2
7
.
8
7
0
9
@
S
Q
R
T
(
D
1
0
*
D
1
0
/
(
1
+
D
1
1
*
D
1
1
)
)
x
P
o
s
it
io
n

o
f

H

a
t

in
t
e
r
c
e
p
t

o
f

b
a
s
e

c
a
m

a
r
c

a
n
d

lin
e

Q
R
3
1
3
.
9
3
1
0

+
D
1
2
*
D
1
1
y
1
.
4
7
3
3
1
1
(
+
D
1
3
-
D
9
)
/
(
D
1
2
-
D
8
)
G
r
a
d
ie
n
t

o
f

lin
e

G
H




4
-
2
7
.
1
2
4
3
1
2
+
G
7
-
G
1
2
*
G
6
c
I
n
t
e
r
c
e
p
t

o
f

lin
e

G
H


c
=
y
-
m
x
5
5
.
8
3
1
3

@
A
T
A
N
(
+
E
1
4
)
*
1
8
0
/
@
P
I
A
n
g
le

o
f

lin
e

G
H














1
0
0
.
8
3
1
4

@
A
T
A
N
(
+
E
1
4
)
*
1
8
0
/
@
P
I
A
n
g
le

o
f

lin
e

G
I















1
0
.
8
3
1
5
+
E
1
5
-
4
5
A
n
g
le

o
f

lin
e

G
J















-
5
.
2
2
6
0
1
6
@
T
A
N
(
+
E
1
6
*
@
P
I
/
1
8
0
)
G
r
a
d
ie
n
t

o
f

lin
e

G
I












0
.
1
9
1
3
1
7
@
T
A
N
(
+
E
1
7
*
@
P
I
/
1
8
0
)
G
r
a
d
ie
n
t

o
f

lin
e

G
J











5
4
.
6
9
5
4
9
.
0
7
0
4
3
.
4
4
5
3
7
.
8
2
0
3
2
.
1
9
5
2
6
.
5
7
0
2
0
.
9
4
5
1
5
.
3
2
0
9
.
6
9
5
4
.
0
7
0
1
8
D
e
t
e
r
m
in
e

e
q
u
a
t
io
n

o
f

lin
e
s

2
.




E
n
t
e
r

a
n
g
le

t
o

h
o
r
iz
o
n
t
a
l
a
x
is
6
1
.
4
1
2
1
1
.
1
5
3
2
0
.
9
4
7
1
0
.
7
7
6
2
0
.
6
2
9
6
0
.
5
0
0
1
0
.
3
8
2
8
0
.
2
7
3
9
0
.
1
7
0
8
0
.
0
7
1
2
1
9
@
T
A
N
(
+
D
2
8
*
@
P
I
/
1
8
0
)

m
G
r
a
d
ie
n
t



2
.
9
9
4
0
2
.
3
2
9
9
1
.
8
9
7
1
1
.
6
0
2
5
1
.
3
9
6
4
1
.
2
5
0
1
1
.
1
4
6
5
1
.
0
7
5
0
1
.
0
2
9
2
1
.
0
0
5
1
2
0
+
1
+
D
2
9
*
D
2
9
a
F
in
d

in
t
e
r
c
e
p
t

o
f

c
a
m

a
r
c

a
n
d

lin
e


2
8
.
4
9
1
7
2
0
.
4
0
4
9
1
3
.
9
6
7
8
8
.
6
2
9
1
4
.
0
4
8
8
0
.
0
0
3
4
-
3
.
6
6
2
2
-
7
.
0
6
1
5
-
1
0
.
2
8
2
2
-
1
3
.
3
9
6
2
2
1
-
2
*
(
$
D
$
8
+
D
2
9
*
$
D
$
9
)
b
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
-
9
7
0
.
8
9
3
8
2
2
(
+
$
D
$
8
*
$
D
$
8
)
+
(
$
D
$
9
*
$
D
$
9
)
-
(
$
D
$
1
0
*
$
D
$
1
0
)
c
1
3
.
8
6
7
6
1
6
.
4
9
9
0
1
9
.
2
3
8
8
2
2
.
0
6
8
4
2
4
.
9
5
8
2
2
7
.
8
6
7
1
3
0
.
7
4
1
2
3
3
.
5
1
5
1
3
6
.
1
1
3
0
3
8
.
4
5
1
4
2
3
(
-
D
3
1
+
@
S
Q
R
T
(
D
3
1
*
D
3
1
-
4
*
D
3
0
*
D
3
2
)
)
/
(
2
*
D
3
0
)
x
1
9
.
5
8
2
4
1
9
.
0
2
6
8
1
8
.
2
2
1
9
1
7
.
1
3
0
3
1
5
.
7
1
4
0
1
3
.
9
3
6
5
1
1
.
7
6
6
6
9
.
1
8
1
3
6
.
1
6
9
7
2
.
7
3
6
0
2
4
+
D
3
3
*
D
2
9
y
2
3
.
9
9
5
4
2
5
.
1
8
4
0
2
6
.
4
9
8
5
2
7
.
9
3
6
8
2
9
.
4
9
3
1
3
1
.
1
5
7
7
3
2
.
9
1
6
2
3
4
.
7
5
0
0
3
6
.
6
3
6
2
3
8
.
5
4
8
6
2
5
@
S
Q
R
T
(
D
3
3
*
D
3
3
+
D
3
4
*
D
3
4
)
T
o
r
q
u
e

c
ir
c
le

r
a
d
iu
s






































7
5
.
8
1
0
5
3
.
9
8
7
1
2
.
9
6
0
9
2
.
2
9
6
8
1
.
8
2
7
1
1
.
4
7
3
5
1
.
1
9
4
2
0
.
9
6
4
8
0
.
7
6
9
8
0
.
5
9
9
0
2
6
(
+
D
3
4
-
$
D
$
9
)
/
(
D
3
3
-
$
D
$
8
)
G
r
a
d
ie
n
t

o
f

lin
e

t
o

c
a
m

c
e
n
t
r
e

-
0
.
1
7
2
1
-
0
.
2
5
0
8
-
0
.
3
3
7
7
-
0
.
4
3
5
4
-
0
.
5
4
7
3
-
0
.
6
7
8
6
-
0
.
8
3
7
4
-
1
.
0
3
6
5
-
1
.
2
9
9
0
-
1
.
6
6
9
4
2
7
-
1
/
D
3
6
m
F
o
r
m
u
la

f
o
r

lin
e

t
o

p
e
r
ip
h
e
r
a
l
c
ir
c
le
2
1
.
9
7
2
3
.
1
6
2
4
.
7
2
2
6
.
7
4
2
9
.
3
7
3
2
.
8
5
3
7
.
5
1
4
3
.
9
2
5
3
.
0
8
6
6
.
9
3
2
8
+
E
3
4
-
E
3
7
*
E
3
3
c
1
.
0
2
9
6
1
.
0
6
2
9
1
.
1
1
4
1
1
.
1
8
9
6
1
.
2
9
9
5
1
.
4
6
0
6
1
.
7
0
1
2
2
.
0
7
4
4
2
.
6
8
7
4
3
.
7
8
6
9
2
9
+
1
+
E
3
7
*
E
3
7
a
I
n
t
e
r
c
e
p
t

o
n

p
e
r
ip
h
e
r
a
l
c
ir
c
le

8
-
7
.
5
6
1
8
-
1
1
.
6
2
0
0
-
1
6
.
6
9
7
5
-
2
3
.
2
8
4
0
-
3
2
.
1
5
3
1
-
4
4
.
5
8
4
5
-
6
2
.
8
1
6
4
-
9
1
.
0
4
7
2
-
1
3
7
.
9
0
4
3
-
2
2
3
.
4
5
8
5
3
0
+
2
*
E
3
7
*
E
3
8
b
-
9
5
1
7
.
3
6
-
9
4
6
3
.
3
9
-
9
3
8
8
.
9
4
-
9
2
8
5
.
0
3
-
9
1
3
7
.
1
8
-
8
9
2
0
.
9
9
-
8
5
9
3
.
1
3
-
8
0
7
1
.
0
3
-
7
1
8
2
.
4
4
-
5
5
2
0
.
7
4
3
1
+
E
3
8
*
E
3
8
-
1
0
0
0
0
c
-
9
2
.
5
4
1
5
-
8
9
.
0
4
9
5
-
8
4
.
6
1
3
5
-
7
9
.
1
0
1
7
-
7
2
.
3
8
8
1
-
6
4
.
3
6
6
9
-
5
4
.
9
6
8
6
-
4
4
.
1
7
8
9
-
3
2
.
0
5
6
7
-
1
8
.
7
4
8
8
3
2
(
-
E
4
0
-
@
S
Q
R
T
(
E
4
0
*
E
4
0
-
4
*
E
3
9
*
E
4
1
)
)
/
(
2
*
E
3
9
)
x
3
7
.
8
9
5
5
4
5
.
4
9
9
4
5
3
.
2
9
6
9
6
1
.
1
7
9
5
6
8
.
9
9
2
4
7
6
.
5
3
0
4
8
3
.
5
3
7
2
8
9
.
7
1
1
9
9
4
.
7
2
2
6
9
8
.
2
2
6
7
3
3
(
+
E
3
7
*
E
4
2
)
+
E
3
8
y
1
0
7
.
9
7
3
5
1
0
8
.
8
1
7
6
1
0
9
.
6
1
5
5
1
1
0
.
3
4
3
6
1
1
0
.
9
7
2
6
1
1
1
.
4
6
7
9
1
1
1
.
7
9
0
8
1
1
1
.
8
9
9
7
1
1
1
.
7
5
3
0
1
1
1
.
3
1
1
9
3
4
@
S
Q
R
T
(
(
E
4
2
-
E
3
3
)
*
(
E
4
2
-
E
3
3
)
+
(
E
4
3
-
E
3
4
)
*
(
E
4
3
-
E
3
4
)
)
d
D
is
t
a
n
c
e

t
o

p
e
r
ip
h
e
r
a
l
c
ir
c
le











9
6
.
7
1
3
0
3
5
@
S
Q
R
T
(
(
+
D
1
0
*
D
1
0
)
/
(
1
+
D
1
9
*
D
1
9
)
)
x
R
e
la
t
iv
e

C
o
o
r
d
in
a
t
e

o
f

I



5
-
3
5
.
0
8
2
3
3
6
+
E
2
0
*
E
1
8
y
1
.
0
9
6
5
3
7
+
F
8
-
F
2
0
x
A
b
s
o
lu
t
e

C
o
o
r
d
in
a
t
e
s

o
f

I

1
9
.
4
6
3
4
3
8
+
F
9
-
F
2
1

y
4
2
.
4
1
8
3
9
R
E
n
t
e
r

r
a
d
iu
s

o
f

c
a
m

a
r
c
3
.
1
7
0
5
4
0
+
1
+
G
1
2
*
G
1
2
a
D
e
t
e
r
m
in
e

c
e
n
t
r
e

o
f

c
a
m

a
r
c

o
n

lin
e

G
H
-
1
3
9
.
4
6
4
1
4
1
+
2
*
(
(
G
1
5
-
G
4
1
)
*
G
1
4
)
-
2
*
G
4
0
b
3
7
2
.
3
2
5
6
4
2
(
G
3
9
-
G
1
3
)
*
(
G
3
9
-
G
1
3
)
+
G
3
8
*
G
3
8
-
G
4
0
*
G
4
0
c
2
.
8
5
5
0
4
3
(
-
G
4
4
-
@
S
Q
R
T
(
G
4
4
*
G
4
4
-
4
*
G
4
3
*
G
4
5
)
)
/
(
2
*
G
4
3
)
x
-
2
2
.
9
1
8
2
4
4
+
G
1
4
*
G
4
6
+
G
1
5
y
5
4
.
6
9
5
4
9
.
0
7
0
4
3
.
4
4
5
3
7
.
8
2
0
3
2
.
1
9
5
2
6
.
5
7
0
2
0
.
9
4
5
1
5
.
3
2
0
9
.
6
9
5
4
.
0
7
0
1
8
D
e
t
e
r
m
in
e

e
q
u
a
t
io
n

o
f

lin
e
s

2
.




E
n
t
e
r

a
n
g
le

t
o

h
o
r
iz
o
n
t
a
l
a
x
is
1
.
4
1
2
1
1
.
1
5
3
2
0
.
9
4
7
1
0
.
7
7
6
2
0
.
6
2
9
6
0
.
5
0
0
1
0
.
3
8
2
8
0
.
2
7
3
9
0
.
1
7
0
8
0
.
0
7
1
2
1
9
@
T
A
N
(
+
D
2
8
*
@
P
I
/
1
8
0
)

m
G
r
a
d
ie
n
t



2
.
9
9
4
0
2
.
3
2
9
9
1
.
8
9
7
1
1
.
6
0
2
5
1
.
3
9
6
4
1
.
2
5
0
1
1
.
1
4
6
5
1
.
0
7
5
0
1
.
0
2
9
2
1
.
0
0
5
1
2
0
+
1
+
D
2
9
*
D
2
9
a
F
in
d

in
t
e
r
c
e
p
t

o
f

c
a
m

a
r
c

a
n
d

lin
e



G
+
5
9
.
0
1
5
0
4
7
.
1
4
9
0
3
7
.
7
0
3
5
2
9
.
8
7
0
0
2
3
.
1
4
9
1
1
7
.
2
1
3
1
1
1
.
8
3
4
5
6
.
8
4
6
6
2
.
1
2
0
9
-
2
.
4
4
8
5
2
1
-
2
*
(
$
D
$
8
+
D
2
9
*
$
D
$
9
)
b
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
-
1
2
6
5
.
8
9
3
8
2
2
(
+
$
D
$
8
*
$
D
$
8
)
+
(
$
D
$
9
*
$
D
$
9
)
-
(
$
D
$
1
0
*
$
D
$
1
0
)
c
1
2
.
9
4
6
7
1
5
.
2
9
2
5
1
7
.
7
4
0
1
2
0
.
2
9
0
9
2
2
.
9
4
0
0
2
5
.
6
7
3
4
2
8
.
4
6
5
8
3
1
.
2
7
8
2
3
4
.
0
5
6
0
3
6
.
7
2
8
7
2
3
(
-
D
3
1
+
@
S
Q
R
T
(
D
3
1
*
D
3
1
-
4
*
D
3
0
*
D
3
2
)
)
/
(
2
*
D
3
0
)
x
1
8
.
2
8
1
9
1
7
.
6
3
5
5
1
6
.
8
0
2
4
1
5
.
7
5
0
6
1
4
.
4
4
3
3
1
2
.
8
3
9
5
1
0
.
8
9
5
7
8
.
5
6
8
5
5
.
8
1
8
2
2
.
6
1
3
4
2
4
+
D
3
3
*
D
2
9
y
2
2
.
4
0
1
9
2
3
.
3
4
2
4
2
4
.
4
3
4
2
2
5
.
6
8
6
7
2
7
.
1
0
8
2
2
8
.
7
0
5
0
3
0
.
4
7
9
8
3
2
.
4
3
0
6
3
4
.
5
4
9
5
3
6
.
8
2
1
5
2
5
@
S
Q
R
T
(
D
3
3
*
D
3
3
+
D
3
4
*
D
3
4
)
T
o
r
q
u
e

c
ir
c
le

r
a
d
iu
s






































4
.
0
8
2
6
3
.
2
6
0
6
2
.
6
6
8
5
2
.
2
1
7
8
1
.
8
6
0
2
1
.
5
6
7
1
1
.
3
2
0
3
1
.
1
0
7
8
0
.
9
2
1
0
0
.
7
5
3
7
2
6
(
+
D
3
4
-
$
D
$
9
)
/
(
D
3
3
-
$
D
$
8
)
G
r
a
d
ie
n
t

o
f

lin
e

t
o

c
a
m

c
e
n
t
r
e

-
0
.
2
4
4
9
-
0
.
3
0
6
7
-
0
.
3
7
4
7
-
0
.
4
5
0
9
-
0
.
5
3
7
6
-
0
.
6
3
8
1
-
0
.
7
5
7
4
-
0
.
9
0
2
7
-
1
.
0
8
5
8
-
1
.
3
2
6
7
2
7
-
1
/
D
3
6
m
F
o
r
m
u
la

