WAVECREST Corporation

MEASURING TIME DOMAIN CHARACTERISTICS

OF TRANSMISSION LINES
Application Note No. 101-A

WAVECREST Corporation continually engages in research related to

product improvement. New material, production methods, and design
refinements are introduced into existing products without notice as a
routine expression of that philosophy. For this reason, any current
WAVECREST product may differ in some respect from its published
description but will always equal or exceed the original design
specifications unless otherwise stated.

Copyright 1993

WAVECREST Corporation
A Technologies Company
7275 Bush Lake Road
Edina, Minnesota 55439
(612) 831-0030
(800) 733-7128
www.wavecrestcorp.com
All Rights Reserved

Contents
Introduction ...................................................................................................................... 5
Why Make Transmission Line Cable Measurements? ................................................. 5
Cable Characteristics ....................................................................................................... 6
Velocity of Propagation.......................................................................................... 6
Velocity .................................................................................................................. 7
Impedance, Capacitance, Inductance...................................................................... 8
Attenuation ............................................................................................................. 9
Additional Cable Information................................................................................. 10
Pulse Response ........................................................................................... 10
Shielding and Cross-talk ............................................................................ 11
Self-Generated Cable Noise ....................................................................... 12
Flexibility ................................................................................................... 12
Mechanical Strength................................................................................... 12
Cable Connectors ............................................................................................................. 12
SMA ....................................................................................................................... 13
APC-3.5.................................................................................................................. 13
APC-7..................................................................................................................... 14
TNC ........................................................................................................................ 14
BNC........................................................................................................................ 14
N ............................................................................................................................. 14
SMB........................................................................................................................ 14
SMC........................................................................................................................ 14
Effects of Discontinuities ................................................................................................. 14
Conclusions ....................................................................................................................... 18
Current Measurement Practices ..................................................................................... 18
(Mechanical, Time Domain Reflectometry)
Digital Time System (DTS) .............................................................................................. 20
DTS Cable Measurement Techniques ............................................................................ 20
Bench Setup............................................................................................................ 20
Test Jig for Balanced Cables (or other impedance)................................................ 21
Golden Cable.......................................................................................................... 22
External Calibration Option on DTS...................................................................... 22
Trimming Cables .............................................................................................................. 23
Trimming Balanced or Twisted Wire Cable .......................................................... 23
Trimming Coaxial Cable ........................................................................................ 24
References ......................................................................................................................... 25

Measuring Time Domain Characteristics

of Transmission Lines
Introduction
This application note provides an in-depth study of the time-domain
characteristics of electrical transmission lines and the techniques used to
measure these characteristics.
A transmission line (or cable) transfers electrical energy from one point to
another. In many instances, the time it takes to transfer this energy (time delay)
is important and must be determined. The physical properties of the cable and
its operating environment influence this time delay.
This note helps the reader understand transmission lines, their environmental
properties, how to control them and how to obtain the most accurate
measurements of the time delay. The WAVECREST Timing Measurement
Instrument is introduced and compared to time domain reflectometers.
Hardware implementation of these techniques is also discussed.

Why Make Transmission Line Measurements?

High-speed computer systems use a master clock to synchronize various
individual computing circuits. This clock must be distributed throughout the
computer system, arriving at thousands of individual destinations at the same
time. Often, this distribution system employs cables, and the accuracy of these
cables affects the speed of the computer.
Typically, a transmission line distribution system is used to feed discrete
antenna elements such as those in phased-array radar systems that are
individually fed with microwave energy. The performance of the antenna, and
consequently that of the radar, is dependent on how well the phase relationship
of the antenna elements can be maintained.
Semiconductor automatic test equipment has hundreds of electronic channels
to exercise the pins of the device-under-test. These channels are distributed
throughout the test system before being connected to the device, and the
overall accuracy of the test is a function of the accuracy of this distribution
system. A more accurate test allows semiconductors to be manufactured more
efficiently.
Aside from these examples, there are hundreds of applications where accurate
measurement of transmission line time delay is critical to system performance.
These applications include nuclear research, fiber optic or laser systems,
communications networks, and calibration laboratories. Often these
measurements must be performed with an accuracy of greater than 100
picoseconds.

Cable Characteristics
Real-life transmission line cables have a number of properties that affect the
time delay through the cable. These effects can vary from cable to cable and
are dependent on environmental characteristics such as temperature, humidity,
and proximity to other objects.
The values of these properties are determined by measurements. Usually these
measurements are made on a large number of cables, of the same cable type,
from various production lots. These values are usually specified as nominal
values meaning the actual value may be substantially different from the
catalog specification. An understanding of these properties and how they
affect time delay can help make cable delay measurement easier and more
accurate.

