Metering of / Eai \

Symmetrical Components
lai Eai

kilöJL

Methods of metering the positive and

negative sequence power and energy in Fig. 1. Positive sequence voltage

3-phase 3-wire systems, and positive, nega-

tive, and zero sequence power and energy Certain metering networks (see references 2, 12, and
13) have been proposed for measuring the voltage
in 3-phase 4-wire systems, are discussed in
and current components of these symmetrical sys-
this article. It is presented not because it tems. Relays have been developed which function
contains a wealth of new material, but be- in response to the negative sequence current; volt-
cause it deals comprehensively with the age regulators utilize the positive sequence voltage
of a system, rather than one of the line voltages.
problem, in a way different from that which No doubt these symmetrical components, particu-
has prevailed in earlier published material, larly the power components, will be put to many
and because it is clear, specific, and readily other uses in electrical control and metering equip-
ment when the limitations, errors, and cost of the
understandable, confining itself strictly to
various networks required to separate them are more
the problem of metering of symmetrical extensively known.
components. It has been shown in these other articles referred to,
that when no zero sequence components are present,
networks may be set up which will meter simul-
By
taneously and separately the positive and negative
sequence voltages as shown in Fig. 3, and the positive
G . R. SHUCK University of
MEMBER A.I.E.E. Washington, Seattle and negative sequence currents as shown in Fig. 6.
It is thought advisable, however, to go into con-
siderable detail in the analysis of the voltage and

c
current networks, to establish more clearly certain
facts useful in the explanation of power and energy
ίERTAIN NET WORKS designed to metering networks.
meter simultaneously and separately the positive In order to simplify the vector diagrams and nota-
and negative sequence components of power or tion, the secondary voltages and currents of all the
energy are described in this article. Tables are instrument transformers are assumed to have the
included showing the location and number of instru- same value and phase position as the corresponding
ment transformers required for each network, and primary values. Obviously, this notation will not
equations involving the design of the impedances of affect the result since all the vectors will be reversed.
the metering systems are given. It is suggested The following symbols will be used in this article :
that the components of power be utilized in relays
E , E , Eb, EQ phase voltages of an unbalanced system
and control equipment, as well as the components he, a c

lea, lab line currents of a 3-phase 3-wire system

of current and voltage. Eai, E i, Eu positive sequence voltages of the system
c

Contents of previous publications on metering of EE , zero a2 E , Eb2 negative sequence voltages of the system
c2

sequence voltage of the system

symmetrical components have been confined largely Io zero sequence current of the system
0

to the components of current and voltage in 3-phase Ibci, Icai, Iabi positive sequence line currents of a 3-phase system
3-wire systems. The method of resolving an un- hci, Icaï, Iabi negative sequence line currents of a 3-wire system
I*, h, h, h line currents of a 3-phase 4-wire system
balanced 3-phase system into the 3 symmetrical Z \, Z \ impedances of the positive sequence potential network
a c

components—a balanced 3-phase positive-sequence Zai, Zci impedances of the negative sequence potential network
R internal resistance in series with the moving element of a stand-
system, a balanced 3-phase negative-sequence sys- ard m

wattmeter or voltmeter
tem, and a uniphase or zero-sequence system—has Z impedance equal to that of the current coil of the wattmeter or
m

been described in a number of articles in current ammeter

R effective resistance of the potential coil of the watthour meter
periodicals and books (see references 1-5). The Wi positive sequence watts
utilization of these components in the solution of W2 negative sequence watts
zero sequence watts
practical problems, such as the calculation of short rWo resistance of the load current network
circuit currents in transmission lines (see references ζ impedance of the load current network
6 and 7). and the behavior of induction motors and X L reactance of the potential coil of the watthour meter
condensive reactance = XL
synchronous machines under unbalanced conditions Xc rhc component of 7& flowing through r
(see references 8-11) is described in other articles. Jbc component of Ibc flowing through z.
c

carl component of I flowing through r

ca

Written especially for E L E C T R I C A L E N G I N E E R I N G and containing the full sub-

z ca I component of I flowing through ζ
ca
stance of a paper presented informally at the Institute's Pacific Coast conveir- Isi current flowing through current coil of any positive sequence
tion, Vancouver, B. C , Aug. 30-Sept. 2, 1932. Not published in pamphlet form. meter