f
o
r

lin
e

t
o

p
e
r
ip
h
e
r
a
l
c
ir
c
le
2
1
.
4
5
2
2
.
3
3
2
3
.
4
5
2
4
.
9
0
2
6
.
7
8
2
9
.
2
2
3
2
.
4
6
3
6
.
8
0
4
2
.
8
0
5
1
.
3
4
2
8
+
E
3
4
-
E
3
7
*
E
3
3
c
1
.
0
6
0
0
1
.
0
9
4
1
1
.
1
4
0
4
1
.
2
0
3
3
1
.
2
8
9
0
1
.
4
0
7
2
1
.
5
7
3
7
1
.
8
1
4
9
2
.
1
7
8
9
2
.
7
6
0
2
2
9
+
1
+
E
3
7
*
E
3
7
a
I
n
t
e
r
c
e
p
t

o
n

p
e
r
ip
h
e
r
a
l
c
ir
c
le

-
1
0
.
5
0
9
6
-
1
3
.
6
9
4
2
-
1
7
.
5
7
5
9
-
2
2
.
4
5
5
0
-
2
8
.
7
8
8
4
-
3
7
.
2
9
6
4
-
4
9
.
1
6
4
5
-
6
6
.
4
4
5
6
-
9
2
.
9
3
1
2
-
1
3
6
.
2
3
6
4
3
0
+
2
*
E
3
7
*
E
3
8
b
-
9
5
3
9
.
7
6
-
9
5
0
1
.
5
7
-
9
4
5
0
.
0
8
-
9
3
7
9
.
9
9
-
9
2
8
3
.
0
7
-
9
1
4
6
.
0
3
-
8
9
4
6
.
6
2
-
8
6
4
5
.
5
0
-
8
1
6
8
.
5
7
-
7
3
6
3
.
9
3
3
1
+
E
3
8
*
E
3
8
-
1
0
0
0
0
c
-
9
0
.
0
3
9
3
-
8
7
.
1
4
3
2
-
8
3
.
6
4
9
4
-
7
9
.
4
5
1
2
-
7
4
.
4
2
7
8
-
6
8
.
4
4
8
6
-
6
1
.
3
8
0
5
-
5
3
.
1
0
0
0
-
4
3
.
5
1
0
8
-
3
2
.
5
6
5
7
3
2
(
-
E
4
0
-
@
S
Q
R
T
(
E
4
0
*
E
4
0
-
4
*
E
3
9
*
E
4
1
)
)
/
(
2
*
E
3
9
)
x
4
3
.
5
0
7
7
4
9
.
0
5
1
7
5
4
.
7
9
7
6
6
0
.
7
2
4
9
6
6
.
7
8
7
0
7
2
.
9
0
2
6
7
8
.
9
4
5
8
8
4
.
7
3
7
2
9
0
.
0
3
7
8
9
4
.
5
4
8
8
3
3
(
+
E
3
7
*
E
4
2
)
+
E
3
8
y
1
0
6
.
0
3
0
5
1
0
7
.
1
4
5
0
1
0
8
.
2
7
4
9
1
0
9
.
4
1
2
9
1
1
0
.
5
4
5
7
1
1
1
.
6
5
3
6
1
1
2
.
7
0
8
3
1
1
3
.
6
7
2
1
1
1
4
.
4
9
7
0
1
1
5
.
1
2
5
3
3
4
@
S
Q
R
T
(
(
E
4
2
-
E
3
3
)
*
(
E
4
2
-
E
3
3
)
+
(
E
4
3
-
E
3
4
)
*
(
E
4
3
-
E
3
4
)
)
d
D
is
t
a
n
c
e

t
o

p
e
r
ip
h
e
r
a
l
c
ir
c
le











C
a
m
D
e
ta
ils
N
e
w
B
.Q
P
w
Chapter 8
Integrator and Case
Now if there were Clocks made as described above, (of Which
I have had experience, even such Clocks as be made to go to 2 or 3
Seconds in a Year) and fixd at Sundry Ports in the World, where
ships resort, or each clock made (but however its Cycloid etc., corrected)
at the Port where it is to be fixd (because nearer the Equator
Pendulums Oscillate slower) these would good standards to set the Sea
Clocks by; when the Ships are ready to sail.
John Harrison, 1730 manuscript
Seconds
The seconds are indicated by four diametrically
opposed blued steel pointers which are attached to a brass
collet fixed to the extremity of the escapement wheel arbor.
Each pointer describes a complete circle moving anticlockwise
in four minutes and appears to make one second discrete
movements. The pointers are seen through an aperture in the
front dial plate. The seconds pointer is thus protected from
any action which might stop its movement which would allow
both pallets to fall away from the escapement wheel resulting
in runaway.
Motion Work
The centre wheel arbor extends through the front dial
plate and the hour pinion fits on this extension. A bent brass
spring rests against a
collar on the centre
wheel pinion. The
minute hands are
fitted with two
tapered pins to a
square section at the
distal end of the
minute pinion. A
brass collet domed underneath fits over the centre of the
minute hands and the centre arbor. A pin through the centre
arbor tensions the spring.
The motion work is all anterior to the front plate and
the gearing is opposite to that of the power system of the
215
clock. To fulfil these different parameters the pitch of the
wheels is different and the lantern pinions have fixed rollers.
Torque is transferred after the line of centres.
The minute pinion drives a engraved and silvered hour
wheel which is viewed through a semicircular aperture in the
front dial plate. There is a central fiducial pointer on the
upper arc of the aperture. The hour wheel is attached to a
brass ferrule which has a lignum vitae bush running on a brass
arbor fixed to the front plate. Gould was able to remove the
arbors of the motion work from the front plate and it is
assumed that they screw into the plate. The other end of the
ferrule is attached to the first drive wheel for the day wheel.
A domed collet with tapered pin holds the combined wheels
to their arbor.
Day Wheel
All of Harrisons domestic pendulum clocks and the
first sea clock are fitted with day wheels. In the RAS
regulator the annular engraved and silvered day wheel is
supported by two grooved friction wheels at the bottom of
the clock equidistant from the centre and a sprung guide
fixed to the escapement arbor accessory plate at the top.
The day drive disc with its single pin and attached
wheel has the same construction as the hour wheel. All the
fixed lantern rollers of the motion work have the same
diameter and all the wheels have the same form and pitch.
Harrison would have divided his day wheel into 31
segments by marking off a chord of 11.61E using a segment.
216
He would then set his dividers to this chord and mark the 31
segments around the circle. Unless he was extremely lucky he
would not be able to divide the circle accurately with this
setting of his dividers. It is impossible to readjust the
dividers to 1/31 part of the error. Therefore the circle is
slightly enlarged or reduced according to the error and
divided using the original setting of the dividers. Circles are
adjusted and drawn until the circumference can be divided
accurately. Radial lines are drawn to the points on this
circle which cut a circle of the desired diameter.
Case
The King portrait shows a glass fronted case revealing
the pendulum. As previously described the clock has a
carefully defined and proportioned shape which should not be
hidden. A case is necessary by providing insulation to
minimise sudden temperature changes and avoid draughts.
Gould made accurate measurements of the RAS regulator for
a case suitable for the RASs premises in Burlington House.
His case dimensions are too large for the modern domestic
environment; suggested dimensions and form are given.
X X
217
1
8
8
1
.
7