Velocity of Propagation

The velocity of propagation, , of a transmission medium is a fundamental

physical property which indicates the speed at which a signal can travel
through a unit length of the medium. The simplest example is a traveling wave
through free space (a vacuum), where the velocity of propagation is equal to
the speed of light, :
= 3 X 10

10

cms/sec

In a medium other than the perfect vacuum, the velocity of propagation is

slowed by an amount related to the relative dielectric constant of the medium,
, according to the following relationship:
= /
The relative dielectric constants of several transmission media are given
below:
Media
Air
Solid Polyethylene
Air Polyethylene
Teflon (PTFE)
Foam Teflon

Velocity of Propagation,
1.00068
1.5174
1.1363
1.4409
1.1111

The time delay through a transmission line, td can be determined from the
velocity of propagation if the length of the line, L, is known.
td = l/
The velocity of propagation is the most important property in determining the
time domain characteristics of transmission lines. In most cases, however, the
actual time delay through the cable is longer than predicted by the above
equation because of many effects on the cable by its environment.

Zo
C
L
al(f)

velocity of propagation
impedance
capacitance /unit length
inductance/unit length
attenuation frequency, f

The above values for a specific cable type are derived from measurements
made on many different runs of that cable. The values can be and usually are
different from those of a single cable sample. They also vary from
manufacturer to manufacturer. The following information simply illustrates
that manufacturers technical specifications are always nominal values, and a
single cable sample may vary. Therefore, for time domain applications,
awareness of these variations must constantly be considered, especially when
applications require specific lengths and specific time delays.
The accuracy that can be achieved in the high performance time measurement
of cables is in direct proportion to the amount of effort expended on good
engineering practices. The following paragraphs describe the above
characteristics in detail.

Velocity
The delay per unit length of a single cable sample is related to the velocity of
propagation, . The velocity of propagation is the transmission velocity of an
electrical signal through a cable, and is expressed as a percentage of the
velocity of light. is the velocity of propagation and E is the dielectric
constant.
= 1
E
Velocity of propagation is also related to the inductance per unit length, L, and
the capacitance per unit length, C, and is expressed in unit length delay per
second.
= 1 unit length/second
LC
The specific time delay, T, of a single cable sample is dependent on the
dielectric constant, E, and is expressed in ns/foot.
T = 1.016 E ns/foot
The following table provides velocity factors and computed time delays for
typical cable dielectric. It illustrates that actual cable lengths differ greatly
between cable types.
Velocity
Time Delay
(feet)
Cable Dielectric
(nsec)
Solid Polyethylene
65.9
1.54
Foam Polyethylene
80.0
1.27
Air Space Polyethylene 84.0
1.21
88.0
1.15
Solid Teflon69.4
1.46
Expanded Teflon
85.0
1.27
Air Space Teflon
85.0
1.20
90.0
1.13

There are advantages and disadvantages associated with each type of

dielectric. Solid polyethylene is easy to process, low in cost, has high
dielectric strength, and a relatively low dielectric constant. Foamed
polyethylene contains extruded materials that have been expanded by
numerous individual air cells. These materials reduce the dielectric constant
significantly and provide greater design flexibility. Irradiating polyethylene
increases the thermal stability and resistance to soldering iron heat. Teflon,
while more expensive, has high dielectric strength, a low dielectric constant,
withstands temperature extremes, and withstands exposure to gases and
liquids that would destroy other materials.

Impedance, Capacitance, Inductance

Expressed in ohms, impedance is important for matching signals and
maintaining the most efficient transfer environment. In coax cable:
Zo = 101600 = 138 log D
vC E
d
Zo
v
E
D
d

ohms

is the nominal impedance

is the velocity factor
is the dielectric constant
is the dielectric outside diameter
is the inner conductor diameter

In all cables, the nominal or high frequency characteristic impedance, Zo, is

dependent upon L and C.
Zo = L ohms
C
However, it is important to understand that the actual impedance of a single
cable sample varies with frequency, and can be different from the characteristic
impedance because of reflections in the cable.
The total of all random and periodic reflections, as well as connector and line
termination reflections, is the voltage standing wave ratio (VSWR). The
VSWR indicates the difference between real input impedance and the average
characteristic impedance. The following diagram illustrates some typical
cable/ connector combinations, and their resulting VSWR over frequency.