536 ELECTRICAL ENGINEERING

 Let Z = V 3 j° R . c 2An m

inspection of the network and

vector diagram shows the
f Ε ά
> validity of the following equa-
kfiJUL tions :

τ — (8)
ί al ry

6 a2
(9)
Fig. 2. Negative sequence voltage ICI = -ry—
Ac!
Fig. 3. Positive and negative sequence
voltage Il = Ia2 + Ici (10)

Γλ current flowing through current coil of any negative sequence The equivalent voltage indicated by the voltmeter
meter
Iai, h i , h i , l a , etc., currents flowing through impedances having is:
corresponding subscripts
7i current flowing through the potential coil of any positive sequence hRm (Π)
meter
Λ current flowing through the potential coil of any negative sequence
Consequently the meter indicates the negative se-
meter. quence voltage when the network is applied to an as-
sumed negative sequence system.
METERING OF Although it has been shown that the component
voltmeters indicate correctly when the networks are
POSITIVE AND NEGATIVE SEQUENCE VOLTAGES applied to the respective component systems, it
In using a standard voltmeter combined with a remains yet to be proved that the voltmeters will
potential network to measure the symmetrical com- indicate correctly when the networks are applied
ponents of voltage, the high internal resistance R m
to an unbalanced set of voltages. A proof of this
is removed, leaving only the negligibly small resist- is shown in the following equations:
ance of the moving element of the meter. Conse- Egl Eçl (12)
quently the network impedances which take the Il =
Zi Zi
place of R , together with the line voltages, entirely
a c

m
and
control the magnitude and phase position of the po-
+ p = ο (13)
tential current of the meter. Zal
The complex components of the network imped- Egl + Eg-, Ed + E Eg £ç (14)
ances may be any value provided the following gen- Il = c2 +

Zgl Zl Zg\ Z\
eral equations are satisfied :
C C

also
Zi = f/tZal (1) (15)
h =
Z2
c

Za2 C

Z a2 = ?/*Ζ Λ (2) and

Zal = Z c 2 = Kj»R m (3) Eal
+ ψ = ο (16)
Z2
If the voltmeter is to be direct reading, the numerical a Z C 2

value of the constant Κ must be equal to V 3 . h =

Eg2 ~h Egi E2
C -h E ] C
Ea_ + Ec
(17)
Z2 Z2
In order to more clearly explain the function of Τ· ' V a C

the network impedances, assume that a positive It is apparent, in comparing eq 12 with 14, and
sequence network can be applied to a positive se- 15 with 17, that the networks produce the same re-
quence set of voltages as shown in Fig. 1. sult, and hence the correct result, when applied to
Let Z i = \ / 3 j° R .
a An inspection of the net-
m
any unbalanced set of line voltages as when they
work and the vector diagram shows the validity of were applied to their respective component voltages.
the following equations: In Fig. 3 is shown the positive and negative se-
quence networks of Figs. 1 and 2 combined and ap-
Egl
1*1 = (4) plied to an unbalanced set of voltages. If the phase
Z\
a
sequence of the line voltages happen to be reversed
/- = ψ (5)
from that shown in Fig. 3 the network would func-
tion, but the voltmeter V\ would become voltmeter
Il = hl + Ici (6) V , and V would become V\. The larger reading
2 2

The equivalent voltage indicated by the voltmeter voltmeter is the positive sequence meter except
is: in very extreme cases of unbalance.
JlRm = j V . £ (7)
e l
POSITIVE AND NEGATIVE SEQUENCE CURRENTS
Consequently the meter indicates the positive se-
quence voltage when the network is applied to an In using a standard ammeter to measure the posi-
assumed positive sequence system. tive and negative sequence current of a 3-phase 3-
Again, assume that a negative sequence network wire system a current network is utilized, which con-
can be applied to a negative sequence set of voltages sists of a low resistance r, a low impedance s, and an
as shown in Fig. 2. impedance Z,„, equivalent to the impedance of the