m
m
486.3 mm
623.1 mm
275.3 mm
343.7 mm
4
5
5
.
3

m
m
461.2 mm
557.7 mm
1
3
7
0
.
5

m
m
38.1 mm
25.4 mm
20.0 mm
Case
The drawing for a case is derived from the style in King's portrait. The front is fully glazed. The
backboard is firmly fixed to a wall but has an aperture to allow the pendulum support to be fixed to the
wall without any soft intervening material. The case frame is attached to the backboard with six brass
removable dowels. The case is necessary to protect the regulator from draughts and sudden changes in
temperature. Gould made a case 86" x 20.7" (2184 x 525.8 mm.).
Case.tcw/1
Scale 1:12
Chapter 9
Adjustment and Errors
There being to be concerned in that Proceeding, [viz. as at
first to get such a clock to be really or truly adjusted], four different
Things, and wherein two of which do as it were to, or in the same
Purpose, viz. the Composers of the Pallats to relative Rest, and the
correspondent Curvature thereto of the Cycloid Cheeks. I say, these two
Things may only as almost be taken at one, viz. in their joint Effects,
for so far as belongs to this Matter, but not quite so; and the other two
particulars in the Case, are the adjusting [viz. by a proper provision]
the Wires, or the redoubling of the Wires for Heat and Cold in the
Pendulum, and the Screw for fast and slow in general at the Bottom of
the Pendulum. And as hence I may say, as Rectifications in these
different Things must pertain to the Clocks adjusting, (viz. as at first
by the Master Workman) it is or must as therefore be a Thing - quite
impossible to do to Exactitude or Perfection, by or from any celestial
Observations whatever,, no, nor indeed by or from any other Means
whatever, unless [as here meaning of Course] the Foundation and
principal Parts, or rather Principles of the Clock be to the Purpose as
I have shewn, and could as thence still farther shew.
John Harrison, Concerning Such Mechanisms, 1775
Harrison regulated and adjusted his 1728 regulators
by comparing the going of two similar clocks. He provided
adjustment of the escapement, remontoire and pendulum by:-
I. The weight of the pallets is counterbalanced by the
lead button on the right side of the frame. The pallet
frame without the pallets, composers and lead button
balances on the knife edge arbor.
II. The torque exerted by the escapement is closely
regulated and adjusted by the helical spring
attachment cams on the periphery of the remontoire
wheel. (See description of Harrisons remontoire
drawing and manuscript description in remontoire
section.)
III. The beat can be adjusted by the annular fixing at the
top of the crutch. The angle of the pallet frame
219
relative to the knife edges of the pallet frame arbor is
fixed by pinning.
IV. Escapement error can be adjusted by altering the mass
of the brass drive pin on the crutch.
V. The distance between the active contact points of the
pallets can be lengthened or shortened by adjusting
the composers. The angle of incidence of the entry
pallet face is less than perpendicular to the pallet
tooth thus allowing fine adjustment by positioning
the composer. The angle of incidence of the exit
pallet face is more perpendicular to the escapement
tooth than the entry pallet face so that even finer
adjustment is possible.
VI. Adjustment of the regulating screw supporting the
pendulum bob.
VII. Adjustment of the suspension cheeks.
VIII. Adjustment of the temperature compensation of the
pendulum.
I have two ways for trying the truth of the Motion, the which
together is very Compleat. One is by the apparent Motion of the fixed
Stars with a very large sort of an Instrument, of about 25 Yards
Radius, cornposd of the East side of my Neighbours Chimney (which
is situated from my House towards the South) and the West side of an
exact place of some one of the upright parts of my own Window
Frames; by which the Rays of a Star are taken from my sight almost in
an Instant: And I have another person to count the seconds of the Clock,
beginning a little before the Star Vanish: So I observe what second is
Mentiond when it Vanisheth; and I have a Table Calculated to show
how much sooner any such Star is to Vanish every Night, or before the
24 Hours of the Pendulum Day is expired.
But my other way is the better part of the Completion: and
that is the two Clocks placd one in one Room and the other in another,
yet so, that I can stand in the Doorstead, and hear the beats of both the
Pendulums, when the Clock Case heads are of, and before or after the
hearing can see the seconds of one Clock, whilst another Person count
the seconds of the other: by which Means I can have the difference of the
Clocks to a small part of a second. And in very Cold and Frosty
Weather, I sometimes make one Room very warm, with a great Fire,
220
whilst the other is very Cold. And again the contrary. And sometimes
the like in Summer by the Suns Rays in at the Windows of one Room,
and also a Fire, whilst the other is close shut up and Cool. Thus I
prove the Operation of the Pendulum Wires, and adjust the same to
what is treated of above. And to prove or adjust the Cycloid to
Vibrations performd in different Arches as required I cause the
Pendulum to describe such by increasing and decreasing the draught of
the Wheels, and that by adding to and taking from the Weight: by which
I can make 8 or 10 Times more difference, than Nature ever will, and
yet the effect be nearly the same as if Nature itself had alterd the
Weight of the Air so much.
John Harrison 1730 Document
It is now much easier and quicker to adjust the
regulator using modern radio controlled clocks. The rising
transition of the seconds pulse to the solenoid of a cheap
analogue MSF or DCF regulated radio clock is accurate to
10-100 milliseconds UTC time. A temperature controlled
quartz crystal or rubidium oscillator can be disciplined with
the L1 GPS radio transmissions to provide timing pulses
accurate to better than 10 nanoseconds. These timing signals
can then be compared to positional signals from the
pendulum on a dual trace oscilloscope and the data saved
using a data-logger attached to a PC. The pendulum signals
are derived using a polarised 2 mW red laser beam focussed
onto a silicon photo-transistor driving a cascode circuit to
improve rise and fall times to 3 s. The laser beam is
reflected by a small mirror on the rear of the pendulum bob.
Signals are recorded for each period which can then be
analysed to determine changes in amplitude and velocity. A
1/100th. second electronic stopwatch is helpful particularly
in the early stages of setting up.
The regulator is setup in the following order:-
I. Adjust composers so that both pallets are in contact
with teeth on the escapement wheel at entry and exit.
II. Adjust the crutch to midline so that it is in beat.
III. Apply torque to the escapement wheel so that the
crutch begins to vibrate.
IV. Reduce torque to the escapement wheel until the
221
crutch ceases to vibrate.
V. Determine the natural period of the pendulum and
adjust to be as near to 1 second as possible.
VI. Drive the pendulum with the escapement and
determine the period of the driven pendulum and
compare it to the undriven thereby finding the
escapement error.
VII. Reduce the escapement error as much as possible by
altering the mass of the crutch pin.
VIII. Alter the mass of the pallet frame lead counterweight,
alter the applied torque to the escapement wheel,
drive the pendulum and determine whether a smaller
escapement error can be attained. Repeat until the
smallest escapement error is achieved.
IX. Run the clock in its case with a closely controlled
ambient temperature. Adjust the rating screw so that
the clock is within 1 second per day to UTC time.
X. Lower the ambient temperature to the lowest that
can be accurately maintained and rate the clock to an
accuracy of at least 100 milliseconds over a period of
at least seven days.
XI. Raise the ambient temperature to a closely
maintained 35EC and measure the rate.
XII. Alter the pendulum temperature compensation by
moving the pins until minimum alteration in rate is
attained. Because of the chaotic nature of pendulum
motion it is important to check the compensation
over at least four temperature cycles without
stopping the clock.
XIII. Adjust the rate of the clock in a temperature
controlled enclosure without stopping it using the
capstan screws acting on the curved cheeks
supporting the pendulum. Care has to be taken to
adjust each screw equally so that the beat is not
altered.
Because the rotation of the earth produces a
222
centrifugal force which varies with latitude it shapes the earth
into an oblate spheroid. The distance to the centre from the
surface varies and thereby alters the value of g. In the same
way the value of g is dependent upon the height above sea
level. Therefore the regulator has to be adjusted in the
position where it is to be permanently operated. There must
be no motion of the support; this can only be achieved by a
non-elastic support of great mass.
A firm Suspension of the Pendulum to be, as in the first
Place, made sure of, viz. from a firm Wall, as not having, or to have
any Dependence from the Clock, the Clock-Case, or any Wainscot; for
as without that, and as hitherto in common, all must be no better than
Chance, as with respect to any true Performance of a clock.
John Harrison, Concerning Such Mechanism, 1775
The difference in length between the theoretical
pendulum and the practical pendulum caused by the
distribution of mass, the added air mass, the finite amplitude
correction for the required amplitude, and the dynamic and
static stretching of the suspension and rods are corrected by
adjustment of the rating screw.
But a Pendulum moving in a Cycloid to Regulate a Clock
true, which is not in Vacuo, must not always keep its same Length,
unless the Air gave always the same resistance; which agrees both with
reason and experiment, viz. the Pendulum must be rather shorter in
Warm than in Cold Weather, which is contrary to the operation of the
wire.
John Harrison, 1730 Document
In this statement Harrison considers that air
resistance is determined by the temperature of the ambient
air. There is no reference to the effect of changes in
barometric pressure on the pendulum in Harrisons writings.
Barometric pressure changes alter the rate of the clock by
change of buoyancy of the pendulum bob and by altering the
viscosity of the ambient air.
By Archimedes principle the effective weight of the
bob is reduced by the buoyancy of the air it displaces, while
the mass is unaltered. The effective gravitational force is
decreased resulting in an increased period.
223
The National Physics Laboratory has records of the
barometric pressure at Teddington. A histogram of the
incidence of pressure shows a skewness (-0.579) towards
high pressure. Errors caused by changes in the barometric
pressure will not average out, but they can be minimised by
careful rating. This skewness in the errors caused by
barometric pressure is the probably largest factor producing
the drift observed by Harrison.
Yet from several observations, I still endeavourd to make further
corrections in this Motion; and in these three last Years have brought a
clock to go nearer the truth than can be well imagind considering the
224
vast Number of seconds of Time there is in a Month, in which space of
time it does not vary above one second, and that mostly the way I
expect; So I am sure I can bring it to the Nicety of 2 or 3 seconds in a
Year. And twill also continue this exactness for 40 or 50 years or
more; however so as not to vary above 2or 3 seconds from what it did
the Year next before; for twill not want Cleaning, and the little it
wears can but alter it insensibly little. This Nicety is owing partly to
the Matter the Clock is made of, partly to the Contrivance it is made
with, and partly to the Nice observations it is tryd by, and the
convenient place it stands in.
However analysis of the recordings of 2009 and
other years show a small correlation of temperature with
barometric pressure. A second order polynomial regression
curve has been drawn through the scatter plot of the monthly
averages of 2009 to illustrate this, indicating that the
barometric errors can be lessened by temperature over-
compensation.
The records of barometric pressure and temperature
records of the Meteorological Office show relatively slow
changes in pressure and rapid and frequent changes in
temperature. As demonstrated previously there is hysteresis
in the response of the supporting rods to a change in
temperature. This of course occurs whether the temperature
is rising or falling and will only introduce significant error is
the change in temperature is frequent and precipitous. It is
beneficial to keep the regulator in a case and to limit changes
of temperature due to radiation.
It may be that this correlation between barometric
225
pressure and temperature is enabling clocks, such as the later
Dent and the Riefler regulators to compensate for
temperature changes in air viscosity. The pendulum of the
Shortt clock, working in an evacuated metal chamber at a
constant maintained pressure of 20mm Hg. (26.7 mbar), is
effectively insulated from temperature changes (cp. Dewar
flask).
Whenas, by a right use of my Clocks [viz. as when with
proper Conveniences, and proper Apparatusses] even that Piece of
Astronomy [the Equation] may hereafter be corrected, because as thence
the Eccentricity of the Earths Orbit, and whatever else may pertain to
the Matter, may be the better known.
John Harrison, Concerning Such Mechanism, 1775
The Shortt master pendulum was compared to a
Caesium atomic oscillator in1985 (Pierre H. Boucheron, Just
How Good was the Shortt Clock, Horological Journal, December
1985, pp 10-14). Tidal changes in g were clearly
demonstrated; preliminary results using similar methods with
Harrisons pendulum and escapement are comparable
vindicating the above quotation from CSM.
226
Chapter 10
Conclusions
The regulator that John Harrison wished to build for
the Royal Greenwich Observatory would have been the
culmination of his life of innovation and study. He was
fortunate to have been born into a literate household and to
have served an apprenticeship. His introduction to
Newtonian mechanics from reading notes of Saundersons
lectures was probably the beginning of his interest in
mathematics and physics which was almost certainly pursued
at Gresham College and in the London coffee houses. His
writings, which have a remarkably large vocabulary for the
period in which they were written, give us a clear insight into
his logical thinking. The design of his early wooden
pendulum clocks and his sea clocks, the choice of a coach
clock form for H4, together with the capitalisation of nouns
in his documents indicate that he was influenced by German
horologists. He considered that his development of a balance
wheel as the oscillator for a marine timepiece and the
development of his pendulum escapement were his two most
notable achievements. Unfortunately he does not give full
details of the escapement in his writings. An understanding
of the escapement can only be obtained by studying his
remaining artefacts in the light of his ideas.
His predicted accuracy of two to three seconds in a
year is so close what is now considered to be the maximum
possible in a pendulum clock performing in free air, that it is
pertinent to examine his methods and ideas.
Harrison designed his ultimate pendulum clock to have:-
< A pendulum having large amplitude with a light bob.
< Closely controlled amplitude.
< Correction of pendulum rod length for changes in
ambient temperature.
< Correction of the changes in air viscosity caused by
varying temperature by altering the length of the
pendulum rod.
227
< Firm fixing of the pendulum suspension to a heavy
mass to minimise resonance .
< Excessively heavy framework and light balanced
wheels to reduce noise (F.R.Moon & P.D.Stiefel Co-
existing Chaotic and Periodic Dynamics in Clock Escapements
Phil. Trans. Roy. Soc. (2006) v.364 pp. 2539-2564)
< Escapement error correction.
< Use of large wheels and pinions to minimise power
transmission fluctuations.
< Use of a remontoire to enable delivery of a smooth
and carefully regulated torque to the escapement
wheel.
< Use of friction wheels and roller bearings to minimise
and regulate frictional losses.
< Use of large rotating rollers on the lantern pinions to
reduce frictional losses.
< Use of materials that did not require oil for
lubrication. Lubricating oils in the eighteenth
century readily oxidised and thickened (rancid). The
problem was solved in the nineteenth century by the
use of whale oil.
< Use of spring driven maintaining power whilst the
clock is being rewound.
< Assessment of ageing effects.
Long and tedious experimentation together with his
artistic talent produced a clock of great beauty and efficient
design, which may still be used to illustrate basic
mathematical and physical principles. The following detailed
drawings have been derived from these principles and the
accuracy with which they compare to the RAS regulator is a
testament to Harrisons technical skills. It was only
considered that the correct sequence of steps had been
determined to construct copies of the drawings on plain paper
when they could be constructed without reference to
Harrisons whilst fulfilling the physical principles.
Examination of his work and drawings reveals his
extensive knowledge of mathematics. In the design of the
plates of the RAS plates he uses irrational numbers including
the quadratrix of Hippias. The quadratrix pictorially
demonstrates the irrationality of B, and can be used to divide
228
an angle into any integral parts. It is a demonstration of how
to trisect an angle.
His drawing of the asymmetrical escapement is
remarkable for his ability to determine the unequal lengths of
the pallets which will both be in contact with a tooth of the
escapement wheel at each and equal extremity of the
pendulums motion. That this is possible can be proven by
computation. His drawing of the remontoire cam is notable
because of his choice of an inclination of the base axis of
26.565E (arctan 0.5) together with 18.435E (arctan
0.3333) making an angle of 45E. He is able to draw an angle
of 68.34 (arctan 2.52) with ruler and dividers utilising the
tangents of circles. Shelton records Grahams use of the 69
angle in the construct of his dead beat escapement.
In his writing and experimentation Harrison pre-
empted the work of many later physicists and mathematicians,
notably in the expansion of metals and the viscosity of air.
The RAS regulator has been associated with many important
personalities and developments since its inception and has
almost certainly prompted many ideas and understanding in
horology.
Attempting to determine the nature of Harrisons
final projected clock from existing artefacts and manuscripts
is analogous to writing the last movement of Schuberts
Unfinished Symphony after studying music and harmony.
The result may be in the same style, but there can be many
interpretations. The drawings prepared as part of this study
and their analysis may be questioned by many, but it is hoped
that they will provide data for further informed discussion of
the work of this remarkable scientific instrument maker and
O
D
E
C
A
P
B
To trisect angle AOB, Draw a
horizontal line EP. Divide the
vertical axis OE into three equal
parts and draw horizontal lines using
straight edge and dividers. Draw the
Hippias curve from C. Lines drawn
through the intersection of the curve
and the horizontal lines divide the
angle AOB into three parts.
229
experimenter of physical science.
. . . . they not wanting to have my true Ingenuity to do;
whenas at Land [thro my continued Diligence, and the nice
Performance of my Clock, and as without any foolish Obstruction or
Interruption,] I discovered [or found Means to discover] pretty much
for the better of late, viz. so as whereby to come up to such great
Exactness as I shall speak of below; nor does the Watch, or will the
Watch, as by Way of Trial, as just here implied, ever need any farther
than what belongs or must belong to its adjusting, viz. by the Help of
such a Clock as mine, or rather of such a Clock as my next will be; and
the Mistakes as here above intimated, must remain to other Workmen,
viz. until they be as by or from me the better instructed; not meaning
that they shall ever be so instructed, until I be [for a Public Good] the
more freely, or the more genteelly rewarded than what I have as hitherto
speakingly been, viz. for what I so highly, or deservingly done; no, the
remaining Part of my Discovery [and as still the more valuable] shall
(excepting better Usage) sleep, save only so far as to be my own
Content or Satisfaction, for not one Stroke as farther will I take; nor
shall I endeavour to seek after a Place at which to prepare an
Observatory, with suitable Conveniencies, for or to the Purpose; but as
being paid short, and that as farther withal, to be attended with a great
deal of Expence, Trouble, and Hindrance, [seurvy Work] I will also
be short, viz. as in a suitable Degree, excepting as intimated, I, or my
Discoveries, can as hereafter meet with better Usage that what I have as
heretofore met withal; for certainly it must be worth all, nay more than
all the Money it was to cost, as being so extraordinary a Matter, or
rather, as taking in the Clock, such extraordinary Matters as were never
to be expected to have appeared in the World; nor, as there is good
Reason to suppose, had it not been from me, would never have so
appeared, [as being so far out of the beaten Path] and still I must be
used ill.
Now, whether my Style of Writing in this affair, be right
proper to the Purpose or not, I thought it better than that the Contents
of this Book should be in Danger of sleeping in Oblivion; yea,
notwithstanding what I had, as verbally, communicated to the World.