If C is specified, a knowledge of v or Zo determines L. By rearranging the

formulas, v and Zo can be specified so that L and C can be determined.
L = Zo
v

C=1
Zo

Capacitance is the ratio of electrostatic charge on a conductor to the potential

difference between the conductors required to maintain that charge. It is
expressed in picoFarads per foot.
C = 7.36E pF/ft
log D
d
The above formulas show that capacitance, C, figures significantly in
determining Zo. Some dielectric materials cause changes in capacitance when
the cable is stressed or bent, if the dielectric moves, or even when the
temperature varies. Careful system design and good engineering practices are
mandatory for maintaining consistent capacitance.
This table shows capacitance values for some typical cable types. Capacitance
is normally specified in picofarads per foot.
Nominal
Capacitance
(pF/foot)
Cable Types
30.8
25.4
29.4
20.6
16.9
19.5
16.3
13.5
15.4
10.0
6.5
12.0
22.0
20.0

50 ohm Solid Polyethylene Coax

50 ohm Foam Polyethylene Coax
50 ohm Solid PTFE Coax
75 ohm Solid Polyethylene Coax
75 ohm Foam Polyethylene Coax
75 ohm Solid PTFE Coax
95 ohm Solid Polyethylene Coax
95 ohm Air space Polyethylene Coax
95 ohm Solid PTFE Coax
125 ohm Air space Polyethylene Coax
185 ohm Air space Polyethylene Coax
Low Capacitance Twisted Pair ELA Data Cable
Foil Shielded Single Twisted Pair (UL#2092)
Foil Shielded Two Twisted Pair (UL#2094)

Several other items also affect cable capacitance and impedance. The
capacitance and impedance of long lengths of cable generally vary less than
2% over their operating temperature range. However, dielectric movement at
the connector interface can cause the VSWR to vary significantly; and adding
an integrated circuit input, or other passive components on any cable path can
change the capacitive load.

Attenuation
Attenuation is the power loss in a cable. It is due to heating loss because of
conductor resistance, skin effect, and dielectric loss caused by poor dielectric
materials. The total loss is expressed in decibels per unit length of cable. The
decibel is a unit that expresses the ratio between two amounts of power
existing at two points.
dB = 10 log PI
P2

The attenuation of a single cable sample can vary as the frequency, rise times,
or pulse widths change. Random and periodic impedance variations give rise
to varying attenuation responses. In addition, attenuation usually increases
with time. This is usually caused by corrosion, contamination of the primary
insulation, by moisture penetration, and/or dielectric deterioration.
Silver plated copper is much more effective long term, than bare copper and
tinned copper. Foam polyethylene dielectric has approximately 15% less
attenuation than solid polyethylene cables. However, when using foam
polyethylene cable, impedance increases if moisture is absorbed. Flexing also
impacts attenuation (see the following section on flexibility and mechanical
strength). The following diagram illustrates the different attenuation factors
that impact a cable as frequency changes.

Additional Cable Information

A variety of other items affects cable measurements. These include pulse
response, shielding and cross-talk, self generated cable noise, flexibility and
mechanical strength.

Pulse Response
Five characteristics affect pulse response when making time domain
measurements on cable. They include impedance and reflection, rise time,
amplitude, overshoot or pre-shoot, and pulse echoes.
Impedance and reflection: the impedance along a length of the cable varies,
sometimes up to +/- 5%. One cannot assume that the correct impedance is
selected throughout the entire system.
Rise time and amplitude: rise times are affected as they pass through a cable.
The output rise time is a function of input rise time, pulse width, and cable
attenuation. The faster the rise time, the greater the energy imparted to a
signal, and the more reliably the signal passes through the cable for
measurement. The following diagram illustrates rise time variations versus
cable lengths for 10, 20, 50, and 100 nanosecond pulses.

10

Amplitude is attenuated as cable length increases. The following diagram

illustrates peak amplitude attenuation versus cable length.

Raising the cable temperature results in increased rise time and decreased
amplitudes. Therefore, short cables are beneficial to most electronic
applications because they not as susceptible to degradation.
Overshoot or pre-shoot: overshoot disrupts signal quality, and is typified by
ringing. It is caused by periodic reflections due to impedance discontinuities
within the cable. Pre-shoot is seen in some balanced delay lines and is
minimized by good cable design and termination.
Pulse echoes: pulse echoes can occur when a narrow pulse is sent through a
cable. In addition, periodic reflections can cause a small pulse of energy to be
created after the initial pulse has arrived. Normally this echo level can be
neglected.

Shielding and Cross-talk

The shielding efficiency of cable depends on the construction of its outer
conductor. These shielded cable include single braid, double braid, triaxial,
strip braids, and solid sheath. The relative shielding efficiencies of the
different cables are rated by the amount of signal that leaks through the outer
conductor, and can be detected outside the shield.
Efficiency depends on the shield type. In cable using a single copper braid, the
leakage typically -30 to -50 dB down from the signal level in the cable. In
solid sheath cable, the leakage is typically -300 to -1000db down. Double
braid, strip braid and triax have -70 to -100 dB down. All of these specs are
over the 10 to 1000 MHz range.
The signal leakage and other factors allow signals to transfer in and out of a
cable, and indicate cross talk susceptibility. Cross talk factors include isolation
and leakage with other cables, relative spacing and positioning of cable runs,
distance from certain objects, and grounding practices.