AUGUST 1933 537

 Positive sequence current Negative sequence current
Fig. 6. Positive and negative sequence
current
current coil of the ammeter. The relations of the
impedances are such as to satisfy the equation :
ζ + Z = f/*(r + Z) (18) Therefore,
m m

j"/<Ibcl + lb*) , 4
This is the fundamental equation for all the current +
m i c a X ICA2)

networks described in this article. Vs Vs

The diagram in Fig. 4 shows a positive sequence (21)
network assumed to be applied to a positive se- Vs
quence set of 3-phase currents. It is apparent, from also
eq 18 and the diagram, that each of the line currents
divides into 2 equal components 60 deg apart. I&2 —
+ Vs
(22)
One component passes through the resistance r and Vs
and
leads the line current 30 deg; the other component
. /./
x :i1
3Ï
passes through the impedance ζ and lags the line J
6 e l

+ = 0 (23)
Vs
3
J-ca
current 30 deg. The vector diagram of Fig. 4 shows
Vs
the components of the line currents. Two com- Therefore,
ponents pass through the ammeter; J a com- caU
j1
X/6c2 4 IbcQ 3
(^ca2 4 I cal)

ponent of Icai, passing through the resistance r, and Vs +

Vs Vs
(24)
*I i, a component of I i, passing through the
bc bc

impedance z. Consequently the ammeter indicates It is apparent in comparing eq 19 with 21, and 22
the sum of these 2 currents which is equal to I . s] with 24, that the component networks produce the
It is seen from the diagram that the current I i is s same result when applied to the unbalanced line
equal and opposite to the current I \. Conse- ba currents as if they were applied to their respective
quently it is apparent that this ammeter indicates component systems.
directly the positive sequence current when the net- In Fig. 6 is shown the positive and negative se-
work is applied to a positive sequence system. quence networks of Figs. 4 and 5 combined to meter
Again, assuming a negative sequence system of simultaneously and separately the 2 components of
line currents and a negative sequence network as current.
shown in Fig. 5, it is seen that the 2 components
which pass through the ammeter are zlca2> & com- METERING OF SYMMETRICAL COMPONENTS OF POWER
ponent of I i passing through the impedance z, and
ca

RIBC2Y a component of I 2 passing through the

BC
In metering power components of a 3-phase sys-
resistance r. The ammeter indicates the sum of tem, standard wattmeters may be used, the current
these 2 components which is equal to I . It is seens2 coil of the wattmeter replacing the ammeter and
from the diagram that I is equal and opposite to
s2 the potential coil replacing the voltmeter. As is
I B2-
A Consequently it is apparent that this ammeter the case with the voltmeters, the internal resistance
indicates directly the negative sequence current R of the wattmeter potential coil is removed, leav-
m
when the network is applied to a negative sequence ing only the negligibly small resistance of the moving
system. element of the wattmeter.
It has now been shown that these component cur- Although it is evident from the previous analysis
rent networks, if it were possible to apply them to of potential and current networks that correct abso-
their respective systems, would cause the proper lute values of currents will flow through the coils
component currents to pass through the respective of the wattmeters, it is not so apparent that the phase
ammeters. It remains yet to be proved that these relations of these currents will be such that each watt-
same networks will meter the respective components meter will indicate correctly its component of power.
if applied to unbalanced systems. The proof of this In order to show these phase angles and current
is shown in the following equations, and the dia- values clearly, balanced component systems will
gram of Figs. 4, 5, and 6. again be assumed, and 2 methods of measuring the
power of each component system will be shown to-
(19)
Isl =
V3 Vs gether with vector diagrams.
and An assumed balanced positive sequence system
metered by 2 separate methods (a) and (b), is shown
= 0 (20)
Vs + Vs in Fig. 7. Method (a) is a well-known standard

538 ELECTRICAL ENGINEERING

 connection consisting of 2 potential transformers, a the networks of_Fig. 3 the voltmeter reading is too
standard wattmeter and either one or no current high by the V 3 factor, and Κ is therefore not the
transformer. For the sake of simplicity the power multiplying constant of the voltmeter. When these
factor is assumed to be 100 per cent. The power same networks are used however in the measure-
indication shown in method (a) is E X I = ment of power as in Fig. 9, and Κ = 1, the watt-
cal cal