FINIS.
John Harrison, Concerning Such Mechanism, 1775
230
Chapter 11
Drawings
The drawings have been made using CAD on a PC and
wherever possible constructed by eighteenth century methods as
described in John Robertsons Treatise of Mathematical Instruments
(1747). They are dimensioned in millimetres and Cartesian co-
ordinates are given with the origin coloured green. Actual measured
dimensions from the RAS Regulator, taken predominantly from
Goulds notebook, are coloured green. All other dimensions have
been derived photogrammetrically or from analysis of Harrisons
intent. Although all dimensions have been given to an accuracy of
0.01mm. it is suggested that they have a tolerance of 0.25mm.
When Harrison clearly used Imperial measurement, they have in many
cases been indicated in blue (i.e. drilled holes are dimensioned in
1/64th inch.). No indication of surface finish has been indicated in
the drawings as it is assumed that the best possible should be
obtained. Harrison did not polish out his construction lines in his
wooden or brass clocks. Many of his original lines are still visible and
are a helpful guide to his methods.
Material List
Plates 10 s.w.g. or 1/8" 2 x 470 x 410 (182" x 163") brass
Dial 16 s.w.g. 1.65 mm. 1 x 287 x 390 (112" x 182") brass
Dial backing 16 s.w.g. 1 x 220 x 225 (9" x 9") brass
Drum 3" x 22" diam. brass
Great wheel 10 s.w.g. 1/8" 1 x 225 x 225 (9" x 9") brass
Great wheel ratchet wheel 1/8" 1 x 86 x 86 (32" x 32") brass
Centre wheel s.w.g. 14 1 x 180 x 160 (7" x 7") brass
Maintaining pwr ratchet wheel s.w.g. 14 1 x 100 x 100 (4" x 4") brass
Third wheel s.w.g. 14 2 mm. 210 x 210 (8" x 8") brass
Remontoire wheel s.w.g. 14 or 2 mm. 220 x 220 (82" x 82") brass
Escapement wheel s.w.g. 16 1/16" 140 x 140 (52" x 52") brass
Roller bearing covers 12" x 3" diameter brass
Front roller bearing cover 3/8" 1 x 100 x 100 (4 x 4)
Rear plates s.w.g. 13 or 3/32" 200 x 130 (8" x 10") brass
Friction wheel supports s.w.g. 14 300 x 150 (12" x 6") brass
Friction wheels s.w.g. 20 x 32" x 32" brass
Girders 15" x 3/4" x 5/16" brass
Cross support 16" x 2" x 3/16" s.w.g.6 brass
Pillars and columns 35" x 7/8" diam. brass,
Dial pillars 12" x 2" diameter brass
Clock support 12" x 3/4" diameter brass
Pendulum support 2 x 13" x 3" x 3" brass
SAE660 leaded gunmetal 300 x 6.35 diameter (12" x 3")
SAE660 leaded gunmetal 300 x 9.53 diameter (12" x 3/8")
SAE660 leaded gunmetal 150 x 12.7 diameter (6" x 2")
231
Measurements from Gould's Notebook 1927
Height of plates 18.36 in. 466.34 mm.
Projection of fly beyond plate 2.75 in. 69.85 mm.
Half width of plates 5.62 in. 142.75 mm.
Half width of girder outside measurement 7.975 in 202.56 mm.
Fly in horizontal position projects beyond girders 0.495 in 12.57 mm.
Wheel projects from plates 2.62 in. 66.55 mm.
Distance between outside of dial and girders 2.355 in 59.82 mm.
Wheel projects beyond girders. 0.265 in 6.73 mm.
Height of horizontal plane to register hole centre 17.91 in. 454.91 mm.
Distant apart of register hole centres 8.23 in. 209.04 mm.
Distance between outside of plates 3.88 in. 98.55 mm.
Suspension spring plate length 13.05 in. 331.47 mm.
Suspension spring depth 2.50 in. 63.50 mm.
Projection of suspension spring plate from backplane 0.65 in. 16.51 mm.
Top of suspension spring holder suspension plate 0.35 in. 8.89 mm.
Overall height of clock 19.91 in. 505.71 mm.
Diameter of weight pulley 2.80 in. 71.12 mm.
Girder ends project beyond front plate 0.32 in. 8.13 mm.
Dial Plate projects 1.23 in. 31.24 mm.
Overhang beyond girders 0.91 in. 23.11 mm.
Length of girders 6.91 in. 175.51 mm.
Actual depth of clock 7.92 in. 201.17 mm.
Winding hole centre above horizontal plane 4.94 in. 125.48 mm.
Breadth of gut barrel approx 1.75 in. 44.45 mm.
Radius of barrel 1 1/16 in. 26.99 mm.
Radius of barrel mount pinion 11/16 in. 17.46 mm.
Great wheel arbor diameter 13/32 in. 10.32 mm.
Winding square length 59/64 in. 23.42 mm.
Winding square major side 19/64 in. 7.54 mm.
Winding square minor side 1/4 in. 6.35 mm.
Coques major diameter 9/16 in. 14.29 mm.
Radius of escapement wheel 2.66 in. 67.56 mm.
Measurements Provided by Jonathan Betts 2005
Distance between outside of main plates 98.55 mm.
Length of Pallet Frame 76.68 mm.
Maximum width of Pallet frame 14.44 mm.
Distance between pallet frame and pallet arbors 23.09 mm.
Radius of escapement wheel 71.44 mm.
232
Dimension Concordance
mm. Inch mm. Inch
13.10 0.516 33/64 0.40 0.016 1/64
13.49 0.531 17/32 0.79 0.031 1/32
13.89 0.547 35/64 1.19 0.047 3/64
14.29 0.563 9/16 1.59 0.063 1/16
14.68 0.578 37/64 1.98 0.078 5/64
15.08 0.594 19/32 2.38 0.094 3/32
15.48 0.609 39/64 2.78 0.109 7/64
15.88 0.625 5/8 3.18 0.125 1/8
16.27 0.641 41/64 3.57 0.141 9/64
16.67 0.656 21/32 3.97 0.156 5/32
17.07 0.672 43/64 4.37 0.172 11/64
17.46 0.688 11/16 4.76 0.188 3/16
17.86 0.703 45/64 5.16 0.203 13/64
18.26 0.719 23/32 5.56 0.219 7/32
18.65 0.734 47/64 5.95 0.234 15/64
19.05 0.750 3/4 6.35 0.250 1/4
19.45 0.766 49/64 6.75 0.266 17/64
19.84 0.781 25/32 7.14 0.281 9/32
20.24 0.797 51/64 7.54 0.297 19/64
20.64 0.813 13/16 7.94 0.313 5/16
21.03 0.828 53/64 8.33 0.328 21/64
21.43 0.844 27/32 8.73 0.344 11/32
21.83 0.859 55/64 9.13 0.359 23/64
22.23 0.875 7/8 9.53 0.375 3/8
22.62 0.891 57/64 9.92 0.391 25/64
23.02 0.906 29/32 10.32 0.406 13/32
23.42 0.922 59/64 10.72 0.422 27/64
23.81 0.938 15/16 11.11 0.438 7/16
24.21 0.953 61/64 11.51 0.453 29/64
24.61 0.969 31/32 11.91 0.469 15/32
25.00 0.984 63/64 12.30 0.484 31/64
25.40 1.000 1 12.70 0.500 1/2
mm. Inch s.w.g.
3.25 0.128 10
2.34 0.092 13
2.03 0.080 14
1.63 0.064 16
1.22 0.048 18
0.91 0.036 20
0.71 0.028 22
0.56 0.022 24
286.48 mm
1
1
1
0
9
8
7
6
405.13 mm
498.16 mm
266.70 mm
285.50 mm
tuart
0
3
0
2
0
1
0
3
0
4
0
5
0 6
243.63 mm
60
6 0
1
0
0
0
0
0
1
0
2
0
2
3
0
3
4
5
0
4
0
5
73.82 mm
23.58
466.34 mm
107.95 mm
0 10 30 50 70 90 10
24
Front Elevation
Scale 1:2.5
235
Wheels
117.47,432.57
117.47,201.92
117.47,29.30
R22.63 mm
R103.05 mm
R109.92 mm
R13.89 mm
R89.90 mm
R18.00 mm
R9.40 mm
R115.67 mm
0 10 30 50 70 90 10
R68.75 mm
A
B
C
D
E F
Scale 1:2.5
236
Wheel and Pinion Dimensions
Distance between Centres Position of Centres
sqrt(x
2
+y
2
) y x y x
125.47 0 Great Wheel
132.44 132.44 0.00 Great to Centre 257.91 0 Centre Wheel
112.33 94.62 60.54 Centre to Third 318.45 94.62 Third Wheel
116.76 -94.62 68.42 Third to Escape 386.87 0 Escape
119.06 -100.09 -64.48 Escape to R. Fly 322.39 -100.09 Remontoire Fly
2p/3 5p/7 2p/7 Chordal Radius Pitch No. Teeth Dist. or Radius
132.44 Great Wheel
6.60 2.64 114.81 9.25 78 114.78 78 tooth Wheel
6.16 17.86 12 17.66 12 tooth Pinion
112.33 Centre Wheel
6.72 2.69 89.90 9.41 60 89.86 60 tooth Wheel
6.27 22.63 15 22.47 15 tooth Pinion
116.76 Third wheel
5.14 2.06 103.05 7.19 90 103.03 90 tooth Wheel
4.80 13.89 12 13.74 12 tooth Pinion
119.06 Remontoire Wheel
5.14 2.06 109.92 7.19 96 109.90 96 tooth Wheel
4.80 9.40 8 9.16 8 tooth Pinion
3.57 120 68.12 Escapement Wheel
Distance between Centres Position of Centres
sqrt(x
2
+y
2
) y x y x
257.91 0 Centre Wheel
64.47 64.47 0.00 Centre to Hour 193.44 0 Hour Wheel
79.08 42.49 -66.69 Hour to Calendar 150.95 66.69 Calendar Wheel
4p/7 2p/3 p/3 Chordal Radius Pitch No. Teeth Dist. or Radius
64.47 Hour Wheel
4.82 2.41 55.30 7.23 48 55.26 48 tooth Wheel
4.13 9.45 8 9.21 8 tooth Pinion
79.08 Calendar Wheel
5.26 2.63 52.77 7.89 42 52.72 42 tooth Wheel
4.51 26.46 21 26.36 21 tooth Pinion
4.13 Fitted fixed roller size
Left Side
175.5 mm
98.6 mm
68.8 mm
8.1 mm
201.17 mm
31.24 mm
44.45 mm
125.48 mm,
54.88 mm
0 10 30 50 70 90 10
Scale 1:2.5
238
Right Side
175.5 mm
98.6 mm
68.8 mm
8.1 mm
201.17 mm
31.24 mm
44.45 mm
125.48 mm,
54.88 mm
Scale 1:2.5
0 10 30 50 70 90 10
239
Front Plate
Scale 1:2.5
Holes dimensioned in red are cut with front and rear plates fixed together. Friction wheel holes are cut
with their associated auxiliary plates. Details of the marking out of the side supporters and bottom
cutout are given in separate drawings.
240
466.34
0,0
0,466.34
133.35,466.34
33.95,353.10
0,386.87
120.65,381.00
127.86,254.00
123.03,130.42
0,257.91
0,125.47
9.52
6.35
6.35 6.35
6.35
405.13
6.35
0,161.06
0,89.82
22.55,102.92
35.79,125.72
22.55,148.02
-107.29,84.46
160.27,63.50
187.44,35.32
202.57,31.10
-189.34,25.94
-194.78,18.90
-199.12,13.26
50.80
3.18
3.18
4.76
0,395.22
0,325.32
94.62,318.45
94.62,333.05
96.59,238.33
4.76
86.43,281.30
128.64,311.00
-127.87,254.00
-100.09,269.10
-84.90,298.81
1.19
-100.09,322.39
6.35
-100.09,331.70
-122.38,308.05
1.19
-26.30,361.76 25.11,360.58
66.69,150.95
266.70 mm
117.47,432.57 -117.47,432.57
9.98 mm
-120.65,381.00
107.29,84.46
117.47,29.30
-117.47,29.30
9.98 mm
117.47,201.92
-117.47,201.92
9.98 mm
9.98 mm
-123.03,130.42
3.18 mm
4.76 mm
104.52,454.91 -104.52,454.91
-95.77,441.08
102.39,346.64
95.77,441.08
-102.39,346.64
50.80 mm
-27.83,230.03
23.55,281.40
-33.95,353.10
0,193.44
6.35 mm
133.35,88.12
4.76 mm
4.76 mm
3.17 mm
9.98 mm
1.19 mm
1.19 mm
4.76 mm
6.35 mm
4.76 mm
4.76 mm
6.35 mm
6.35 mm
11.11 mm
1.19 mm
11.11 mm
1.19 mm
1.19 mm
6.35 mm
4.76 mm
4.76 mm
1.19 mm
9.98 mm
15.87 mm
4.76 mm
15.88 mm
4.76 mm
209.04 mm
0 10 30 50 70 90 10
466.34
0,0
0,466.34
-133.35,466.34
0,386.87
0,257.91
0,125.47
405.13
-189.34,25.94
-194.78,18.90
-199.12,13.26
-94.62,318.45
100.09,322.39
266.70 mm
-117.47,432.57
117.47,432.57
117.47,29.30 -117.47,29.30
117.47,201.92 -117.47,201.92
9.53 mm
3.18 mm
104.52,454.91
-104.52,454.91
27.00/228.49
-34.12/217.76
-22.41/210.04
75.76/377.41
-58.53/377.41
86.94/415.12
109.87/405.35
3.18 mm
1.19 mm
6.35 mm
1.16 mm
-14.74,257.91
62.27,257.91
-17.40,346.18
1.19 mm
33.91,372.47
1.19 mm
4.76 mm
86.99,394.03
1.50 mm
1.50 mm
2.00 mm
1.19 mm
29.34,284.76
24.95,150.47
23.89,125.47
14.82,140.37
14.82,110.63
24.95,100.53
0,146.43
0,104.51
3.96 mm
3.96 mm
3.97 mm
-35.42,455.30
35.42,455.30
82.45,416.30
-82.45,416.30
4.76 mm
4.76 mm
9.53 mm
12.70 mm
12.70 mm
9.53 mm
6.35 mm
9.52 mm
4.76 mm
9.53 mm
9.52 mm 9.52 mm
3.50 mm
82.45,357.08
4.76 mm
-82.45,357.08
4.76 mm
6.35 mm
4.76 mm
-94.62,333.05
3.97 mm
-128.64,311.00
1.19 mm
98.68,341.19
3.97 mm
83.40,302.44
1.19 mm
98.65,272.48
4.76 mm
121.75,304.28
-86.43,281.30
1.19 mm
-94.62,238.98
3.97 mm
3.17 mm
0 10 30 50 70 90 10
31.75 mm
Rear Plate
Holes dimensioned in red are cut with front and rear plates fixed together. Friction wheel holes and pillar
holes for auxiliary plates are cut with their associated components.
Scale 1:2.5
This drawing is the posterior aspect of the rear plate
241
60.00
31.75 mm
31.75 mm
24.64 mm
R24.64 mm
R34.13 mm
7.94 mm
7.94 mm
0 10 30 50 70 90 10
Directions for marking out Supporters
1. Draw line from centre arbor point to E.
2. Extend line from edge of plate A to base at D and divide into eleven equal parts.
3. Draw a horizontal line FB 15.08 mm.
19
32
inch long 1
1
4
inches above base line. This is four parts up
the line DB and its vertex is on the line DH which subtends an angle of 60 to the line AD.
4. Draw horizontal line JC 2
1
2
inches 63.5 mm. above base line. This is eight parts up the line DC.
5. Draw quadrant radius AC centred on C for upper arc of supporter.
6. Draw a
9
64
inch, 3.57 mm. vertical line from G, and a
1
8
inch, 3.175 mm.
horizontal line from I. Draw the line GI and draw the perpendicular LM from its midpoint.
7. Define the intercept M by drawing an arc from K centred on C. An arc from K to J centred on M
produces the lower arc of the supporter.
D
A
G
E
B
K
J
C
L
M
F
H
I
Scale 1:2
242
A
B
C
D E
F
G
H
I
J
K
Directions for marking out Cutout
1. Draw an arc BC from centre arbor point of radius centre point to top of front plate (208.43mm.).
2. Scribe a horizontal line, DE, tangent to this arc and parallel to the base. This marks the bottom of the
cutout.
3. Draw horizontal line KK 1/5 width of base (405.13 mm.)above baseline to mark top of cutout.
4. Mark points K-K from mid-line an equal distance to distance between top and bottom lines.
5. Draw arcs of radii FA centred on F cutting the arc BC. These intersections I-J mark the lateral
extremities of the cutout
6. Draw perpendicular from midpoint of line joining K-J. Mark intersection M where perpendicular
intersects with line GH
7. Draw arc between K and J centred on M to mark the upper side of the cutout.
8. Scribe lower arc of cutout which is the quadrant of a circle from the lateral extremity J to the bottom
line.
n.b. In the photographs of the base of the front plate it can be seen that Harrison has inscribed a line 45
from great wheel arbor centre to the baseline. The significance of this line has not been determined.
K
L
M
Scale 1:2
7.94 mm
R208.43 mm
128.96 mm
49.48 mm
81.03 mm
87.46 mm
99.39 mm
63.09 mm
7.94 mm
31.54 mm
0 10 30 50 70 90 10
243
Scale 1:2.5
Front Anterior Auxiliary Plates and Friction Wheels
244
0,395.22
0,325.32
-100.09,331.70
-100.09,269.10
-38.21,317.64
-38.21,186.00
94.62,333.05
94.62,238.98
0,193.44
66.69,150.95
II
III
IV
I
0 10 30 50 70 90 10
-107.29,84.46
107.29,84.46
Scale 1:2.5
Front Posterior Auxiliary Plates and Friction Wheels
245
This drawing is the posterior aspect of the front plate
0/124.49
76.40,392.25
-97.39,221.70
-117.47,201.92
117.47,432.57
33.80,352.92 -33.73,352.92
-25.11,360.58 26.30,361.76
0,257.91
100.09,269.10
-94.62,318.45
0,193.44
-66.69,150.95
38.22,186.00
38.22,317.65
0 10 30 50 70 90 10
100.09,331.70
-107.29,84.46
107.29,84.46
Scale 1:2.5
Rear Anterior Auxiliary Plates and Friction Wheels
246
XII
58.53,377.41 -75.76,377.41
-98.68,341.19
-98.65,272.48
94.62,333.05
94.62,238.98
34.12,217.76
22.41,210.04
-86.94,415.12
V
VI
VII
0 10 30 50 70 90 10
35.42,455.30 -35.42,455.30
82.45,416.30
82.45,357.08
-82.45,357.08
-82.45,416.30
62.27,257.91
-14.74,257.91
-34.12,217.76
-22.41,210.04
VIII
IX
X
XI
0 10 30 50 70 90 10
Scale 1:2.5
Rear Posterior Auxiliary Plates and Friction Wheels
247
This drawing is the posterior aspect of the rear plate
The arbor positions are fixed by the layout as described previously.
For centre wheel and third wheel draw line between wheel centres and draw perpendiculars. Friction wheels are
on lines 45 apart from these perpendiculars. Draw line from escapement wheel centre to remontoire fly centre.
Draw perpendicular to this line at remontoire fly centre. Friction wheels for remontoire fly are on this extended
line and its perpendicular. Draw line from third wheel centre to escapement wheel centre. Bisect the angle
between this line and the line between the escapement and remontoire fly centres. The friction wheels for the
escapement wheel arbor are on lines 45 to this bisector.
The alternate facing arrangement of brass and wood of the collets on the friction wheels is indicated by the
coloured circles.
Front Plate Geometry
Scale 1:2.5
248
A
B
C D
E
F
G
H
90
1.26
90
0/257.91
-100.09,322.39
0/386.87
128.64,311.00
86.43,281.30
-27.83/230.03
-23.55/281.40
25.11/360.57
-26.30/361.76
-84.90,298.81
-122.38,308.05
94.62,318.45
3.97
2.38
2.78
1.98
90
0,0
50.63
62.55
74.82
73.31
70.75
70.75
90
107.29,84.46
123.03,130.42
127.28,256.39
120.65,381.00
117.47,432.57
104.52,454.91
117.47,201.92
117.47,29.30
260.35 mm
77.24 mm
0,193.44
0,124.48
15.88 mm
-104.52,454.91
15.88 mm
4.76 mm 4.76 mm
4.76 mm
-95.77,441.08
-102.39,346.64
95.77,441.08
4.76 mm
102.39,346.64
66.69,150.95
4.76 mm
4.76 mm
R119.06 mm
0 10 30 50 70 90 10
53.75 mm
66.88 mm
Scale 1:2.5
Front Posterior Geometry
249
This drawing is the posterior aspect of the front plate
0/124.49
90
0/192.79
33.95/353.10 -33.95,353.10
I J
86.06
86.06
-128.64,311.00
-86.43,281.30
27.83/230.03
23.55/281.40
-94.62,318.45 100.09,322.39
84.90,298.81
122.38,308.05
107.29,84.46
26.30,361.76
-25.11,360.57
0 10 20 30 40 50 60 70 80 90 10
0,0
Scale 1:2.5
Rear Anterior Geometry
250
-34.12/217.76
-22.41/210.04
9.70
90 -33.91/372.47
17.40/346.18
58.53/377.39
0/386.87
-86.94/415.12
-109.87/405.35
-86.99/394.01
K L
M
N
O
P
78.95
64.03
128.64,311.00
86.43,281.30
94.62,318.45
2.78
90
73.31 mm
66.88 mm
-100.09,322.39
-83.40,302.01
-121.75,304.28
2.38
54.08
49.66 mm
90
12.70 mm
98.68,341.19
-75.76,377.41
0 10 20 30 40 50 60 70 80 90 10
35.42,455.30 35.42,455.30
0,0
Scale 1:2.5
Rear Posterior Geometry
251
This drawing is the posterior aspect of the rear plate
0/124.49
69.09/416.30
82.45/416.30
69.09/366.16
82.45/357.08
109.85/406.99
-14.74/257.80 62.27/257.79
27.00/228.49
-34.12/217.76
-22.41/210.04
0/402.17
90.00
90
Q
R
S
T
68.38
68.38
75.73
75.73
0 10 30 50 70 90 10
29.33,284.76
25.24,362.37
-24.51,361.62
0,0
92.20
9.98 mm
7.94 mm
9.53 mm
3.37 mm
R5.77 mm
23.81 mm
43.66 mm
20.50 mm
15.88 mm
12.28 mm
11.68 mm
12.33 mm
4.76 mm 5.56 mm
4.68 mm
R2.96 mm
1.65 mm
5.95 mm
6.61 mm 12.53 mm
5.68 mm
20.71 mm
29.21 mm
6.35 mm
12.28 mm
Main Pillars
252
Pillars B,C,and F
Pillar E
Pillars A and D
Front on left to show
pinning
Pillars riveted to
rear plate
Dial Pillar
Dial Backing Pillar
Riveted to dial
Riveted to front plate
Scale 1:1.5
0 10 30 50 70 90 10
B,C, and F
E
A and D
Front on left to show
pinning to front plate
Front Auxiliary Plates
82.60 mm
12.94 mm
12.7 mm
12.7 mm
2.65 mm for
1
8" Whit.
or 3.00 M
75.49 mm
12.75 mm
9.52 mm
12.70 mm
6.70 mm
12.70 mm
9.52 mm
12.82 mm
144.32 mm
12.70 mm
9.53 mm
12.82 mm
Scale 1:2
R2.52 mm
3.97 mm 9.93 mm
4.52 mm
17.32 mm
II, III. & IV
I
0 10 20
0 10 20
Scale 1:1
Scale 1:1
R2.52 mm
3.97 mm 9.93 mm
16.49 mm
4.52 mm
All front anterior plates 2.03 mm thick.
0 10 30 50 10
253
Crosspiece
22.21 mm
16.27 mm
262.41 mm
Scale 1:2
15.87 mm
12.33 mm
4.76 mm
3.37 mm
5.68 mm
R2.66 mm
0 10 20
Scale 1:1
76.40,392.25
-97.39,221.70
-117.47,201.92
117.47,432.57
33.80,352.92
-33.73,352.92
25.42 mm
16.63 mm
354.74 mm
0 10 30 50 70 90 10
254
This drawing is the posterior aspect of the front plate
Rear Auxiliary Plates
Scale 1:2
15.97 mm
153.33 mm
75.76 mm
134.29 mm
02 mm
12.82 mm
9.41 mm
106.77 mm
94.07 mm
12.50 mm
81.39 mm
12.66 mm
68.71 mm
17.64 mm
19.00 mm
95.98 mm
77.02 mm
17.55 mm
12.71 mm
12.71 mm
18.89 mm
91.41 mm
72.36 mm
13.55 mm
12.33 mm
4.76 mm
2.5 mm
5.68 mm
R2.66 mm
16.49 mm
0 10 20
V
Scale 1:1
R2.52 mm
3.97 mm 9.93 mm
16.49 mm
4.52 mm
0 10 20
Scale 1:1
VI & VII
0 10 20
0 10 20
R2.52 mm
3.97 mm 9.93 mm
15.15 mm
4.52 mm 2.54 mm
12.33 mm
4.76 mm
3.37 mm
5.68 mm
R2.66 mm
20.71 mm
4.76 mm
Scale 1:1
Scale 1:1
VIII
IX
2.34 mm thick
2.03 mm thick
2.03 mm thick
2.34 mm thick
3.175 mm thick
0 10 30 50 70 90 10
255
The drawings VIII and IX are the posterior aspect
Escapement Plates
0 10 30 50 70 90 10
Scale 1:2
116.39 mm
159.96 mm
50.14 mm
46.63 mm
20.64 mm
106.17 mm
138.18 mm
183.05 mm
59.39 mm
77.36 mm
138.18 mm
164.91 mm
50.14 mm
59.22 mm
70.84 mm
83.61 mm
12.70 mm
R2.52 mm
3.97 mm 9.93 mm
15.15 mm
4.52 mm
Scale 1:1
4.76 mm
5.68 mm
R2.66 mm
14.82 mm
R2.52 mm
11.96 mm
3.97 mm
4.52 mm
10.90 mm
9.47 mm
0 10 20
256
Friction Wheels
12 B.A. or M1.5
0.91 mm
1.82 mm
0.91 mm
16.53 mm
7.14 mm
15.88 mm
7.56 mm
4.90 mm
15.88 mm
9.29 mm
3.20 mm
I and J
The flanges on the wooden collets of all the friction wheels except I and J are on the
inside so that the associated wheel is above them. The flanges on I and J are on the
outside. The plates are all 16.67 mm. apart.
0 10 20 30 40 50 60 70 80 90 10
Outside Diameters
66.88 mm. P 54.08 mm. K 74.82 mm. F 50.63 mm. A
75.73 mm. Q 49.66 mm. L 73.31 mm. G 53.75 mm. B
75.73 mm. R 64.03 mm. M 66.88 mm. H 70.75 mm. C
68.38 mm. S 78.95 mm. N 86.06 mm. I 70.75 mm. D
68.38 mm. T 73.31 mm. O 86.06 mm. J 62.55 mm. E
257
Rear Roller Bearing
Front Roller Bearing