11

Self-Generated Cable Noise

When a cable is flexed, it generates acoustical and electrical noise. Acoustical
noise is a function of the mechanical motion within the cable, and is minimized
by good cable design. Electrical noise is a function of an electrostatic effect; it
is minimized by preventing motion between dielectric and conductors and by
dissipating electrostatic charges.

Flexibility
Cables are classified as flexible or semi-flexible. Flexible cables are intended
for applications where the cables flex repeatedly while in service. These
cables are made with a stranded center conductor and braided outer conductors.
They typically withstand over 1000 flexes through 180 degrees with a bend
radius equal to 20 times the outside diameter of the cable. The minimum
recommended bend radius is five times the cable outside diameter.
Semi-flexible cables are intended for applications where the cable remains
flexed while in service. These cables are made with a tubular outer conductor.
They typically withstand only 10 flexes through 180 degrees with a bend
radius equal to 20 times the outside diameter of the cable. The minimum
recommended bend radius is ten times the cable outside diameter.

Mechanical Strength
The break strength of coax cable is dependent on the strength of the outer
conductor. Coax cables typically have a break strength of 70% of their outer
conductor. Other cables depend on the strength of their inner conductors. In all
cases, cable strength is enhanced if the center conductor stretches up to 10%
before actually breaking. The following table illustrates typical conductor
materials and their properties.
Tensile
Conductor
Strength
Material
Conductivity (Psi)
Elongation
Annealed copper
Copper covered steel
bard drawn
soft drawn
ITT Alloy 63

100%

35,000

20%

40%
40%
90%

120,000
60,000
60,000

2%
12%
12%

Care must be exercised using conductor sizes of less than 26 AWG, since
breakage easily occurs, especially during assembly.

Cable Connectors
Connectors and the cable/connector attachments are ultimately the greatest
limitation on performance of a system. The same rules that apply to cables also
apply to connectors. Resistance, inductance, capacitance, and impedance all
impact the quality of signal integrity. Oftentimes skew from cable assembly is
derived from problems associated with connectors rather than the cable itself
(for example, coaxial cable typically holds specifications within 2% to 5%).
Also, like cable, silver and gold plated parts maintain impedance better, and
provide a mechanical interface that lasts longer.

12

Manufacturers assembly instructions are important to follow. Manufacturers

tool sets for specific connectors can also be used when they are available.
Soldering, rather than clamping or crimping, is the preferred method of
attachment. It ensures a constant impedance connection and a low VSWR
when done correctly. For best results, attention must also be given to nut
tightening torque.
All of these techniques are important in obtaining specific lengths or delays in
cable measurement. The standard mechanism for measuring cable length is to
measure from connector face to connector face. The face is usually the plane
where the two connectors mate. This face varies among different types of
connectors. Check the specifications for specific connectors.
Standard cable measurements are usually made with one male and one female
connector installed. As opposed to a male/male or female/female cable, this
preferred cable configuration prevents any ambiguity. The standard
male/female cable configuration also allows easy insertion into a test setup for
substitution measurement.
Nonstandard cables can be measured, but not as easily. For a male/male cable
configuration, a female/female adapter is inserted on one end of the cable.
Likewise, for a female/female cable configuration, a male/male adapter is
inserted on one end. The adapter must be physically measured to determine the
delay through the adapter; then it is divided by two and subtracted from the
reading of the entire cable being measured. This is the most accurate technique
available for measuring the absolute length of non-standard cables;
unfortunately, it relies on measuring the adapter mechanically. In many
applications, it is not necessary to determine the exact length of a cable, but
only to assure that two cables are electrically the same. This requirement
simplifies the use of nonstandard cable since the adapter can be ignored.
The most important factor in cable and connector interface, is constant
impedance. There is a variety of constant impedance connectors availablein
typical impedance of 50 and 70 ohms (many others are also available)and
for all shapes of cable (including coaxial, triaxial, parallel, flat, and twisted).
The following connectors are typical constant impedance connectors.
SMA connectors are semi-precision, 50 ohm subminiature connectors
designed to exhibit low loss, low VSWR, and operate to 18GHz with
semi-rigid and flexible cables. They are threaded (1/4-36) connectors,
originally designed to duplicate the performance characteristics of 0.141 inch
diameter semi-rigid cable. SMA connectors are the preferred connector for
high-performance applications.
APC-3.5 connectors are precision, 50 ohm coaxial connectors designed to
exhibit low VSWR, low loss, and operate to 34GHz with rigid air line and
semi-rigid cable. They are threaded (1/4-36) connectors and mate with SMA
connectors (providing VSWR performance typical of SMA mated pairs).
These connectors also provide an air dielectric mating face and thicker outer
conductor shoulders.