V 3 Εχ X Ii positive sequence power. Method (b) meters read direct and Κ is the multiplying factor of
involves the component network system consisting the wattmeter assuming a 1 to 1 ratio of instrument
of a standard wattmeter with its potential internal transformers.
resistance R removed, a positive sequence current
m In order to determine definitely the location and
network equivalent to that shown in Fig. 4, a posi- number of instrument transformers to be used in con-
tive sequence voltage network equivalent to that junction with the various networks which are shown
shown in Fig. 1 with the exception that the value in this article, numerical values of an unbalanced
Κ = 1. A comparison of the 2 vector diagrams of 3-phase 3-wire system will be assumed, and the com-
Fig. 7 reveals that the 2 wattmeters will indicate ponent values calculated.
the same values and consequently the correct posi-
tive sequence watts. E 100 + j c a Iah 32 - j 13.5
E -25 + j 5Ce lea - 2 2 . 3 5 + j 15.165
An assumed balanced negative sequence system, E - 7 5 - j 50b I - 9.65 - j 1.665 bc

metered by 2 separate methods (a) and (b) is shown

Eei 78.8 - j 14.45 Iabi 20.99 - j 3.066
in Fig. 8. Method (a), as in Fig. 7, is a standard Ed - 2 6 . 8 8 + j 75.525 Icai - 7.845 4 j 19.69
connection. Method (b) involves a component £ i - 5 1 . 9 2 - j 61.075 6 hei - 1 3 . 1 4 5 - j 16.627
network system, consisting of a standard wattmeter E 21.15 +'j 14.45 7 62 11.002 - j 10.421
with its potential internal resistance R removed, E
a2 0

1.93 - j 25.525
m c2 I a2 -14.53 - j 4.34 C

a negative sequence current network equivalent to E - 2 3 . 0 9 + j 11.075 b2 hc2 3.53 + j 14.76

that of Fig. 5 with the exception that the current Total power 3,048 watts
transformers are located in different lines, a negative Positive sequence power 2,596 watts
sequence potential network equivalent to that shown Zero Negative sequence power 452 watts
sequence power 0 watts
in Fig. 2 with the exception that Κ = 1. A com-
parison of the 2 vector diagrams of Fig. 8 reveals Applying the network of Fig. 9 to the unbalanced
that the 2 wattmeters of this figure will indicate the set of voltages, complex values of currents may be
same, and therefore the total negative sequence calculated which pass through the current coils of the
power of the system. wattmeters, and equivalent voltages may be calcu-
It has already been shown that component net- lated which produce currents through the potential
works, both current and potential, when applied to coils of the wattmeters. From these values the indi-
an unbalanced system, give the same results as if they cation of the wattmeters may be computed. Table I
were applied to their respective component system. contains these values for different locations of the in-
Consequently the networks shown in Figs. 7 (b) and 8 strument transformers. For example, E bi is the C

(b) may be combined as shown in Fig. 9, with the equivalent voltage of wattmeter Wi when the poten-
assurance that each wattmeter will indicate its re- tial transformers are located across voltages E and c

spective component of power. The relative com- E ; and I (ab) (ca) is the current passing through the
b 2

plex values of the impedances for this network are current coil of wattmeter W when the current trans- 2

given in the following equations : formers are located in lines ab and ca. Only those
combinations may be used which give approximately
Z = Z = Kj*/*R (25)
el a2 m
2,596 watts in the positive sequence wattmeter and
Zl = Z = Kj R (26) 452 watts in the negative sequence wattmeter. It is
Q
a e2 m

It is important to observe that when Κ = 1 in

l
bc2

Iab2
leal
ΠΠΓΊ 7ΠΓ
'abl

GT UUL
TM.