4
3
.
1
4

1
4
.
3
5

5
0
.
8
0

1
4
.
3
5
13.34 mm
6.37 mm
31.75 1
1
4
"

8
.4
7

8
.4
7

2
5
.
4
0
13.17 mm
3.91 mm
0 10 30 50 10
Scale 1:1
N.b. cheese head machine screws used where hidden
by the dust covers and panhead where visible.
258
108.23 mm
2.78 mm
6.35 mm
2.78 mm
5.57 mm
6.79 mm 6.12 mm
3.97 mm
7.90 mm
Winding Square
10.32 mm
2.38 mm
2.38 mm
Fly
3.97 mm
3.65 mm
4.49 mm
6.54 mm
9.70 mm
3.65 mm
approx. 148 mm
1.98 mm
2.54 mm
69.58 mm
1.98 mm 0.98 mm
3.97 mm
2.38 mm
105.04 mm
Arbors
Remontoire Fly
Remontoire and Escapement
Third Wheel
Centre Wheel
Great Wheel
0 10 20 30 40 50 60 70 80 90 10
Scale 1:1
259
2P/3
2P/7
P
5P/7
5P/7
2P/7
Wheel Details and Tooth Form
All wheels have similar tooth form and proportions but different pitch.
P 360
Chordal radius of Lantern Pinions where
2sin 2N
q
q
= =
Figure 1, 1730 Manuscript
Remontoire Wheel
10 mm.
Going Train Proportions
Motion Work Proportions
Harrisons Drawing
Hour Wheel
P
P
/3
4 P/7
4 . 1 3 m m
260
G
e
n
e
r
a
l

D
r
a
w
i
n
g

o
f

G
r
e
a
t

W
h
e
e
l

a
n
d

C
r
o
s
s

G
i
r
d
e
r
R
6
0
.
2
5

m
m
R
5
5
.
2
3

m
m
R
1
2
.
8
6

m
m
R
8
.
7
3

m
m
R
1
0
5
.
5
8

m
m
R
5
0
.
8
0

m
m
R
1
1
5
.
6
7

m
m
1
3
3
.
4
3

m
m
0
1
0
3
0
5
0
7
0
9
0
1
0
W
h
e
e
l

t
h
i
c
k
n
e
s
s

1
/
8
"

3
.
1
7
5

m
m
.
S
c
a
l
e

1
:
2
4
4
.
4
4

m
m
4
.
7
6

m
m

t
h
i
c
k
3
9
9
.
8
0

m
m
6
3
.
5
1

m
m
1
0
7
.
9
5

m
m
0
,
0
4
8
.
1
9
,
2
2
.
2
2
8
2
.
5
5

m
m
5
.
5
6

m
m
S
t
o
p
w
o
r
k
0
1
0
2
0
3
0
4
0
5
0
2
6
1
N
.
b
.

G
r
e
a
t

w
h
e
e
l

l
o
o
s
e

o
n

a
r
b
o
r

w
i
t
h

a
n


e
x
t
e
n
d
e
d

c
o
l
l
e
t

w
i
t
h
i
n

d
r
u
m
.


D
i
a
m
e
t
e
r

r
e
c
o
r
d
e
d

b
y

G
o
u
l
d

a
s

1
1
/
1
6
"
,

1
7
.
4
6

m
m
.

b
u
t

l
e
n
g
t
h

u
n
k
n
o
w
n
.


G
r
o
o
v
e
s

o
n

d
r
u
m

a
r
e

a

l
e
f
t

h
a
n
d

h
e
l
i
x

w
i
t
h

2
.
5
4

m
m
.

1
/
1
0
"

p
i
t
c
h
.
D
r
u
m

D
e
t
a
i
l
S
c
a
l
e

1
:
1
4
4
.
4
7

m
m
1
7
.
4
6
m
m
5
3
.
9
7
5
m
m
5
7
.
7
1

m
m
8
.
6
8

m
m

3
1
.
7
5

m
m
3
0
.
5
6

m
m
6
5
.
8
8
m
m
0
1
0
2
0
3
0
4
0
5
0
6
0
7
0
8
0
9
0
1
0
2
6
2
1
7
.
5

T
u
r
n
s
4
4
.
4
4

m
m
6
5
.
3
3

m
m
19.05 mm
4.76 mm
13.78 mm
114.3 mm
3.47 mm
13.78 mm
16.77 mm
0 10 20 30 40 10
Frame Pillars
Scale 1:1
263
Maintaining Power
R42.92 mm
46.4 sq. mm.
69.6 sq. mm.
Three and a half turns spiral spring
tensioned so that its torque is less
than that exerted by the driving weight
0 10 20 30 40 50 60 70 80 90 10
2.03 mm
9.52 mm
Rear Plate
3.17 mm
5.16 mm
6.35 mm
2.03 mm
R38.84 mm
R22.42 mm
4.37 mm
3.00 mm
R22.77 mm
Ratchet wheel and lantern pinion loose on
centre arbor. Drive plate pinned to arbor.
Ensure that both contact faces of the drive
plate are in contact with the extended pins of
the lantern pinion.
264
46.25 mm
3.18 mm
5.75 mm
40.22 mm
33.74 sq. mm. 51.27 sq. mm. 47.89 sq. mm. 67.02 sq. mm.
Speculum Metal
Lignum vitae or PTFE
Great Wheel and Maintaining Power Ratchets
0 10 20 30 40 10
Great Wheel Maintaining Power
Scale 1:1
265
2.07 mm
4.12 mm
10.34 mm
22.21 mm
R103.05 mm
R13.89 mm
33.34 mm
16.67 mm
R8.14 mm
116.76 mm
112.33 mm
5.38 mm
2.65 mm
R89.90 mm
R22.63 mm 52.62 mm
35.21 mm
6.35 mm
22.21 mm
R8.14 mm
10.34 mm
Centre and Third Wheels
Centre Wheel and Third Wheel Lantern Pinion
Third Wheel and Remontoire Wheel Lantern Pinion
0 10 30 50 70 90 10
Scale 1:2
266
Wheel thickness s.w.g.14 2.032 mm.
Wheel thickness s.w.g. 14 2.032 mm.
Escapement Wheel
Scale 1:1
Scale 3:1
Tangent of
11.35 mm circle
Tangent of
34.06 mm circle
0 10
267
12.7 mm
1
2
"
2
0

3
1
.
7
5

m
m

1
5
.
9

m
m

5
8
"

1
1
4
"
0 10 30 50 10
R68.12 mm
R34.06 mm
R11.35 mm
Wheel thickness s.w.g. 16 1.63 mm.

2
2
.
4
5

m
m

1
9
.0
5
m
m

15.88 m
m
12.70 mm
4.76 mm
1.59 mm
12 B.A.
Escapement Wheel Collet
0 10 30 50 10
Scale 1:1
268
P
a
l
l
e
t

F
r
a
m
e

a
n
d

C
r
u
t
c
h
2
.
5
2

m
m
1
5
.
1
3

m
m
7
6
.
6
8

m
m
4
1
.
4
7

m
m
3
5
.
2
1

m
m
2
3
.
0
9

m
m
1
4
2
.
4
2

m
m
9
.
2
2

m
m
5
0
.
8
0

m
m
6
9
.
2
4

m
m
9
.
2
2

m
m
4
.
2
0

m
m
6
.
3
5

m
m
6
.
0
2

m
m
4
.
2
0

m
m
1
4
.
4
4

m
m
2
2
.
7
1

m
m
2
.
0
0

m
m
1
9
.
1
0

m
m
3
2

m
m
0
1
0
2
0
3
0
4
0
1
0
S
c
a
l
e

1
:
1
2
6
9
Pallets and Composers
5
.
4
1

m
m
8
.
2
7

m
m
39.60 mm 54.24 mm
36.02 mm 31.67 mm
R33.94 mm
1
.
4
2

m
m
R51.72 mm
4.90 mm
32.17 mm
4.30 mm 10.82 mm
33.89 mm
0 10 20 30 40 10
Pallets
Pallets made from Snakewood with lead weights.
Note angle of pallet contact faces.
Scale 1:1
Composers
270
Pallet Frame Knife Edge
55.39 mm
13.50 mm
7.31 mm
2.03 mm
15.15 mm
55.39 mm
70.89 mm
Scale 1:1
0 10 30 50 10
271
1.Cut out frame pattern from 16 swg brass.
2.Ensure balance point is in the correct position using knife edge.
3. Cut 15 mm length of 1/4" brass rod and bore 3.5 mm through centre.
4.Mill 2.5 mm slots at each end of rod using a jig made of square stock so that slots are parallel.
5.Mill 3 mm slot in one side of pallet frame centred on balance point.
6.Rod can now be eased into pallet frame and silver soldered.
7.Central hole drilled and reamed 4.2 mm to fit arbor
8. Arbor fitted and pinned.
Pallet Frame Construction
272
R
e
m
o
n
t
o
i
r
e

L
a
y
o
u
t
S
c
a
l
e

1
:
2
.
5
0
1
0
3
0
5
0
7
0
9
0
1
0
2
7
3
0/386.95
-86.94/415.12
-109.87/405.35
-86.99/394.01
-91.90/332.39
142.00 mm
0 10 30 50 10
Remontoire Release and Fly
Gould replaced the crutch with a heavier adjustable brass one. This resulted in the clock requiring a
much heavier driving weight and more importantly a larger remontoire fly. The drawings are for a fly
as originally fitted and as seen in Cottingham's photograph. The fly is constructed of thin copper
which is corrugated to impart rigidity. It is carefully balanced with a small lead counterweight
sufficient only to ensure that the restraining arm comes to rest against its detent. The smaller fly
enables the remontoire springs to be tensioned in minimum time thereby reducing the time that the
escapement is powered directly by the driving weight.
274
Scale 1:1
0
.
9
1

m
m
S
c
a
l
e

1
:
1
5
0
.
4
6

m
m
0
.
9
1

m
m
1
2
.
7
7

m
m
4
9
.
2
0

m
m
3
.
2
4

m
m
2
5
.
4
8

m
m
2
.
8
9

m
m
4
2
.
3
5

m
m
6
4
.
8
3

m
m
8
1
.
7
8

m
m

1
5
.
6
3

m
m
3
2
.
1
1

m
m
8
7
.
8
2

m
m
0
1
0
3
0
5
0
1
0
R
e
m
o
n
t
o
i
r
e

R
e
l
e
a
s
e

C
o
m
p
o
n
e
n
t
s
C
a
m

R
o
l
l
e
r

C
o
u
n
t
e
r
p
o
i
s
e
C
a
m

R
o
l
l
e
r
R
e
l
e
a
s
e

F
o
r
k

C
o
m
p
o
s
e
r
R
e
l
e
a
s
e

F
o
r
k
A
d
j
u
s
t
a
b
l
e

s
t
o
p

f
o
r
C
a
m

R
o
l
l
e
r

C
o
u
n
t
e
r
p
o
i
s
e
A
r
b
o
r
A
d
j
u
s
t
m
e
n
t
s
a
r
e
m
a
d
e
w
h
e
n
s
e
t
t
i
n
g
u
p
w
i
t
h
t
h
e
s
c
r
e
w
r
e
s
t
o
n
t
h
e
r
e
l
e
a
s
e
f
o
r
k
a
n
d
t
h
e
p
o
s
i
t
i
o
n
o
f
t
h
e
r
e
l
e
a
s
e
f
o
r
k
c
o
m
p
o
s
e
r
o
n
i
t
s
a
r
b
o
r
c
o
l
l
e
t
.
T
h
e
r
e
l
e
a
s
e
f
o
r
k
c
o
m
p
o
s
e
r
h
a
s
e
l
l
i
p
t
i
c
a
l
h
o
l
e
s
;
i
t
i
s
f
i
x
e
d
w
i
t
h
c
h
e
e
s
e
h
e
a
d
m
a
c
h
i
n
e
s
c
r
e
w
s
.
A
l
l
o
t
h
e
r
w
h
e
e
l
s
i
n
t
h
e
c
l
o
c
k
a
r
e
f
i
x
e
d
w
i
t
h
d
o
m
e
d
c
o
u
n
t
e
r
s
u
n
k
m
a
c
h
i
n
e
s
c
r
e
w
s
.
2
7
5
Remontoire Fly and Arbor
3
.
6
6