13

APC-7 connectors are precision, 50-ohm coaxial connectors designed to

exhibit low VSWR, low loss, and operate to 18GHz with rigid air line (7 mm
line size), precision semi-rigid coax and some flexible coaxial cables. These
connectors are sexless, coplanar connectors, with an air dielectric mating face
and thicker outer conductor shoulders. Therefore, any two connectors can be
connected.
TNC connectors are weatherproof, 50 ohm miniature connectors designed for
extreme vibration or where safety is paramount, such as in medical and test
equipment. They operate to 11GHz with flexible cables, and have a threaded
(7/16-28) coupling.
BNC connectors are lightweight, 50 ohm miniature connectors designed for
extreme vibration or where safety is paramount, such as in medical and test
equipment. They operate to 4GHz (11GHz usable) with flexible cables, and
have a bayonet-type coupling for quick connect/disconnect.
N connectors are weatherproof, 50 or 70 ohm medium size, higher power
connectors designed to exhibit a consistently low VSWR and operate to
11GHz with flexible cables. They are threaded (5/8-24) connectors designed
to impedance match 50 or 70 ohm cables.
SMB connectors are semi-precision, 50 ohm subminiature connectors
designed for system OEM use in video and IF systems (4GHz) with flexible
cables. They are snap-on connectors.
SMC connectors are semi-precision, 50 ohm subminiature connectors
designed for low frequency applications, such as system OEM use in video
and IF systems (10GHz) with flexible cables. They are screw-on connectors.

Effects of Discontinuities
The following data and diagrams illustrate the impact of discontinuities on
time measurement. The four models are the result of computerized circuit
emulation using Micro-Cap II (Spectrum Software, Sunnyvale, CA). The fifth
model is an actual example.
The ideal model, which is the starting point for all of the variations, is
illustrated in the following.

14

The following diagram illustrates the delay in cable measurement that results
from placing a cable of mismatched impedance into the ideal model.

The following diagram illustrates the change in delay readings when the
terminating impedance is mismatched.

15

Termination Impedance

The following diagram illustrates the change in delay readings when the
terminating capacitance is mismatched.

Load Capacitance

The following diagram illustrates the change in delay readings when a

capacitive element is introduced into the middle of the ideally-matched cable.

16

Center Tap Capacitance

The following diagram illustrates an actual series of measurements, showing

the delay in cable measurement that results from placing a cable of
mismatched impedance in a test setup.

17

Conclusions
Manufacturers technical specifications are always representative values and a
single cable sample may vary. Therefore, all of the nominal values of a cable
can be, and usually are, different from those of a single cable sample. They
also vary from manufacturer to manufacturer.
For time domain applications, awareness of these variations must constantly
be taken into consideration, especially when applications require specific
lengths and specific time delays. The accuracy that can be achieved in the
high performance time measurement of cables, is in direct proportion to the
amount of effort expended on good engineering practices.

Current Measurement Practices

Mechanical
The easiest, yet least accurate, way to measure cable lengths is the mechanical
method. The time delay required from a single cable sample is computed using
the dielectric constant and the mechanical information supplied by the connector
manufacturer. The cable is mechanically cut to length and the connectors are
installed. The total calculation must consider not only the time delay through the
cable, but also the delay added by the connectors. The total time delay is measured
from the face of one connector to the face of the other connector of the cable (see
the above paragraphs on cable connector installation).
It is very difficult to mechanically measure a piece of cable to thousandths of
an inch. This difficulty in cutting the length correctly, and cutting it perfectly
perpendicular, affects the total accuracy. Additional errors occur when installing
the connectors, since the cable conductors can be slid into a variety of positions
within the connectors. Soldering and crimping move and bend conductors and
dielectric, which cause changes in impedance and capacitance. Add to this the
variations that occur in impedance, capacitance, inductance, and velocity from
cable sample to cable sample, and it is easy to understand the limitations of
this method.

Time Domain Reflectometry

Time Domain Reflectometry (TDR) is the most common means of measuring
cables. It is also a good way to perform strip line evaluations, computer
backplane measurements, and printed circuit board testing. The following
discussion deals specifically with cable measurement.
TDR equipment sends a pulse through a transmission line and measures the
reflections caused by any impedance change. The pulse sent into the cable is
called the incidental traveling wavefront. The magnitude of the reflected
wave, called the reflection coefficient, is dependent on the impedance
discontinuities encountered as the incident wave passes through the cable.