Ica2

Z'bcl'
Ibci

Fig. 7. Positive sequence power Fig. 8 Negative sequence power

AUGUST 1933 539

important to notice that 4 current transformers are wherein the high resistance of the potential circuit
necessary when using the network of Fig. 9. may be separated from the negligibly small resistance
The load current network would be greatly simpli- of the moving element, will not be operative in con-
fied if only 2 current transformers could serve both junction with induction type watthour meters
positive and negative sequence networks as in Fig. 6, wherein the highly inductive impedance of the po-
and yet give the proper phase relations. This is tential coil exists as one inseparable unit.
accomplished by a network shown in Fig. 10. The In Fig. 11 is revealed a network designed to be used
negative sequence potential impedances are the same with standard induction type watthour meters. The
as those in the network of Fig. 9. The positive se- load current network is the same as in Fig. 10. The
quence potential impedances are changed; one is a potential networks have been altered to produce the
pure resistance, the other a condensive impedance. proper phase relations. X is a condensive reactance
c

equal to and placed in series with the inductive reac-

= Kj"/>R m
(27) tance X of the potential coil of the watthour meter.
L

Z = Kj°R (28)
The relations of impedances in this energy net work
c l m
are shown in eqs 29 and 30 :
The vector diagram of Fig. 10 and eqs 27 and 28 (29)
show the phase relations of the positive sequence sys-
Zd = Z Kj^X (30)
tem. Comparing this diagram with that of Fig. 8 a2 L

(b) it is seen that 2 series transformers located in

y In this network as in the network for metering
lines ca and ab will give the proper phase relations for power, the potential coil currents are entirely con-
the 2 systems. The simplified network of Fig, 10 trolled by the network impedances, since the conden-
therefore will meter simultaneously and separately sive reactance Xc balances out the effect of the in-
the 2 components of power. Table II gives the pos- ductive reactance X , and the effective resistance of
L

sible combinations of transformers for this network. the potential coil is assumed to be negligible. The
vector diagrams of Fig. 11 show that when the poten-
POSITIVE AND NEGATIVE SEQUENCE ENERGY tial transformers are located across voltages E and a

E , and the current transformers are located in lines

c

It is apparent that a network which will be opera- be and ab the watthour meters will register the true
tive with electrodynamometer type wattmeters, components of energy. Table III shows the possible

Fig. 9 . Positive and negative sequence Fig. 1 0 . Positive and negative se- Fig. 1 2 . Vector diagram of modified
power quence power energy network

90°

E W / " ™ / E C 2

Fig. 1 1 . Positive and negative sequence energy Fig. 13. Three-phase 4-wire network

540 ELECTRICAL ENGINEERING

combinations of instrument transformers to be used rents may flow. The potential networks, however,
with this network. will function if line voltages are utilized to feed the
In the design of the network of Fig. 11, it was as- network, since they contain no zero sequence com-
sumed that the effective resistance of the potential ponents.
coil was small enough to be neglected without appre- A diagram of a network designed to meter positive
ciable error in the registration. In some watthour and negative sequence components of power in the
meters this resistance may be of sufficient magnitude
to cause errors too great to tolerate. A slight modi-
fication of the positive sequence impedance values in
accordance with the vector diagram of Fig. 12 and eqs Table III—Location of Potential and Series Transformers to
Meter Simultaneously Positive and Negative Energy (Watt-
hours) When Using Network Shown in Fig. 11. Κ = 1

Table I—Location of Potential and Series Transformers to

Potential Transformers Series Transformers Watts
Meter Simultaneously, Positive and Negative Sequence
Power When Using Network Shown in Fig. 9. Κ = 1
2,599
1. Eacl h(bc)(ab)
Eact h(bc)(ab) 452.5
Potential Transformers Series Transformers Watts
3,051.5

2. Eabl Il(ca)(bc) 2,597

Eacl h(ca)(bc) 2,597 Eabl h(ca)(bc) 454
Eacl h(ca)(ab) 451.6
3,051
3,048.6
3. Ecbl h(ca)(ab) 2,591
2. E bia /i(eo)(ob) 2,591 451.6
Eabi h(bc)(ab) 452.5 E L2
C It(ca)(ab)
3.042.6
3,043.5