m
m
21.36 mm
142.00 mm
4
.
8
3

m
m
6
.
7
5

m
m
2.54 mm
2
5
.
3
8

m
m
105.04 mm
17.23 mm
18.24 mm
30.76 mm
47.71 mm
2.71 mm
3.66 mm
5.88 mm
7.10 mm
22.94 mm
0 10 30 50 10
Scale 1:1
276
Remontoire Control Bracket
26.86 mm
36.59 mm
7.78 mm
19.01 mm
0 10 30 50 10
Scale 1:1
277
0 10 20 30 40 50 60 70 80 90 10
5.88 mm
Remontoire Wheel
Scale 1:1
Scale 1:1.25
Wheel thickness 2.03 mm.
Cam thickness 3.175 mm.
Springs made of 15 closely wound 24 swg
piano wire 10.71 mm. diameter. 100 grms
should extend spring to 100 mm.
0 10 30 50 10
278
A
B
C
D
E
F
G
H
A
B
C&D
E
F
G
H
Pb
F
G
H
1
5
9

m
m
1
8
4

m
m
7
9
4

m
m
Portrait Pendulum
76.20 mm
14.5 mm
9 mm
R124 mm
22.25 mm.
Constructed of 2 mm. steel and
3.175 mm. brass rods with nickel
silver cross pieces. A, B, E, F, G,
and H 3.175 mm thick. C and D
1.588 thick.
Scale 1:4.5
5.56 mm
44.45 mm
38.10 mm
20.64 mm
0 10 30 50 70 90 10
0 10 20 30 40 50 60 10
279
A
B
Scale 1:1
Pendulum Suspension Block
A
273.05 mm
9.53 mm
Radial Cheeks
25.40 mm
0 10 30 50 70 90 10
Scale 1:2.5
Scale 1:1.25
280
0 10 30 50 10
9 ct Gold spring suspension
70 x 6 x 0.5 mm.
Calibrated capstan adjusting screw
56 t.p.i. x 3/32"
Rolled brass or bronze cheeks
P
e
n
d
u
l
u
m

S
u
s
p
e
n
s
i
o
n

R
a
d
i
a
l

C
h
e
e
k
s

L
a
y
o
u
t
2
8
1
R
7
3
.
0
3

m
m
R
7
3
.
0
3

m
m
6
3
.
5
0

m
m
1
3
6
.
6
2

m
m
0
1
0
2
0
3
0
4
0
5
0
6
0
7
0
8
0
9
0
1
0
0
,
0
9
.
5
2
1
8
.
2
6
4
4
.
4
5
6
3
.
5
0 6
7
.
6
8
8
8
.
6
9
1
0
9
.
0
3
1
1
7
.
9
3
1
1
9
.
1
1
3
2
.
0
0
,
2
1
.
6
3

2
5
.
4
0

m
m
5
.
5
6
1
0
.
6
1
1
4
.
2
9
1
9
.
4
5
2
9
.
4
2
3
4
.
9
2
3
8
.
1
0
4
1
.
2
7
5
1
.
9
7
5
3
.
9
7
5
9
.
1
3
R
7
1
.
8
1

m
m
R
7
2
.
4
6

m
m
3
.
5
8

s
i
n
3
.
5
8

=
1
1
6
Pendulum Suspension Block
Tap small holes 4 BA, M3.5 or 5/32" Whit.
331.47 mm
63.50 mm
9.53 mm
1.00 mm. or 24 tpi thread
Mounting Plate
209.04 mm
41.91 mm
11.91 mm
11.95 mm
11.91 mm
159.24 mm
23.02 mm
11.95 mm
159.24 mm
16.95 mm
9.46 mm
11.91 mm
Coques
16.49 mm 24.00 mm
1.00 mm. or 24 tpi thread
Scale 1:2.5
282
0 10 30 50 70 90 10
286.48 mm
286.47 mm
60
6 0
0
3
0
3
tuart
R121.81 mm
R120.87 mm
R119.26 mm
R106.28 mm
R105.25 mm
R103.94 mm
7.14 mm
0 10 30 50 70 90 10
Front Dial Plate
283
O/381.13
0/272.13
0/143.17
0/9.75
154.77/339.77
23.61
R
0
.
4
R
286.47 mm
R7.94 mm
R
5
0
.8
0
m
m
R
3
0
.7
1
m
m
4
8
.
6
2

-44.17/103.79
-26.70/93.86
0/129.47
0/114.46
0/78.69
0/230.95
0/217.65
61.63/217.06
80.46/200.40
33.77/196.62
5/8"
120
0/164.29
R
41.18 m
m
R
8
2
.5
5
m
m
R
1
0
7
.
9
5

m
m
0 10 30 50 70 90 10
R54.51 mm
R13.30 mm
-120.65,265.28
-127.87,138.28
-123.03,14.70
123.03,14.70
127.87,138.28
120.65,265.28
0,0
286.47 mm
R
8
.9
7
m
m

3
.9
7
m
m
D
C
B
E
I I I
II
II
III
IV
II
I I
III
Layout of Front Dial
284
Layout of Front Dial
Outline
Outline drawn as described previously for RAS Regulator.
Seconds and Hour Aperture
I. Draw the arcs R107.95 mm.(4), R82.55mm. (3) and
R41.18mm. (1e).
II. Draw lines 48.59E from centre-line from D to outer arc. The angle is
constructed by drawing a perpendicular 0.75R from the centre line on
the outer arc (i.e. sin 48.59E = 0.75). Bisect outer arc from this line
to centre and draw line to D.
III. Bisect centre-line between D and outer arc. Draw circle from this point
with radius to touch the inner arc.
IV. Draw line from apex of circle III to inner line crossing middle arc I.
Bisect this line and bisect its lower half. Draw a perpendicular through
the lower half bisection point and draw a circle diameter 3.97mm.
(
5
/
32
).
V. Draw an arc centered on point where line III crosses middle arc of radius
half line IV. Draw an arc centered on midpoint of line IV on opposite
side of radius where perpendicular IV cuts middle arc. At the
intersection of these arcs draw a circle of radius equal to circle III.
VI. Bisect the outer segment of the middle arc and draw an arc to pass
through this point centered on the centre-line of radius from this point
to the centre of circle V.
VII. The relevant parts of these constructions are joined to mark out the
aperture.
Hour Aperture
I. Mark point E 63.50 mm. (2) below the centre point C and draw arcs
R50.80 mm. (2") and 29.36 mm . (1
5
/
32
)
II. Draw arc of equal radius from centre line on larger arc. Draw lines from
the intersection to E.
III. Draw two circles 6.35 mm. () diameter tangentially to upper arc
crossing midline for index.
IV. The relevant parts of these constructions are joined to mark out the
aperture.
285
Front Seconds Dial Plate.
Scale 1 : 2
286
R25.4 mm 6.5"
165.10 mm
0/207.59
10.25"
260.35 mm
10.00
0/257.91
1"
3.63
0/193.01
R102.62 mm
4
"
99.50
222.45 mm
0
3
0
2
0
1
0
3
0
4
0
5
0 6
3
5
8
"
109.76
R92.32 mm
0/386.87
0/0
12.70 mm
1
2
"
102.39,-40.23
95.77,54.21
110.01,63.40
110.01,-48.18
0,110.01
0 10 30 50 70 90 10
R110.01 mm
110.01 mm
111.58 mm
20.04 mm
8.39 mm
220.01 mm
7
8
9
1
1
1
0
1
2
1
3
1
4
1
5
1
6 1 7
1
8
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
1
2
3
4
5
2
9
3
0
3
1
6
2
8
Dial Work
1"
R110.01 mm
0 10 30 50 70 90 10
2
3
1
2
3
4
5
6
7
8
9
1
1
1
0
1
2
1
3
1
4
1
5
1 6 1 7
1
8
1
9
2
0
2
1
2
2
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3
1
287
266.70 mm
405.13 mm
209.04 mm
0,193.44
R55.27 mm
R9.21 mm
R52.71 mm
R19.36 mm
4.13 mm
4.13 mm
66.69,150.95
0,257.91
107.29,84.46
-107.29,84.46
0 10 30 50 70 90 10
0,0
Dial Motion Layout
Scale 1:2.5
288
0
/
2
5
7
.
9
1
1
0
.
0
0

1
0
.
2
5
"
0
/
2
0
9
.
8
3
6
.
5
"
R
2
5
.
4

m
m
1
"
0
1
0
3
0
5
0
7
0
9
0
1
0

1
6
5
.
1
0

m
m

1
9
4
.
6
2

m
m

2
6
0
.
3
5

m
m
1
2
3
4
5
6
7
8
9
1
1
1
0
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
2
0
2
1
2
2
2
3
24
2
5
2
6
2
7
2
8
2
9
3
0
3
1
D
a
y

W
h
e
e
l

a
n
d

D
i
a
l

P
i
l
l
a
r
s
1
2
.
3
3

m
m
4
.
7
6
m
m
5
.
5
6

m
m
4
.
6
8

m
m
R
2
.
9
6

m
m
1
.
6
5
m
m
9
5

m
m
6
.
6
1

m
m
2
.
5
3

m
m
D
i
a
l

B
a
c
k
i
n
g
D
i
a
l
5
.
6
8

m
m
2
0
.
7
1

m
m
2
9
.
2
1

m
m
6
.
3
5
m
m
0
1
0
2
0
3
0
4
0
1
0
H
o
u
r

D
i
a
l

a
n
d

D
a
y

W
h
e
e
l

D
r
i
v
e
6
4
.
4
8

m
m
R
5
5
.
3
1

m
m
R
9
.
5
5

m
m
7
8
.
6
6

m
m
R
2
4
.
6
8

m
m
R
1
8
.
4
6

m
m
R
5
3
.
7
2

m
m
R
3
0
.
0
5

m
m
R
2
6
.
4
9

m
m
R
1
2
.
0
0

m
m
0
1
0
2
0
3
0
4
0
5
0
6
0
7
0
8
0
9
0
1
0
3
.
1
7

m
m
4
.
5
1

m
m
2
9
0
Gould's Notebook p. 14
Cannon Pinion
18.22 mm
26.10 mm
4.95 mm
0 10 20 30 40 10
11.20 mm
The spring is made of hard brass by rolling 16 s.w.g to approximately 18 s.w.g. thickness
and bending into the required shape. The brass rollers are soldered onto their backing plate
using lead free tin solder cream. The solder is prevented from spreading beyond the
junction by painting the surrounding areas with Plaster of Paris. Grooves are machined in
the bronze arbor to enable the front friction wheels to accommodate side thrust.
291
207.50 mm
12.70 mm
3.97 mm
Scale 2:1
9.53 mm
13.85 mm
Thury 13 or
13 B.A.
Scale 2:1
12.70 mm
242.95 mm
1.98 mm
0.80 mm
Minute and Seconds Pointers
0 10 20 30 10
292
50.80 mm
33.34 mm
8.58 mm
6.35 mm
84.32 mm
90.00
86.86 mm
19.05 mm
12.06 mm
29.46 mm
105.69 mm
81.53 mm
33.34 mm
0 10 20 30 40 50 60 70 80 90 10
76.20 mm
Drive Weight and Pulley
Scale 1:1.5
Brass Weight
152.01 cu.cms.
1.27 kgms.
293
3
6
8
.
3

m
m
.
3
2
7
.
0
m
m
6
4
T
4
5
T
8
R
3
6
T
6
R
6
R
6
0
T
5
2
T
1
0
R
8
R
4
8
T
6
R
3
6
T
6
R
1
2
7 8
"
1
4
9
.
2
m
m
0
1
0
3
0
5
0
7
0
9
0
1
0
1
7
2
6
-
8

R
e
g
u
l
a
t
o
r

G
e
n
e
r
a
l

A
r
r
a
n
g
e
m
e
n
t
8
x
9
x
3
.
6
x
S
c
a
l
e

1
:
4
2
9
4
1
2
3
4
5
6
7
8
9
1
1
1
0
1
2
1
3
1
4
1
5
16 17
18
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3 1
3 0
4
0
5
0
6
0
1
0
2
0
30
4
0
5
0
6
0
1
0
2
0
1
5
0 10 30 50 70 90 10
40T
20T
30T
45T
48T
6L
1
2
3
4
5
6
7
8
9
1
1
1
0
1
2
1
3
1
4
1
5
16 17
18
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3 1
Pitch P
5P/7
3P/7
78.4 mm
9.5 mm
1726-8 Regulator Motion Work and Escapement
24.71 mm
32.20 mm
76.2 mm
0 10 20 30 40 10
The dimensions in the drawings in the 1730 document have been compared to the published dimensions of one of the
1726-8 regulators. The escapement in this clock at present is a replacement which was most probably copied from the
Guidhall regulator.
Pallet lengths and escapement circle as calculated.
Scale 1:4
295
1728 Wheel Details
GOING TRAIN
5p/7 3p/7 Chordal Radius Pitch No. Teeth Dist. or Radius
90.06 Great Wheel
5.61 3.37 80.08 7.86 64 80.05 64 tooth Wheel
10.27 8 10.01 8 tooth Pinion
60.25 Centre Wheel
5.30 3.18 53.21 7.42 45 53.17 45 tooth Wheel
7.42 6 7.09 6 tooth Pinion
49.00 Third wheel
5.24 3.14 42.05 7.33 36 42.00 36 tooth Wheel
7.33 6 7.00 6 tooth Pinion
3.60 60 68.76 Escapement Wheel
STRIKING TRAIN
5p/7 3p/7 Chordal Radius Pitch No. Teeth Dist. or Radius
94.12 Great Wheel
6.03 3.62 80.71 8.45 60 80.67 60 tooth Wheel
13.67 10 13.45 10 tooth Pinion
70.79 Centre Wheel
5.30 3.18 61.39 7.41 52 61.35 52 tooth Wheel
9.69 8 9.44 8 tooth Pinion
53.14 Third wheel
4.42 2.65 47.27 6.18 48 47.23 48 tooth Wheel
6.18 6 5.90 6 tooth Pinion
40.14 Warning Wheel
4.29 2.57 34.45 6.00 36 34.40 36 tooth Wheel
6.00 6 5.73 6 tooth Pinion
5p/7 3p/7 Chordal Radius Pitch No. Teeth Dist. or Radius
59.30 Hour Wheel
4.75 2.85 50.86 6.65 48 50.83 48 tooth Wheel
8.69 8 8.47 8 tooth Pinion
70.46 Calendar Wheel
5.02 3.01 47.02 7.03 42 46.97 42 tooth Wheel
23.58 21 23.49 21 tooth Pinion
Appendix A.
Some Notable Names in Isaac Wares
Subscribers List
Isaac Wares translation of
Palladios Four Books of
Architecture was dedicated to
and revised by Richard
Boyle, 3
rd
Earl of
Burlington. He inherited
Burlington House, home of
the Royal Astronomical
Society in 1715. Following
his Grand Tour he fostered
the Neo-Palladian style,
building Chiswick House
with the assistance of
William Kent.
James Douglas, 14
th
Earl of Morton(1702-1768) Scottish representative
peer who became President of Royal Society in 1764. Antagonist
of Harrison.
Thomas Watson-Wentworth, Earl of Malton, (1693-1750) M.P. for
Malton, Yorkshire 1715-27, Lord Lieutenant of Yorkshire 1733-1750
Galfridus Mann (1706-1756), an Army clothier in London for whom
Walpole procured a military contract.
Sir William Stanhope, 2
nd
Earl of Harrington, (1719-1779) M.P. for
Aylesbury and Bury St. Edmunds.
William Stanhope, Rt. Hon. Lord Harrington, (1683-1756) Secretary
of State for the Northern Department, and Lord President of Council.
Political ally of the Pelhams. Minister in Walpoles government. Half
brother of Philip Stanhope, Earl of Chesterfield and brother of Charles
Stanhope.
Francis Hillyard , carpenter - development of Berkeley Square.
297
Sir Robert Corbett, one of the Commissioners of the Customs.
Lord Charles Cavendish, Vice President of the Royal Society, instrument
maker who was awarded the Copley Medal in 1757 for his thermometer.
Thomas Coke, Vice chamberlain of the Household, Privy Councillor who
became 1
st
Earl of Leicester.