18

The TDR is calibrated for length or time on the horizontal axis of a display.
The reflection coefficient is displayed on the vertical axis. The TDR can detect
and display any impedance discontinuity, including opens, shorts, and step
changes. This allows cable lengths to be measured with moderate accuracy.
The primary disadvantage of the TDR technique is that the TDR signals
seldom represent the actual signals that are ultimately sent through the cable.
Second, the signal is not only passed through the cable once, but its reflection
is passed back through. This doubles the signals exposure to attenuation. The
following figure illustrates a typical TDR display.
Point A illustrates a theoretically ideal signal. Point B shows how the signal
has been attenuated after one pass through the cable. Point C is the signal after
it has passed through the cable twice. Notice that attenuation reduces the rise
time of the signal. Point A is the ideal signal. Point C is the actual signal. A
slow rise time delays the signal because of the slope introduced. The signal at
point D illustrates that further reading errors can occur. In this case, an
impedance mismatch has changed the rising edge and caused an even more
erroneous reading.

Another fact to consider is that the triggering point is not the typical 50%
point. The incidental signal must be measured at 25% of the total signal, and
the reflected signal must be measured at 75% of the total signal. The
measurement is further complicated by the total signal amplitude. While the
incidental signal is 1/2 the intended total signal, the reflected signal is
attenuated by passing through the cable twice, and therefore, is actually less
than 1/2 of the total signal intended. The 50% point of the reflected signal is
not necessarily 150% of the incidental pulse, but is actually 50% of the
amplitude of the reflected signal after it has fully settled.
The most practical cable measuring technique calls for a signal to be measured
at its 50% point, at the input and at the output. The amplitude as it enters the
cable is the amplitude measured on its output. This can be the trip point for
most digital electronic applications, even if the signal is attenuated.

19

The WAVECREST Digital Time System (DTS)

The WAVECREST DTS allows a simple mechanical mechanism to size cables,
and at the same time provides greater timing accuracy over the TDR. Several
practical mechanical means of cable measurement are discussed as follows.
The DTS not only provides greater timing accuracy over the TDR method,
but also greatly simplifies and speeds the measurement process. As with most
techniques, there are advantages and disadvantages.
The disadvantages to the TDR method include difficulty in setup and use
among others. The TDR test measurement signals applied to the cable usually
bear no resemblance to the actual signals used, and therefore, introduce errors.
The reflected signal edge is always different from the initial signal, which
introduces further potential for error.
Using the DTS measurement method holds only two minor disadvantages.
First, the DTS cannot measure impedance directly, since the DTS measures
the actual time through the cable without known impedance. If impedance
must be determined, the formulas specified earlier in this document can be
used for computation. Second, the total length of cable that can be measured is
limited by the source signal to less than 25 feet for picosecond accuracy. In
most applications, this is not a problem. Picosecond accuracy is typically only
required in cables that are less than 10 feet long.
Because the DTS is an instrument designed to measure the quantity of a
signal, rather than the quality, it does not present a picture of the waveform. A
precision voltage measurement is similar in that an oscilloscope is not used,
because it is not accurate enough. Rather, a precision volt meter is used, usually
with several decimal places of accuracy. Likewise, while an oscilloscope can
provide a qualitative look at time delay, when real precision is required, it is not
accurate enough. The DTS provides that added precision needed for a true
quantitative repeatable measurement.

DTS Cable Measurement Techniques

The following examples describe real and practical methods for measuring
cables. These examples take into consideration that all test setups use the
shortest possible cable fixtures to maintain good wave forms and that good
engineering practices are followed.

Bench Setup
The DTS allows two methods of measuring coax cables. The first method
uses the internal 20MHz signal source built into the DTS, and the second
method uses an external signal source. See the "Ext Cal Option on the DTS"
section for details on method one.
Method two requires a DTS, a programmable signal generator, and the
appropriate cabling/fixturing including the correct cable connectors for the
cable to be tested. The signal generator must be set to provide 0 to 5 volt into
a power splitter at a rep rate equal to or greater than the TPD of the cable to be
measured at point A. The connectors at points B and C must be connected to
provide a signal to the "Stop" input on the DTS.

20

The output of the signal generator is very important in assuring that the
generator output has a fast, clean rising edge. The slew rate must be 1.25
volts/nsec or faster to obtain good readings. A good frequency for the signal
generator is around 2.5KHz. This is less than the DTS sampling capability,
but simplifies the setup and use of the DTS (it does not require an external
trigger or gate signals). Setting the pulse width at the 2.5KHz rate is not
critical, but a few hundred nanoseconds is consistent and easy to use with
cables under a few feet long.
The DTS should be set to Burst mode with automatic trigger. Adjust the
number of samples to around 500. Allow the split signal from the signal
generator to operate the DTS in the TPD++ FCN mode. Press the "Func" key
to invoke the "Pulse Find" function. The DTS is now ready to take readings.
The lower cable is designed with two connectors that match the cable ends
of the cable to be tested. Insert the cable under test between points B and C.
Operate the DTS to obtain a new reading on the display. The difference
between the two obtained readings represents the time delay of the cable
under test.
This technique of adding in an unknown length of cable and measuring the
change in time is called the substitution method. It is the most accurate method
for measuring cables. Ultimate accuracy of less than a few picoseconds is
achievable for cable lengths under 10 feet.