3. Ecbi Il(bc)(ab). .2,599

EcU h(ca)(bc) . . 454

3,043
presence of zero sequence currents is shown in Fig. 13.
The potential network is similar in all respects to
those used in the other circuits previously described.
Table II—Location of Potential and Series Transformers to The current network is also similar to the others with
Meter Simultaneously, Positive and Negative Sequence the exception that a double secondary current trans-
Power When Using Network Shown in Fig. 1 0 . Κ = 1 former is placed in the fourth conductor. The ratio
of transformation of this transformer is 3 times that
Potential Transformers Series Transformers Watts
of the other 2 current transformers. The second-
aries are connected in the circuit differentially with
1. Eacl /i(ea)(ob) 2,591
respect to the secondaries of the others. This con-
El ac h(ea)(ab) 451.6 nection causes only the positive and negative se-
3,042.6 quence components to flow into the network imped-
Eabl .h(bc)(ab) 2,599 ances, and the latter functions as in all the other
Eab2 • h(bc)(ab) 452.5 networks.
3,051.5 Attention is called to the fact that zero sequence
Eebl. • Ii(ca)(bc) 2,597 currents may be eliminated by another method of
E b2. C .h(ca)(bc) 454
connecting 2 differentially connected pairs of series
3,051 transformers together so that currents equal to the
difference of the line currents flow through the net-
work. This method, however, causes currents to
flow through the meters equal to V 3 times their true
31 and 32 will tend to decrease errors due to the pres- value and 30 deg displaced from the true values, and
ence of resistance of the potential coil. therefore would not be suitable for the networks
shown in this article.
The vector diagram of Fig. 13 shows an assumed
_ + - — + - (3D
7 l I a l 7 c l Έ
unbalanced current system containing positive, nega-
k { r + J X l )

tive, and zero sequence components. The diagram

la = J / / 2
3 a l (32) also shows 3 unbalanced currents containing only
positive and negative sequence components, the zero
METERING OF COMPONENTS OF POWER sequence components having been removed.
OR ENERGY IN 3-PHASE 4-WIRE SYSTEMS It is important to notice that the total power of a
3-phase 3-wire system is metered by the 2 wattmeters
Since the fundamental principle of operation of the Wi and Wi. In 3-phase 4-wire systems the total
networks previously described is based on the fact power is the sum of the positive, negative, and zero
that the 3 unbalanced vectors must add to zero and sequence components, the latter component not being
therefore contain no zero sequence components, it is metered. The zero sequence component may be
apparent that the current networks will not function metered by standard instrument transformers and
in 3-phase 4-wire systems wherein zero sequence cur- meters involving no network. It is only necessary to

AUGUST 1933 541

use 3 potential transformers and one current trans- systems having balanced voltages E may be equal to 0

former connected as shown in Fig. 14. zero, and although I is present, the total power of
0

Zero sequence power-3Eo/o cos φ (33)

the system is equal to the sum of W\ and W . 2

As in 3-wire systems, the wattmeter Wi will indi-

POSITIVE AND NEGATIVE SEQUENCE cate the total power of the 4-wire system at the ter-
REACTIVE VOLT-AMPERES minals of the alternator, if the no-load voltages are
balanced. This suggests the possibility of metering
Although the networks shown in Figs. 10 and 11 large blocks of power at the alternator terminals with
were designed primarily to meter, respectively, power one wattmeter, a positive sequence network, 2 poten-
and energy components, it is obvious that the reac- tial transformers, 2 current transformers replacing 3
tive volt-ampere components also may be metered. wattmeters, 3 current transformers and 3 potential
For example, if electrodynamometer type watt- transformers.
meters are used in Fig. 11 to replace the watthour When a line to line short circuit occurs on a 3-phase
meters, it is seen from the vector diagrams of this system, the negative sequence voltage and the nega-
figure that the equivalent voltages E \ and E will ac ac2
tive sequence power increases from approximately
be in phase, respectively, with Ii and I , and conse- 2
zero at the generator terminals to a maximum at the
quently in quadrature with I and I . These are the
si s2
point of short circuit. On the other hand, the posi-
proper phase relations for metering reactive volt- tive sequence voltage and positive sequence power
amperes in a system of 100 per cent power factor decrease from a maximum at the generator terminals
to a minimum at the point of short circuit. Positive
and negative sequence components of power might be
Fig. 14. Zero utilized in relays where definiteness of action de-
sequence power pends upon the amount of voltage available at the
relay. One relay utilizing the balanced components
of current and voltage will function regardless of
which phase is short circuited. This fact is a point
in favor of an investigation into the possibilities of
component current, voltage, and power relays.
It is apparent that a network must be designed for
one frequency and will be in error at any other fre-
quency. In large interconnected systems having
synchronous timing devices as part of their load, the
which the diagrams represent. The condensive react- frequency is not allowed to depart appreciably from
ances Xc will be omitted in this case. Again, if 60 cycles, even under short circuit conditions, and
watthour meters together with proper values of X c metering errors due to this cause would not be
are used to replace electrodynamometer type watt- serious.
meters, it is also apparent that the potential coil cur- A network similar to that shown in Fig. 10 was
rents Ii and I would be in phase, respectively, with
2 constructed in the laboratories of the University of
and I . These are the proper phase relations for
s2 Washington. In tests conducted with correct fre-
metering reactive volt-ampere-hours with watthour quency errors no greater than 2 / per cent were re-
1
2

