The setup for maktng time-domain reflectometer

measurements. The TDR circuit i s cmntained in a

Minobox, center. Its power supply is visible a t the
left. The a h l e being tested, coiled up in the
background, is termmated with a potentiometer
mounted for testing. foreground.

An l nexpensive
Time-Domain Reflectometer
BY WARREN JOCIIEM.* WBZlPF

line with a
C VOhl and SWR metertransmission
HSCKING YO~JR
is but locating the
sasy.
scope starts sweeping to the right t o display the
top of n single pulse, Fig. 1H. If the voltage step
problem may not he. There is a simple technique encclunters any discontinuities (shorts, npms, or
called time-domain reflectometry (TDR) that mismatches) along the line, a ret'lectcd step will h e
works like a closed-circuit radar system t o physical- x t op a t that point and will travel back u p the line.
ly locate a faolt. More than that, it can find If the line is not too long, the reflected step (or
multiple fnults along a single line, measure SWR, steps) will rehlrn soon enouph to add algebraically
and Even tell you the characteristic impedance of to the forward voltage pulse. Fig. IC. 'This pro-
your cable. ducts a series of bumps on the top of thr displayed
Industry has used 'I'DR for sume time, hut pulse. 'lhe position of these bumps on the scope
vummercldl 'rDK systems are very expensive. I:or h c e is dircctly related t o the position of the faults
amateur applicalinns a hirnple I D R can be assem- along thc Line. A nice sdvantagc of this over the old
bled for only a few dollars, provided a wlde-band VOhl is that all faults on the line will be shown.
~ i ~ i l l o s c o p( 1e 0 MHz or better) is available, not just the fiist one. Src Fig. 2.
Fig. .1 shews a hlack diagratn fcrr a basic 1 D R . If a fdult is too fnr down the line, the rrflectcd
It consists of a fast risa-time pulse generator, a step may not return to the scope in time to add t o
wideband oscilloxope, und a '1 oonnector. l l e the positive pulse, and this information mpillbe lost.
%opt rhould be connected t o t h e 1 t t v o u ~ hs Therelore, the widex the p u k , t h e longer a cable
short length cnf RG-62/U. '93-ohm coax. h li1:l you can test. Ilsing a I-MHz square-wave souice
probe &:an slsn he oscd hut s short piece of cable (500 nanosecond duration of the positive portion
will make the system more sensitive. of the wave), cablrs up to abuut I 5 0 fest in lcngth
In operation a vulfagr stcp ( t h e lending edge of may he tested.
the pulrel from the senerator ir w.nt down the The pulse generator used here is r simple
transmission line under test. i\t &e rame time crystalsontrolled astable multivibrator. It was con-
* 6 1 Montague Ave.. Trenton. N J 08628. structed from a 7400 sriies 'LTL quad NANDINOR

Fig. 1 - At A. the basic setup for time.domain reflectometry measure-

ments. At B, the initial pulse as displayed on the o$cilloscope. and a: C, the
effect of a reflected n31tage o n the scope display (where the load
impedance is smaller than the characteristic impedance of the test linel.

(A) (8)

OUi001M REFLECTLO
PlJLSE (Vil PULSE I V l l

1 I-- L . L rcl

March 1973
 Fig. 2 - The "bumps" on t h e pulse as dlrplaved on
the orcillosmpe are related t o the porltlon of the
faults along the test lane.

with a high-frequency tlip-tlup 17490) %:I1 signifi-

vantly improve the rewlution.

I n operation the test line is connected as shown

in E.ig. I . Here 1s an important lim~tattonuf this
\imple I'DR. The test l ~ n emufit be termtmted in s
pure icristance. For example. it must he iliscon-
neuted from an ;antenna, or a tuner, ur a loading
gate. I l c r:ircult (shown in Fig. 31 i.; rtralght- <mil. The display i9 simply too complex whcn
fnrward and inexpcnsivc.l We have burlt several o n reactances arc pre=nt. Once the $"stem is set u p
pertorated hoard and others no [rc board u ~ t hno dnd the t e t line resistively terminated, the rcvpe
payout problems. R1 should equal the characteris rweep should he t t o display the top of a single
tic impedance at' the line heing tested 1i.e.. Yor pulst. across the entzre rcreen tree Fig. 4). In thir
KC-59iII. K1 75 ~ t h m s )'Ibis
; . prevents mult~ple test a length uf KC-59iU was terminated in 150
s the line, rwhich can distort thc scope uhms.
~ ~ t l c c t l o non tind the length o f this line, l, we use the
d~splay. equation
I' T
Multiple retlcctiuns rcault r\ihcn both the gener- ,.e wq. 11
ator and the load cr~dr,t a line :(re ~niismatched.A
voltage step partially ieilected by the ioaci wulild where VP is the vclacity ut' propagation on the line
b? partldly reretlected nt the generator and will and T Is the lime it touk the pulse to go doum tu
~:ontinuetraveiine back m d forth on the line until the lrlad and hack. t> equaL~thc velocity factor
IVF) of the 1:aItle in use timcs the speed of light.
VF wries fn,m cdble tu ,.able Jepcnding primarily
upon the dielectric constant of thc rahle. '[he VF
with mare than one h u l t on a lioa, so it i- best i;?r ~:.~mmnna b l e r can be forrnd in most elcc.
when terting a line t o i i x the fault closer1 to the tronics handbo(~kr.zFor KG59/U, VF =- i1.60.
gcnciator and then work nn down the line. r; - ( i l 6 6 119.84 X 1tl8 ft/r)
I'hc resolution o i this system is limited chiefly
hy the rifr time af the pulsc. Even a,,pronounced , ;
L = 6.5 x 1118ft/s
taulis as close t u p t h e r as nnr foot can bc rasulved. T is rend directly from the *:ape face. R e d i n g
llsine the generator shown and a It)-Mliz rcope, from the midpoint o f hnth leading edges ll'ie. 3). I'
the s w p r is the limiting factor. If a better scope is * 4.6 divisions on the X axis. with the scope set t o
ava~lable,using 7411lJtl gates or shaping the pulse 0 5 W/div.
I For this and subsequent r e f ~ r e n c ~scrs , tisting
at the m d of this article.