Test Jig for Balanced Cable

(or impedance other than 50 ohm)
The DTS and the following illustrated test jigs, operate in 50 ohm environments.
WAVECREST designed and built the following cable test jig, which can test
balanced or unbalanced cables, and provides the ability to change impedance.

21

This example is another variation on the previously illustrated substitution

techniques (see Bench Setup). The input of the cable test set, A, fed by an
ECL level signal, is terminated to -2 volts, and provides a 50 ohm impedance
match. This input can be fed with a signal generator. The 100112 buffer splits
the signal, one for the Start input of the DTS, and the other for the Stop input.
Select and install the correct resistor values for RI and R2. For example, if the
test cable is 70 ohm cable, use 70 ohm resistors for RI and R2.
Again, the optimum frequency for the signal generator is around 2.5KHz.
This is less than the DTS sampling capability, but simplifies the setup and use
of the DTS (it does not require external trigger or gate signals). The pulse
width at the 2.5KHz rate is not critical, but a few hundred nanoseconds is
consistent and easy to use with cables under a few feet long.
The DTS should be set to the Burst mode with automatic trigger. Adjust the
number of samples required to obtain the accuracy desired. Allow the source
signal to operate the DTS, and note the displayed reading. During this initial
measurement, ensure that points B and C are connected to provide a signal to
the "Stop" input of the DTS. After the initial measurement is made, the cable
to be measured is inserted between connectors at B and C. A new measurement
is then made, and the difference is noted. The difference between the two
obtained readings represents the time delay of the cable under test.

Golden Cable
The golden cable concept is a simple way to compare similar cables. One
cable, designated the golden cable, can be a cable that has a "known" time
delay. This can be a NIST traceable cable, or a known good working cable
that is "the standard" by which all other cables are measured. Any of the test
setups described above can be used to accomplish the golden cable measurement.
To start this measurement in any of the above test setups, insert the golden
cable between connectors B and C. Make the initial measurement, the value
of golden unit. Then insert the cables to be measured. Take the new reading
and calculate the difference between the two readings. Then add or subtract
this difference to the "known" value of the golden cable.
For example, suppose a golden cable exists which has been calibrated by the
NIST, and has a "known" delay value stamped on it of 1.042 nanoseconds.
This golden cable is inserted between connectors B and C in one of the test
setups. The DTS displays 1.035 nanoseconds. The difference is 0.007 nanoseconds.
Then, the new cable to be measured is inserted and measured. The new reading
is 1.050 nanoseconds. Since the difference established with the golden cable
is 0.007 nanoseconds, it is added to 1.050, establishing a value of 1.057 nanoseconds
as the value for the new cable. Accuracy of less than a few picoseconds is achievable
using this golden cable technique.

Ext Cal Option On the DTS

For testing short cables under 10-feet in a 50 ohm environment, an option is
available on the DTS for self-contained cable measurement.
This DTS cable measurement option is simple and easy to use. It is a variation
of the DTS external calibration software. Complete operating instructions are
provided with the DTS. The main limitations are the fixed 50 ohm environment
and cables less than 10 feet in length.

22

During normal cable measurement use, the aforementioned cable fixture is

connected between the Ext Cal output connectors and the "Start/Stop" input
connectors. The DTS then makes a measurement of this fixture. After this
measurement, the cable to be measured is inserted. If the same connectors are
used on Ext Cal, and on the cable to be measured, the cable to be measured is
simply inserted between the cable fixture connector and the Ext Cal connector.
If special connectors are required, they are installed in the cable fixture at
points B and C. If they is used, points B and C must be connected to provide
the signal to the start input on the DTS during the initial measurement. After
inserting the cable to be measured, allow the DTS to operate and note the
displayed time delay reading. Refer to your DTS operating guide for details
on using the "Cable Delay Menu.

Trimming Cables
Measuring a cable is one task. Cutting or trimming a cable to the correct
length is entirely another. However, WAVECREST has worked with a variety
of vendors and users of cables and has developed at least three trimming
techniques that work well in practice. The first technique presented is
primarily for balanced or twisted wire cable. The second technique is for
coaxial cable. The third can be used on either.