Thomas Ripley, 1683-1758, ex carpenter who became an architect in
Office of Works. Worked on site at Houghton Hall for Sir Robert
Walpole, succeeding Grinling Gibbons as master carpenter. In 1726
succeeded Vanbrugh as Comptroller of the Kings Works. Old Admiralty
Building in Whitehall now known as Ripley Building.
Gustavus Hamilton, 2
nd
Viscount Boyne, (1710-1746), M.P. for
Newport, Isle of Wight.
Mrs. Mary Chandler, (1687-1745), poet, friend of Pope.
Lord Charles Cornwallis, (1700-1762), lived in Grosvenor Square. His
wife, Elizabeth Townsend was daughter of 2
nd
Viscount Townsend and
niece of Robert Walpole.
Lord Viscount Townsend (1674-1738) Secretary of State. First wife ,
Elizabeth Pelham, second wife Walpoles sister.
Thomas Townsend, (1701-1780) M.P. for Winchester. Secretary to
father.
Sir John Evelyn, (1706-1767), grandson of the diarist, M.P. for Helston,
Cornwall.
Thomas Gladwin (1710-1799) composer and musician, organist at
Vauxhall Gardens.
Dr. Edward Hody, (1698-1759), obstetrician and member of Royal
Society. He published a revision of Giffords Cases in Midwifery in 1734.

William Watson, 3
rd
Earl Fitzwilliam (1685-1756) married Ann
Wentworth, daughter of the Marquis of Rockingham. Built Wentworth
Woodhouse, architect Henry Flitcroft.
Thomas Belasyse, 4
th
Viscount Fauconberg (1699-1774) created Earl
Fauconberg in 1756 and was a predecessor of the Earl of Yarborough of
298
Brocklesby House.
Francis Hayman (1708-1776) painter and illustrator. Painted scenery for
Vauxhall Gardens in the 1740's. Influential in the introduction of the
rococo style to English painting and a founder member of the Royal
Academy.
Thomas Gray (1716-1771) poet with sense of style, humour and beauty.
Friend of Walpoles son, Horace. Writer of the Elegy.
Westby Gill (1679-1746), architect and master carpenter who succeeded
William Kent. Responsible for the circular Ionic Temple at Shirburn
Castle, Oxon.
John Boyle, 5
th
Earl of Orrery, (1707-1762) Fellow of the Royal Society, friend
of Alexander Pope, Jonathan Swift and Samuel Johnson.
Graham made an orrery for the 4
th
Earl (1674-1731).
John Montagu, 2
nd
Duke of Montagu, (1690-1749), Fellow of the Royal Society
and a founding governor of the Foundling Hospital.
Sir Henry Bedingfield, brother-in-law to Lord Burlington.
William Kent, (1685-1748), painter and architect. Edited Designs of Inigo Jones
in 1727.
Matthew Brettingham (1699-1769), architect who supervised the building of
Holkham Hall and created a viable interior arrangement for town houses based
upon the Palladian villas.
Kellom Tomlinson (1690-1753), dancing master who wrote The Art of Dancing
published in 1735.
Willem Anne van Keppel, 2
nd
Earl of Albemarle (1702-1754), a founding
Governor of the Foundling Hospital, Governor of Virginia, 1737-1754.
William Hogarth (1697-1764) engraver and painter.
Samuel Johnson (1709-1784) essayist and lexicographer, who wrote for the
Gentlemans Magazine.
Henry Flitcroft (1697-1769), nicknamed Burlington Harry. Master carpenter
and mason who became Comptroller of Works in 1758.
299
Jonathan Sisson, (1690-1747), instrument maker who made backstaffs before the
development of the sextant.
Henry Pelham (1694-1754), Prime Minister 1743-54, succeeded by his brother,
Duke of Newcastle-upon-Tyne. A founding governor of the Foundling Hospital,
who introduced the Calendar Act and the Marriage Act 1753, which legislated a
minimum age of consent for marriage.
Henry Fox (1705-1774) Secretary of War and Paymaster of the Forces. Father
of Charles James Fox. He was succeeded at the War Office by William Barrington,
2
nd
Viscount Barrington, a loyal supporter of Harrison.
Sir Andrew Fontaine (1676-1753) amateur architect who succeeded Newton as
Warden of the Mint.
Thomas Heath, instrument maker, Secretary of the Royal Society.
George Parker, 2
nd
Earl of Macclesfield, (1695-1764), Astronomer who observed
at Shirburn Castle. President of the Royal society and Vice President of the
Foundling Hospital. Collected many of Newtons manuscripts and letters. These
documents and the Portsmouth collection are now in the University of Cambridge
Library.
William Jones (1675-1749), Vice President of the Royal Society, mathematician
who after private tutoring in North Wales moved to London under the patronage
of Lord Macclesfield. He was a close friend of Newton and Halley. Introduced the
symbol pi in Synopsis Palmariorum Matheseos 1706. Published A New Compendium on
the Whole Art of Navigation before 1706.
300
Appendix B
Musical Temperament
The second half of Concerning Such Mechanism is devoted to Harrisons
interest in music, particularly musical temperament. A lot of the material is the
same as that found in the manuscript A True and Full Account of the Foundation of Music
in the Library of Congress, Washington. In these documents he describes tuning
bells, training church choirs and his friendship with Dr. Robert Smith, (1689-
1768), Master of Trinity College, Cambridge. The friendship ended when Smith
published Harmonics, or, The Philosophy of Musical Sounds , in 1749 because Harrison
considered that Smith had plagiarized his work on counting beats when setting up
a temperament. (n.b. Harrison was awarded the Copley Medal in 1749.) Sir John
Hawkins (1719-1789) describes Harrisons assistance to Smith in his General
History of the Science and Practice of Music published in 1776. William Harrison
continued the musical tradition; his English first edition of Haydns Creation is
in the library of the Wiener Musikverein, together with a invitation ticket for a
subscription concert that he arranged in Williss Rooms, Hanover Square on Friday
18
th
May 1792.
Harrisons development of musical temperament was influenced by his
knowledge of bells, tuning of his viola da gamba, and vocal training. Bells resonate
in two modes, longitudinal and circular. They have five dominant harmonics of
which three are in unison, a minor third and a fifth. The viola da gamba has tied
frets like a lute which means that the sharps and flats are not enharmonic, i.e. they
are the same. The players of fretless string instruments like the violin family,
players of wind instruments and singers can slightly adjust the pitch of the sharp
and flat notes to make them sound in tune. This is impossible with keyboard
instruments, so that a compromise has to be sought.
The problem is caused by the Pythagorean comma. The ratio of pitch for
a pure fifth is 3/2 and an octave 2/1. We would therefore consider by counting the
notes that twelve fifths would equal seven octaves returning to the original note.
= =



12 7
3 2
129.746 128.0
2 1
Unfortunately the resulting pitch has a ratio of 129.746:128, 1.014:1 to
the primary pitch amounting to approximately a quarter of a semitone too high. If
the musician is to be enabled to play in all keys i.e. starting the diatonic scale on any
301
note, this error or comma has to be distributed between the notes. The thirds of
the Pythagorean scale, having its complex ratios, sound too sharp. The just
temperament has the ratios adjusted to give smaller whole numbers and the notes
are approximately the 8
th
to 16
th
harmonic of the base note C, omitting only the
14
th
harmonic. This just scale was proposed by the astronomer Ptolemy (Claudius
Ptolemaeus, 70-147), suggesting that the best tuning is where the theoretical ratios
and the ear are in agreement.
In order to play in all keys the error can be equally divided between all the
twelve notes, as in the now universally used Equal Temperament, making the major
thirds slightly sharp and the fifths slightly flat. Three thirds will now make an
octave. All keys sound equally bad and out of tune. Equal temperament was
popularized by the Scottish piano maker John Broadwood (1732-1812).
During the eighteenth century many mean tone temperaments were devised
to provide a compromise to this problem resulting in a restriction of usable keys
because of wolf notes, that is notes in the temperament which are markedly out of
tune. A mean tone temperament allows three differently tempered thirds to make
up an octave. J.S.Bach chose to flatten six fifths from C and then flatten the
remaining fifths to a lesser extent or not at all. He placed cryptic tuning
instructions at the top of the front page of the Wohltempierte Clavier ; the number
of turns in the diagram indicate the degree of flattening of the fifths. This
produces an irregular mean tone scale having five 1/6th comma fifths, three pure
fifths and three 1/12 comma fifths.
Harrison devised his scale by sharpening the major thirds of the scale so
that they had the ratio . This produced a regular mean temperament scale
1
2 :1

where there were eleven fifths having the same flattened interval with one producing
a wolf note.
Let the Ratio of the Octave, or, as even here, as well below to the Purpose, the Octave
itself, be represented by the Logarithm of 2 [viz. by ,30103]; and let that same number be also
taken or supposed as the Circumference of a Circle. And then [as in the Margin] let the Space
of Quantity of two Octaves and a sharp 3
rd
be taken, or be as chiefly, or rather as primary to
the Purpose notified, viz. when [as according to my Discovery] the said sharp 3
rd
is in its most
strictly musical Proportion, and that is as when , with Respect to the Octave, the same is taken
as the Diameter of the Circle [viz. here, as ,09582]: For the Proportion which the
Circumference of a Circle bears to the Diameter (and as true enough to this Purpose, as well
as to others) is about as 3,1416 to 1: So, as 3,1416 is to 1, so is ,30103 to ,09582.
302
And then, as five larger Notes [but not with Tones Major and Minor,
as hath been imagined, and that from of old] and as together with two of the
lesser Notes [as all along foolishly styled half Notes Major] are, or must be,
exactly contained in the Octave: So, therefore, as in taking Half the Diameter
for the larger Note, viz. ,04791, as I from strict, or proper Experience, found
it to be - as an Interval of Melody, right truly pleasant [although, as barely in itself, as well
as the lesser Note, nothing to do with Harmony], and that four 5
ths
, thence as below to be
generated [viz. of each containing ,17447], and as when, as I am Proof sure, to be then near
in their most strictly musical Proportion, will, as according to Nature, be equal to the two
Octaves and a sharp 3
rd
; and at the same Time, as already intimated, each one of the four 5
ths
will also be as without any Infringement in any Case [viz. as with Respect to the Product of
Nature] so generated by subtracting five times the Radius from the Circumference, where there
will be left such a Quantity or Space, as the two lesser notes must with equal Shares take up;
and that will be ,06148, so the Half of which, viz. ,03074, must be the lesser Note; and the
lesser Note subtracted from the greater will leave ,01717, properly to be called a Flat or a
Sharp [or the Difference of the Notes], and not nonsensically the Half-Note Minor; the lesser
Note having withal the same Authority to be called a whole Note as what the other has; but they
may respectively or properly be styled Tone Major and Tone Minor, viz. without meaning the
fictitious Nonsense as of old; and (as well understood) a 5
th
must contain three of the larger
notes and one of the lesser [viz. as in the Case or Cases here ,17447].
But as notwithstanding, that from what is here above, are indeed the real
Steps or Intervals of Tune, or of natural Melody, exactly pointed out, or are to
be as thence truly generated [viz. according as they are taken by the Voice or by
Voices]; so there must, as in Consequence thereof, be also the real Consonances,
or Chords of natural Harmony, truly limited or described;\ nay as so, in both
Respects [viz. as touching both Melody and Harmony] I found, to my great
Surprize, to be confirmed upon strict Instrumental Musick, as I have shewn above.
John Harrison, Concerning Such Mechanism, 1775

Harrison describes the theory of his temperament but does not give any
indication of how he implemented it. The description is difficult for us to read
because the term semitone had not been defined when it was written. Dr. Smith
describes the implementation and his criticism in the Preface to his Harmonics.
He told me he took a thin ruler equal in length to the smallest string of his base viol. and
divided it as a monochord, by taking the interval of the major IIId, to that of the VIIIth, as the
diameter of a circle, to its circumference. The by the divisions on the ruler applied to that string,
he adjusted the frets upon the neck of the viol. and found the harmony of the consonances so
extremely fine that after a very small and gradual lengthening of the other strings at the nut,
by reason of their greater stiffness he acquiesced in that manner the placing of the frets.
It follows from Mr. Harrison's assumption that his IIId major is tempered flat by