meters. corded. In tests for accuracy with change of fre-

Consequently the networks shown in Figs. 10 and quency; 59 to 61 cycles showed errors no greater than
11 may be designed to meter either the real or reac- 0.5 per cent; and 55 to 65 cycles showed errors no
tive components of power and energy. greater than 4.5 per cent. The various tests con-
ducted as a whole sppaks well for the ultimate pos-
CONCLUSION sible accuracy of such networks when designed and
constructed by meter manufacturers, experienced
Since the zero sequence current does not exist in a with this type of equipment.
3-wire 3-phase system, there is no zero sequence
power, and the wattmeters W\ and W indicate the REFERENCES 2

total watts of the unbalanced system. In 3-wire sys-

tems having unbalanced load currents but balanced 1. E L E C T R I C C I R C U I T S — T H E O R Y A N D A P P L I C A T I O N S (book), O . G . C . Dahl,
line voltages, the negative sequence power is zero, p. 77.
and the wattmeter W\ will indicate the total power of 2. E L E C T R I C A L E N G I N E E R I N G , V . I I (book), V . Karapetoff, 3rd ed., p. 118.
the system. This case will occur when the power is 3.4. AMLETTEHRON DA T O I N G C U R R E N T S (book), C . E. Magnusson, 4th ed., p. 602.
FSYMMETRICAL COORDINATES APPLIED TO T H E SOLUTION O F
metered directly at the terminals of the alternator P O L Y P H A S E N E T W O R K S , C . L . Fortescue. A.I.E.E. T R A N S . , V . 37, 1918, p.
the no-load line voltages of which are balanced. The 5.1027-1115.
MEASUREMENT OF POWER I N P O L Y P H A S E C I R C U I T S , C . L. Fortescue.
positive sequence wattmeter will also indicate a small A.I.E.E. T R A N S . , V. 42, 1923, p. 358-71.
negative sequence copper loss. The negative se- March 1928, p. 151-7; Apr. 1928,C .p.F.194-7;
6. S Y M M E T R I C A L C O M P O N E N T S , Wagner and R. D. Evans. Elec. JL,
June 1928, p. 307-11; July 1928,
quence meter will indicate only negative sequence p. 359-63; Sept. 1929, p. 425-31; Dec. 1929, p. 571-81.
copper loss of the alternator. 7. CALCULATION O FSINGLE PHASE SHORT CIRCUITS B Y M E T H O D OF SYM-
M E T R I C A L C O M P O N E N T S , A. P. Mackerras. Genl. Elec. Rev., Apr. 1926, p. 218-
Since zero sequence power may exist in a 3-phase 31; July 1926, p. 468-81.
4-wire system, the 2 wattmeters Wi and W may not 8. E F F E C T O F U N B A L A N C E D V O L T A G E S O N O P E R A T I O N O F I N D U C T I O N M O T O R S ,
2

meter the total power of the unbalanced system. 9.O. CI.NSchoenfeld. Elec. JL, Jan. 1925, p. 30-3.
D U C T I O N M O T O R S O N U N B A L A N C E D C I R C U I T S , A. M . Dudley. Elec. JL,
This is the case when both E and I are present. In July 1924, p. 339-43.
0 0