Trimming Balanced or Twisted Wire Cable

This method works best with the test jig for balanced cable, or cable impedance
other than 50 ohm described above. The following descriptions refer to test
points in the previous diagram associated with that test jig.
Installation begins with the appropriate cable connector placed on one end of
the cable only. This is typically plugged into the test jig at point B. Point C is
modified with the cable piercing mechanism shown below. Point C is set up
to allow both the cable piercing mechanism and the cable connectors to provide
signal into the stop input of the DTS. This allows the cable insulation to be
breached with minimum damage to the cable.

The cable is inserted and reinserted into the mechanism until the correct
reading is obtained. The cable is cut at this point and the correct connector is
installed. When the cable piercing mechanism is connected into the system test
jig, it is connected to equal the delay introduced by the connector to be
installed. This results in the finished cable being electrically identical in length
to the cable as inserted into the piercing mechanism. Finally, the completed
cable is inserted into the test jig to assure the final assembly results in the
correct length.

23

Trimming Coaxial Cable

This method works best with the simple 50 ohm test jig described previously.
The following descriptions refer to test points in the diagram associated with
that test jig.
Installation begins with of the appropriate cable connector placed on one end
of the cable only. It is typically plugged into the testing jig at point B. Point C
is modified with the cable probe mechanism shown below. It is set up to allow
either the cable probe mechanism or the cable connectors, to provide signal
into the stop input of the DTS. This allows the cable insulation to be breached
with minimum damage to the cable.

The cable is inserted into the mechanism half way, and then the cable is forced
onto the barb located on the side of the mechanism. Because the center
conductor and shield are both in mechanical contact, a reading can be taken.
The cable can be removed and cut until it is "close" to the required length.
Then the cable can be slid in and out of the mechanism until the exact reading
is obtained. The cable is cut at this point and the correct connector is installed.
When the cable probe mechanism is connected into the system test jig, it can
be to equal the delay introduced by the connector to be installed. This results
in the finished cable being electrically identical in length to the cable inserted
into the probe mechanism. Finally, the completed cable is inserted into the test
jig to assure the final assembly results in the correct length.

24

References
Amphenol. General Line Catalog.
Blair, Byron E.. editor. Time and Frequency: Theory and Fundamentals, NIST Nomograph 140, US Department
of Commerce/National Bureau of Standards, May, 1974.
Gans, W.L. and N. S. Nahman, Shielded Balanced and Coaxial Transmission Lines - Parametric Measurements
and Instrumentation Relevant to Signal Waveform Transmission in Digital Service, NIST Technical Note 1042,
US Department of Commerce /National Bureau of Standards, June, 1981.
Harris, I. A. and R. E. Spinney. The Realization of High-Frequency Impedance Standards Using Air-Spaced
Coaxial Lines, IEEE Transactions on Instrumentation and Measurement, December, 1964.
Hold, Donald R. and Norris S. Nahman. Coaxial-Line Pulse-Response Error Due to a Planar Skin-Effect
Approximation. IEEE Transactions on Instrumentation and Measurement, Vol. IM-21. No. 4, November, 1972.
Howard W. Sams & Co., Inc., Reference Data for Radio Engineers, Indianapolis, IN 46268.
MacKenzie, T. E. and A. E. Sanderson. Some Fundamental Design Principles for the Development of Precision
Coaxial Standards and Components, IEEE Transactions on Microwave Theory and Techniques, Vol. MTT-14,
No. 1. January, 1966.
Scott, Waymond R. Jr. and Glenn S. Smith. Error Corrections for an Automated Time-Domain Network Analyzer,
IEEE Transactions on Instrumentation and Measurement, Vol. IM-35 No. 3, September, 1986.
ITT Surprenant, Coaxial Cable Design Data, Section IV, Clinton, MA, 01510.
Times Fiber Communications, Inc. RF Transmission Lines-The Complete Catalog & Handbook. Wallingford, CN.
Sealectro Corporation. RF/Microwave Coaxial Cable Assemblies (CA-1/15), Mamaroneck, NY, 10543.
Zorzy, John. Skin-Effect Corrections in Immittance and Scattering Coefficient Standards Employing Precision
Air-Dielectric Coaxial Lines. IEEE Transactions on Instrumentation and Measurement, Vol. IM-15, No. 4,
December, 1966.

25

WAVECREST Corporation

WAVECREST Corporation

World Headquarters
7275 Bush Lake Road
Edina, MN 55439
(612) 831-0030
FAX: (612) 831-4474
Toll Free: 1-800-733-7128
www.wavecrestcorp.com

West Coast Office:

1735 Technology Drive, Suite 400
San Jose, CA 95110
(408) 436-9000
FAX: (408) 436-9001
1-800-821-2272

200101-01

REV A