,17447
4
,69788
,30103
2
,60206
,09582
,69788
303
a full comma. My IIId determined by theory upon the principle of making all the concords
within the extent of every three octaves as equally harmonious as possible, is tempered flat by
one ninth of a comma; or almost one eighth, when no more concords are taken into the
calculation that what are contained within one octave.
That theory is therefore supported on one hand by Harrison's experiment, and on the
other by the common practice of musicians, who make the major IIId either perfect, or generally
sharper than perfect, with a design I suppose, to improve the false concords, though to the
manifest detriment of all the rest. We may gather from the construction of the base viol, that
Mr Harrison attended chiefly, if not solely to the harmony of the consonances contained within
the octave; in which case the difference between his and my temperaments of the Major IId, VIth
and Vth and their several dependents, are respectively no greater than 4, 3 and 1 fiftieth parts
of a comma. And considering that any assigned differences in temperaments of a system, will
have the least affect in altering the harmony of the whole when at the best, I think a nearer
agreement of that experiment with the theory could not be reasonably expected.
Upon asking him why he took the interval of the major IIId to that of the VIIIth as
the diameter to the circumference of a circle, he answered that a gentleman lately deceased had
told him it would bring out the best division of a monochord. Whoever was the author of that
hypothesis for so it must be called, he took the hint, no doubt, from observing that as the octave,
consisting of five meantones and two limmas is a little bigger than six such tones, or three
perfect major IIIds, so the circumference of a circle is a little bigger than three of its diameters.
When the monochord was divided upon the principle of making the major IIId perfect, or but
very little sharper, as in Mr Huygen's system resulting from the octave divided into 31 equal
intervals,
Mr. Harrison told me that the major VIths were very bad and much worse than the
Vths and VIths major when equally tempered should differ so in their harmony, after various
attempts I satisfied my curiosity; and this gave me the first insight into the theory of imperfect
consonances. With a view to some other inquiries I will conclude with the following
observation. That, as almost all sorts of substances are perpetually subject to very minute
vibrating motions, and all our senses and faculties, seem chiefly to depend upon such motions
excited in the proper organs, either by outward objects or the power of the will, there is reason
to expect that the theory of vibrations here given will not prove useless in promoting the
philosophy of other things besides musical sounds. Such readers as can only dip into this treatise
must remember, that by the word vibration so often repeated I mean the time of a single
vibration, which I notified once for all in sect I art. 8 31/12/1748.
If two musical strings have the same thickness, density and tension and differ in length
only, (which for the future I shall always suppose) mathematicians have demonstrated that the
times of their vibrations are proportional to their lengths (f). Hence if a string of a musical
instrument is stopt in the middle, and the sound of the half be compared to the sound of the whole,
we may acquire the idea of the interval of two sounds, whole single vibrations (always meaning
the times) are in the ratio of 1 to 2; and by comparing the sounds of 2/3, 3/4, 3/5, 4/5, 5/6.
8/9, 9/10, etc of the string with the sound of the whole, we may acquire the ideas of the intervals
of the two sounds, whole single vibrations are in the ratio of 2 to 3, 3 to 4, 3 to 5, 4 to 5, 5 to
6, 8 to 9, 9 to 10, etc.
Robert Smith Harmonics, or, The Philosophy of Musical Sounds, 1749
304
Hermann von Helmholtz (1821-1894) was a German physician and
surgeon who was appointed Professor of Physics at Berlin in 1871. In 1887 he
founded the first institute of pure scientific research, the Physikalisch-
Technische Reichsanstalt in Berlin. He had previously been appointed Professor
of Medicine at Bonn and Heidelberg. Whilst at Bonn he began his interest in
acoustic research resulting in the publication of his classic text Die Lehre von
dem Tonempfindungen (On the Sensations of Tone) in 1863. Helmholtz makes
no reference to Smith or Harrison in this treatise which describes the nature of
beats and their role in consonance and dissonance, the role of harmonics in
timbre, the nature of combination tones and summation tones. He did refer to
Smiths A Compleat System of Opticks published in 1738.
For a single refracting surface of sufficiently small aperture, the product of the index of refraction,
distance from the optical axis, and the angle which a light ray makes with the optical axis at the
object point is equal to the corresponding product at the image point. This equation is known
as the Smith-Helmholtz equation in Britain and the Lagrange equation in
continental Europe.
A.J. Ellis translated On the Sensations of Tone into English in 1875
in his writings suggested that twelve equal semitones to the octave can each be
given a value of 100 degrees or cents. George Bernard Shaw acknowledged him
in his portrayal of Professor Henry Higgins, the fictional character in Pygmalion
or My Fair Lady . The octave is divided into 1200 cents on a logarithmic scale
derived as:-
or more conveniently
2 1 2
1200 log ( / ) = C f f
10 1 2 10 1 2
10
1200
log ( / ) 3986.314 log ( / )
log 2
= = C f f f f
The similarity to the logic used by Harrison in describing his
temperament to the modern cent system is striking.
305
306
Bibliography and References.
Preface
Thomas S. Kuhn, The Structure of Scientific Revolutions, Univ. Chicago, ISBN 0 226 45807 5
John Harrison, 1730 Manuscript, MS 6026/1 Guildhall Library
Allan Cook, Edmond Halley, Clarenden Press, ISBN 0 19 850031 9
Andrea Palladio, Four Books of Architecture, Dover Books, ISBN 0 486 21308-0
Timothy Mowl, William Kent, Pimlico, ISBN 978 184 4135394
John Hadley, Hadleys Octant, Phil. Trans. Roy. Soc., (1731), v. 37, article 25, p. 147
Chapter 1 John Harrisons Astronomical Regulator
Samuel Smiles, Men of Invention and Industry, Bibliobazaar. ISBN 1 4264 0365 8
Humphry Quill, John Harrison, 1965, John Barker
Dava Sobel, Longitude, Fourth Estate ISBN 1 85702 502 4
George Rude Hanoverian London Sutton ISBN 0 7509 3333 X
Paul Langford, A Polite and Commercial People OUP 978-0192852533
Jenny Uglow Hogarth Faber and Faber ISBN 0 571 19376 5
Liza Picard Dr Johnsons London Weidenfeld and Nicolson ISBN 0 297 84218 8
Derek Howse, Nevil Maskelyne: The Seaman's Astronomer. CUP ISBN 0978 0521 36261 0
A.D.Patel Language, Music, Syntax and the Brain Nature Neuroscience (2003) v.6(7) pp. 674-681
Explanatory Supplement to the Astronomical Ephemeris and Nautical Almanac. HMSO 1961
T.S. Lecky, Wrinkles in Practical Navigation Elibron Classics ISBN 1-4021-4885-2
Philip Woodward Woodward on Time Bill Taylor and BHI ISBN 0 95096216 3
Philip Woodward My Own Right Time OUP ISBN 0 19 856522 4
Papers of William Stukeley, Bodleian Library MSS Eng. misc. b. 65-7 etc.
W.M.H.Greaves & L.S.T.Symms Short Period Erratics of Free Pendulum Clocks MNRAS (1943) v.103
pp196-209
Johann Bayer Uranometria, omnium asterismorum.. British Library, Nat. bib. no. GB8950094
John Flamsteed 1729 Atlas Coelestis British Library 001248649 (48.i.17)
Gillian Wagner Thomas Coram, Gent Boydell Press, ISBN 1 84383 057 4
William Hogarth, The Analysis of Beauty 1753, Brit. Lib. General Reference Collection 555.c.36
Paul Glennie and Nigel Thrift Shaping the Day OUP ISBN 978 0b 19 927820 6
Christiaan Huygens De Ratiociniis in Ludo Aleae 1714 tr. W.Browne
Anders Hald History of Probability and Statistics Wiley ISBN 978-0471502302
Chapter 2 Documentary Sources
J.R.Smith John Harrison mezzotint engraving, Science Museum, Kensington
J.Wright John Harrison photograph of painting, Derby Museum and Art Gallery
J.Wright A Philosopher giving that Lecture on the Orrery painting, Derby Museum and Art Gallery
Thomas King John Harrison painting, National Maritime Museum
John Harrison, 1730 Manuscript, MS 6026/1 Guildhall Library
John Harrison Concerning Such Mechanism 1775, ESTC 006056929 British Library
John Harrison An Explanation of my Watch 1763, MS 3972/2 Guildhall Library
John Harrison A True and Full Account of the Foundation of Music US Lib. of Congress MS
Thomas Bradley et al Drawings of H2 and H3' MS RGO 6/586f213-217 Cambridge University Lib.
Sotheran Auction Catalogue Bibliotheca Chemico-Mathematica US Lib. of Congress
Nicholas Saunderson Lecture Notes British Library Add.57488
307
Rupert T. Gould Rough Notebook RAS Harrison Clock James Caird Library, Greenwich GOU/15
Charles H. Cotter A History of Nautical Astronomy Hollis and Carter ISBN 370 00460 4
R.T.Gould Marine Chronometer Holland ISBN 0-900-470-17-8
W.M.Smart Textbook on Spherical Astronomy CUP ISBN 0-521-06491-0
Edmund Beckett, Lord Grimthorpe Clocks, Watches and Bells General Books ISBN 978-1459013797
Johann Horrins Memoirs in a Trait in the Character of George III Kessinger ISBN 0548564922
Chapter 3 Development of his Pendulum Clocks
Herbert Jttemann Die Schwarzwalduhr Badenia ISBN 3 7617 0360 0
Derek Roberts Precision Pendulum Clocks Vol.1 Schiffer ISBN 0 7643 1636 2
Rupert T Gould John Harrison and his Timekeepers 1958 National Maritime Museum
Rupert T Gould Marine Chronometers Holland Press ISBN 0-900-470-17-8
Rupert T. Gould Restoration of John Harrisons Third Timekeeper HJ March-July 1932
Don Unwin A Replica of John Harrisons Sea Clock H3' HJ September 2002- January 2003
Mackensen, Bertele & Leopold Die erste Sternwarte Europas Callwey ISBN 3-7667-0875-9
John Campbell Ships Log HMS Essex 1758-9 Public Record Office Miscellanea (H.C.A. 30)
ed William J.H.Andrewes The Quest for Longitude Harvard University ISBN 0-9644329- 0-0
Jonathan Betts Time Restored OUP ISBN 978-0-19 856802-5
Karsten Gaulke et al Der Ptolemus von Kassel MHK ISBN 978-3-931787-43-1
David Thompson Clocks British Museum ISBN 0-7141-2812-0
Allan Chapman Englands Leonardo Institute of Physics ISBN 0-7503-0987-3
Chapter 4 Royal Astronomical Society Regulator
E.T.Cottingham A Description of the RASs Harrison Clock MNRAS May 1909
William S. Laycock The Lost Science of John Longitude Harrison Brent Wright ISBN 0-903512076
K.G.McLare & D.Tabor The Frictional Properties of Lignum Vitae Br.J.App.Phys, (1961), v. 12, pp118-
120
Rupert T. Gould The Societys Harrison Clock MNRAS 1929, Vol.89, pp. 398-401
Chapter 5 Logical Design
Vitruvius The Ten Books on Architecture Dover ISBN 486-20645 9
J.A.Bennett The Mathematical Science of Christopher Wren CUP ISBN 0-521-24608 3
Matilda Ghyka The Geometry of Art and Life Dover ISBN 0-486-23542-4
Richard Padovan Proportion Taylor and Francis ISBN 0-419-22780-8
Albrecht Drer Unterweisung der Messung Dr Alfons Uhl ISBN 3-921503-17-5
Albrecht Drer Vier Bcher von menschlicher Proportion Nordlingen ISBN 978-3921503560
Jrme Lelande Diary of a Trip to England 1963 tr. Richard Watkins ISBN 0-9581369-1-2
John Robertson A Treatise of Mathematical Instruments Flower-de-Luce ISBN 1-931468-1107
Chapter 6 Oscillator
J.A.Bennett The Mathematical Science of Christopher Wren CUP ISBN 0-521-24608-3
Stephen Hawking God Created the Integers Running Press ISBN 978-0-7624-1922-7
Stephen Hawking On the Shoulders of Giants Running Press ISBN 0-7624-1348-4
Christiaan Huygens Horologium Oscillatorium 1678
A.E.Bell The Horologium Oscillatorium of Christian Huygens Nature (1941) v. 148 pp.245-248
G.Rousseaux et al Robert Hookes Conical Pendulum from the Modern Viewpoint of Amplitude Equations and its
Optical Analogues Proc. Roy. Soc. A (2006) v.462 pp.531 - 540
T.W.B.Kibble & F.H.Berkshire Classical Mechanics Imperial College Press ISBN 186-0944248
308
S.T.Thornton & J.B.Marion Classical Dynamics Thomson ISBN 0-534-40896-6
G. L. Baker & J. A. Blackburn The Pendulum, a Case Study in Physics OUP ISBN 0-19 856754-5
ed. Michael R. Mathews et al The Pendulum Springer ISBN 978-1-4020-3525 8
Robert J. Matthys Accurate Clock Pendulums OUP ISBN 0-19-852971-6
Moshe Gitterman The Noisy Pendulum World Scientific ISBN 978-981-283-299-3
Mervyn K. Hobden Harrison and Airy HJ (1990) v.133 pp.74-75
James Gleick Chaos, Making a New Science William Heineman ISBN 978-0434295548
G.L.Baker & J.B.Gollub Chaotic Dynamics CUP ISBN 0-521-47106 0
R.Gilmore & C.Letellier The Symmetry of Chaos OUP ISBN 978-0-19 531065-8
M.K.Hobden John Harrison, Balthasar van der Pol and the Non-linear Oscillator HJ (1981) v.124 pp.16-
F.R.Moon & P.D.Stiefel Co-existing Chaotic and Periodic Dynamics in Clock Escapements Phil. Trans.
Roy. Soc. (2006) v.364 pp. 2539-2564
Mark Denny The Pendulum Clock: a venerable dynamical system Eur. J. Phys. (2002) v.23(4), pp.449-
458
T.W.B.Kibble &F.H.Berkshire Classical Mechanics Imperial College ISBN 1860944248
John Shelton Mr Grahams Escapement Gentlemans Magazine 1754
Joseph-Louis Lagrange Mecanique Analytique (1811) CUP 2009 ISBN 978-1108001748
J.Clerk Maxwell On the Viscosity of Air and Other Gases Phil.Trans.Roy.Soc. (1866) v.156 pp.249-268
W.N.Bond The Viscosity of Air Proc. Phys. Soc. (1937) 49 205
G.K.Batchelor An Introduction to Fluid Dynamics CUP ISBN 978-0-521-66396-0
Henry Kater An account of experiments for determining the length of the pendulum vibrating seconds in the latitude
of London Phil.Trans.Roy.Soc. (1818) v. 104 pp.83-85
F.W.Bessel Untersuchungen ber die Lnge des einfachen Secundenpendels Abhandlungen der Berliner
Akademie d. Wis. Math. Cl. (1826) pp. 1-172
D.Candela, K.M.Martini, R.V.Krotkov & K.H.Langley Bessels improved Katers pendulum in the teaching
laboratory Am.J.Phys (2001) v. 69(6) pp. 714-720
D.W. Allan Time and Frequency, Characterization, Estimation, and Prediction of Precision Clocks and Oscillators,
IEEE Transactions (1987) UFFC-34, pp.647-654
George G. Stokes On the Effect of the Internal Friction of Fluids on the Motion of Pendulums Trans.Camb.
Phil.Soc. (1850) v.9, p.8
J. M. Bloxam On the mathematical theory and practical defects of clock escapements, with a description of a new
escapement Memoirs R. Astron. Soc.(1854) v. XXII, pp.103150
A.L.Rawlings The Science of Clocks and Watches BHI ISBN 0-950-9621-3-9
K.James Precision Pendulum Clocks - Circular Error and the Suspension Spring Antiq. Horol. v.8 pp 868-
883
George B. Airy On the Disturbances of Pendulums and Balances and on the Theory of Escapements Camb. Phil.
Trans. (1927) v.3 part 1, pp. 105-128
ed Abraham Rees Clocks, Watches and Chronometers 1819-20' David & Charles ISBN 7153-4915-5
R.A.Nelson & M.G.Olsson The Pendulum - Rich Physics from a Simple System Am. J. Phys.(1986)
v.54(2) pp.112-121
Humphrey Quill The Grasshopper Escapement Antiq. Horol. (1971)v.7(4) pp.288-296
Martin Burgess The Grasshopper Escapement - its Geometry and its Properties Antiq. Hor. (1971) v.7(5)
pp. 416-422
Henry Casson The Geometry of the Grasshopper Escapement - a Recalculation of Burgesss Figure 4' NAWCC
Science Chapter Newsletter (2000) v.1 pp. 22-24
Guy D. Aydlett The Anatomy of the Grasshopper NAWCC Bulletin Supplement April 1972
Philip Woodward A New Look at Escapement Theory HJ (1989)v.131 pp.9 -
Kenneth James The Harrison Grasshopper Escapement HJ (1985) v.127 pp.25 -
Peter R.Hastings A Look at the Grasshopper Escapement HJ (1993) v.136 pp.48-53
309
Peter R.Hastings Another Look at the Grasshopper Escapement HJ (1996) v.138 pp.412-3
Philip Woodward Harrisons Weighted Escapement Torque HJ (1997)v.139 pp.196-197
Jeffrey Short Harrisons Grasshopper Escapement NAWCC Bulletin, August 1998
Chapter 7 Power Supply
D. Tabor The Friction Properties of Some White Metal Bearing Alloys, the Role of the Matrix and the Hard
Particles J.Appl.Phys. (1945) v.16, pp.325-337
Peter Hastings John Harrisons Roller Gearing HJ (1981)v.123 pp.17-20
F.P.Bowden & D. Tabor "The Friction and Lubrication of Solids. Part" (1986) ISBN 978-0198520269
James Newlands The Carpenters Assistant (1860) rep. Bracken Books ISBN 1-85170-933-9
Jeremy Thacker The Longitude examined,..... British Library 530.e.8.(1.)
M.Ashby, H.Shercliff & D.Cebon Materials, Engineering, Science, Processing and Design Butterworth-
Heineman ISBN 978-0-7506-8391-3
A.L.James Harrisons Remontoire HJ (1993) v. 136 pp.147-148
Peter R Hastings John Harrison - Spacecraft Engineer? HJ (1993) v. 136 pp. 193-7
Chapter 9 Adjustment and Errors
D.A.Bateman Vibration Theory and Clocks HJ (1977) v.120 pp.3 -
D.A.Bateman The Accuracy of Pendulum Timekeepers HJ (1989) v.132 pp.83 -
D.W.Allan, N.Ashby & C.C.Hodge The Science of Timekeeping H.P. Application Note 1289
Anon GPS and Precision Timing Applications H.P. Application Note 1272
P.H.Boucheron Just How Good Was the Shortt Clock HJ (1985) v.128(6) pp.10-14
Appendix B
J.S.Bach Well Tempered Clavier, 48 Preludes and Fugues Breitkopf & Haertel
Charles A.Padgham The Well Tempered Organ Positif ISBN 0-906894-13-1
Bradley Lehman Bachs Extraordinary Temperament Early Music (2005) v.33(1)pp. 3-24, (2)pp.211-231
Hermann Helmholtz On the Sensations of Tone Dover ISBN 0-486-60753-4
Robert Smith Harmonics, Or the Philosophy of Musical Sounds Bibliolife ISBN 978-1-113-91958-8
Dave Benson Music, a Mathematical Offering CUP ISBN 978-0521853873
C.E.H.Lucy Pitch,Pi, and Other Musical Paradoxes Music Teacher Magazine, London January 1988

Typeset in Bruce Rogers Centaur as prepared for the Metropolitan Museum, New York using
WordPerfect.
Drawings created with TurboCAD.
310