542 ELECTRICAL ENGINEERING

10. J. Slepian. Elec. World,
INDUCTION MOTORS
Feb. 1920, p. 313-5.
ON UNBALANCED VOLTAGES,
esses and resulting physiological and psychological
11. ELECTRIC CIRCUITS, T H E O R Y A N D A P P L I C A T I O N S (book), O . G . C . Dahl, effects, is not included in this figure. It has been
p. 138.
said that man can see better than any other animal—
12. A N E W S E Q U E N C E S Y S T E M Evans. Elec.
World, Feb. 1923, p. 333-4.
OF POLYPHASE M E T E R S , R . D .
because of his intelligence. The problem of proper
13. M E A S U R E M E N T O F SYMMETRICAL COMPONENTS, A. P. Hayward. Elec. lighting for seeing is affected by such incidental
JL, June 1931, p. 351-7.
conditions—the subject's vision, perception, inter-
pretation, and reactions—but these are elements
beyond the direct control of the illuminating engi-
neer. It is indeed true that oculists may heal sick
eyes and fit crutches to crippled eyes, and educators
and experience can train the mind and body to react,
but all the oculists and educators in the country

Lighting for cannot make even a normal man see without light.
Providing the proper lighting is the job of the engi-
neer, and Fig. 1 indicates some of the characteristics

Effective Seeing
which must be considered.
Among these characteristics of lighting we find
certain positive factors such as level of illumination
or foot-candles, and distribution and diffusion of the
light; certain negative factors such as glare, direct
A non-technical explanation of the funda- and reflected ; and other characteristics, which may
mental conceptions and relations under- be helpful or otherwise, such as direction of light,
source brightness, type of object to be seen, and
lying the application of lighting for effec- background about that object.
tive seeing is given in this article. The The seriousness of one of the least appreciated of
tables and diagrams included should be these characteristics, glare, is illustrated by the
useful in any study of illumination and in- following example. An operator is working before
a machine illuminated by a bare lamp bulb and sus-
tensity. pended on a drop cord above the machine. As-
suming that seeing conditions other than glare are
satisfactory, the seeing ability of the operator
By will be reduced to 5 8 per cent if this lamp bulb,
L. V . J A M E S General Elec. Co., the source of glare, is at an angle 4 0 deg above the
Chiosso, III. line of vision; this simply means that 4 2 per cent
of the light originally furnished is neutralized or
rendered ineffective and wasted. As the glare source
is lowered further toward the line of vision, its im-
portance as a waster of energy and an interference
I HE INFLUENCE of good lighting
with seeing greatly increases. At any angle of 2 0
on seeing has been demonstrated by various labora-
deg above the line of vision, 5 3 per cent of the
tory studies ; the successful application of the results
light is wasted, at 10 deg 69 per cent is wasted, and
brought forth in these studies, however, depends
at 5 deg 8 4 per cent of the light furnished is wasted.
upon an understanding of the fundamental principles
of lighting and seeing. So far, the maximum eco-
nomic and human benefits from the manufacturing WHAT CONSTITUTES BETTER SEEING?
and technical developments in artificial lighting have
not been secured, due to the fact that the funda- Let us examine more in detail the results in better
mental principles are not sufficiently well known. seeing through the elimination of glare. Without
Certain of these principles are outlined in the
following paragraphs.
An individual retains a mental picture of each
visual observation. Whether that picture is clear
and distinct, or is hazy and easily forgotten or
easily replaced, depends largely upon whether the
light was proper, the lens (or eye) was accurate,
and the film (retina and receptiveness) fresh and
sensitive, with proper time allowance for good im-
pression. Of these 3 elements in seeing, 2 are
suggested in Fig. 1; namely, first, the external
physical realm including light and lighting, and
second, the eye or visual sense. The third element,
the internal realm consisting of the mental proc-
Essentially full text of an address presented before a conference on economics
of applied lighting, Case School of Applied Science, Cleveland, Ohio, Feb. 8-9,
1933, and based upon a pamphlet "Lighting for Seeing," by M. Luckiesh
Fig. 1. Lighting and vision are both necessary
(ΜΊ5) and F. K. Moss. Not published in pamphlet form. for seeing

AUGUST 1933 543