Fig. 3 - Schematic diagram of the 1'DR pulse

rt generator. Resistances are in ohms: all resirtors are
%?o 112 W 10-percent mlerance. c o m p o r ~ t ~ o ntJnlers
.
r\ " " " " ~roted.all capacbtance values are picofarads IpFI.
TO
y;OpT

C1 - Elestrolytic.
2109 J 1 , J2 - C o a ~ i a lConnectors. type BNC. LlNE
R 1 -See text. UNDER
111 - allad NANDINOR gate, type 7400 (Slgneticr
N7400A or equivalenrl. one section unused.

QST for
Fig. 4 -TDR display wlth a 7Cfoot length of Fig. 5 - Multiple-exposure photograph showing
RG-59lU (75-ohm coax1 terminated in a 150-ohm the TDR display with three different values of
resistance. terminating resistance at the end of the test line.
The trace which appears as almost a stra!ght.line
extension of the initial pulse amplitude results
from a matched condition at the far end af the
l'hcrcfure:
- 1238x&-3
1 6 . ~ f t~i r ~~
line.

2
1 = 7s ft
We divide by two since we want the c;iBlr lcngll, -
AT-?..-
A
+ ~ L

....f.-, A.;
i-
.

not the distance the prilsr h:!s traveled, \which is

down and back. 'This 75 fect is in exact agirclncnr
with the m~;gsured length uf the <,able. I t j u r t
happencd that tlle fault was the cnd o f the cable. If
ZL.. zo ZL :: zu

i t h;d heen in the ruiddlc a snlall bump i n the pulse Flq. 6 - TDR d~splays tor SWR measurement.
would appcar, as st~owni n Fig. 2. where Ll, is purely resisttvr.
It' you do riut know the velocity o f prupxyation
rlrnply warrange equation 1.
llsinr nnutilur ivrrnnla Vrom rrshich the SWR
l.i, : .$
, m y he fbund if 21, is known .ind is purely
resistive,
I= known cahle length
-
'I' time down and hack. read directly SWR =i!:
, L!!
: !.i(l .-
:
7.i
Now for thoas skeotirs who d o riot bclievr in
~'haracleristic irnpcdancc. connect a 2lIlI-ohm po- h r yoll ?;in *ee, thcsc twu vllucs are i n very &:lose
tentiometer to thc i:nd rlt' the line. ,\djustine, i t ;agreetncnf.
through its range st~<,uldprovide s displsy like that 'There are m m y ~lrt~er:ipplications uf this
uf Fig. 5 . \Vhm the display i s coropletely flat the onusual mea$urine trchnique hut I'll lcavr them
pot ran be rrrnoved frum the circuit and its crpcn 18, ;mateor insenuity. 1 ivuuld like to thank
resistance rdue measured ou ;an c~hmmetcr.The Ijr. :\lien Ki$tz, K?UYH. Heatl ot' the b,lectronic
pot resivtance shc,uid be wfthin *one nltm of the 'le~.hnolocy prograrr, at 'l'rcnton State ('ollc~e, for
~:harscteristicittlpr,toncc o f the line. his llclp in this project and Paul I.arnhut.
I f y<>u ever wanted to *ctu:~lly risui~lize th? WH2HIiH. IC the ,photography.
I
VSlVR <rn .r line. *el op a t r r t l o pmduce thr
displays shown in 1.i~. 6 . and !n;ikr rslculations
with rhip equation.
I ) Potlock. "TTL Uryl;tnl Ilsrillator." l'vrh-
llicd Correspondence. QST. Oetoher. 1971. p. 54,
Also see Yredhack, QS7: December. 1971. p. 47.
and Ertes, "More 3.m the TTl. Oreillator." Tech-
n i c a l Correspondence. QST. Febvuary, 1972. p. 55.
where I.,; = incident voltage :and I.,'" =: reflected 2 ) Antennas and Feed Svdemr chapter, llle
vultagc. Rodio .AmarrfcrS FIIF &f6nuul, csr Transmission
llring Pig. 4 ns an elample I , i m c:haa,trr, r ~ t h e r'rir l K K L ;Inrenna Book. o r
fi., - 3.8
. 'Thr liadio d r,?ufetrr tfunrlhr8oK.
~ ~ ~ ~
a) Allen. '.A hactirai Esprtim~nter's Ap-
. ~ -
proach to Time-l)oma~n Krflertometw;' ifam
Radio. May. 1971, 0. 22.
4) I.mk. "Time Domain Reflertometrr,"Flec-
cronies IVorld,Septemher, 1966, 0. 48.

March 1973 21