Trans Lines
Trans Lines
Trans Lines
The difference between electricity and electronics is the difference between a toaster and
a television set.
Isaac Asimov (1920 - 1992)
OBJECTIVES
To investigate some properties of electromagnetic waves in transmission lines.
THEORY
Electromagnetic radiation can be contained and guided within structures called
transmission lines. There are many examples of transmission lines in common use. Twin lead,
often used as television antenna cable, consists of two parallel conductors separated by a plastic
spacer. A better system for transmitting radio frequency signals (up to about 10 GHz) is the
coaxial cable. The coaxial cable has cylindrical symmetry: a center conductor is surrounded by a
"shield" conductor which is usually at ground potential. The shield helps to reduce noise pickup
in the signal line. The BNC connectors on our oscilloscopes also have this coaxial geometry. At
microwave frequencies, 3 GHz ~ 300 GHz, radiation can be transmitted in hollow metal pipes
called wave guides. Wave guides usually have a rectangular cross section whose sides are
approximately one half wavelength (~1 mm ~ 10 cm) in length. In the optical domain, waves can
be transmitted by "optical fiber". Optical fiber is a transparent material which has a higher index
of refraction than the surroundings so that propagating waves are totally internally reflected.
In addition to their technological importance, there is a wealth of interesting physics
which can be studied in transmission lines. Propagating waves in transmission lines will reflect
from abrupt discontinuities (i.e. changes which occur on length scales shorter than a wavelength)
just as optical waves partially reflect off boundaries between two materials of different index of
refraction. If two boundaries are present, standing waves can occur, as on the vibrating string you
studied earlier. Alternatively, we can view the transmission line as an electrical circuit problem.
Then we would want to know the current and voltage at each point in the line. The wave and
circuit approaches are both useful, as you will see in carrying out this experiment.
The electrical properties of a transmission line can be derived directly from Maxwell's
equations, but the mathematics is very complicated. We will, instead, model the line as though it
were an electrical circuit. To do this, we imagine dividing it into sections of length x
characterized by a combination of capacitance, inductance and resistance per unit length, as
shown in Fig. 1. The capacitance is due to the electric fields between the two conductors, the
x x
2 2
i i + i
Input v cx v + v Output
x
Fig. 1 Equivalent circuit of a small length of transmission line. As indicated, the model for the
whole transmission line consists of a large number of these small sections. The quantities and c
are given per unit length of the line.
inductance describes the effect of the magnetic fields, and the resistance models the energy
losses in the metal wires and insulating materials. For technically useful transmission lines the
energy losses are small, so we will henceforth neglect the resistance terms.
In keeping with our circuit model, we will analyze the response of the line by applying
Kirchoff's laws to each segment x. Consider one section, as marked in the figure, for which the
input is v(x), i(x) and the output is v(x)+v, i(x)+i. The voltage equation for the loop through
the two inductors is
1 di 1 d
v x x (i + i) (v + v) = 0 (1)
2 dt 2 dt
i 1 v
= (2)
t x
For the loop through one inductor and the capacitor, we get
1 di q
v x =0 (3)
2 dt cx
Differentiating with respect to time and using the fact that dq/dt = i - (i + i) yields
dv 1 d 2 i i
x 2 + =0 (4)
dt 2 dt cx
v 1 i
= (5)
t c x
Differentiating Eq. 2 with respect to x and Eq. 5 with respect to t gives two equations which can
be combined to give a wave equation
2v 2v
c =0 (6)
x2 t 2
with propagation speed s = ( c)-1/2. By substituting back into Eq. 2 and 5, one can show that the
corresponding solution for the current is
1 1
i(x, t) = f (x st) g(x + st) (8)
Zc Zc
where Zc = ( /c)1/2 is called the characteristic impedance of the cable. Evidently the transmission
line can support waves of current and voltage traveling in either direction. The quantity Zc is an
impedance because it tells us how much current flows for a given applied voltage. For our
lossless line Zc is real, that is a pure resistance. If we had included energy losses in the
conductors and dielectric it would have an imaginary part and the wave equation would be much
more complicated.
To proceed further we need to specify some particular situation, such as the set-up of
Fig. 2. The signal generator starts a pulse down a transmission line of length d which is
terminated with a load ZL. When the pulse reaches the far end we expect it to reflect and travel
back toward the source. We can find the reflected signal using the general solutions for v(x,t) and
i(x,t) at the far end of the transmission line. Applying Kirchoffs Laws to the loop at the output
end we obtain
i(0,t)
d
vs
ZL
i(d,t)
x=0 x=d
Fig. 2 Finite length of transmission line driven by a voltage source and terminated at x = d with
the load impedance ZL.
where v(d,t) and i(d,t) are given by Eq. 7 and 8. Substituting and solving for g(d + st) we obtain
Z L ZC
r= (11)
Z L + ZC
The interpretation of this result is that the reflected wave is a copy of the incident pulse, but
"flipped over" to travel in the opposite direction and with a different amplitude.
Not surprisingly, the amount of reflection depends on the terminating impedance,
reaching total reflection (r = 1) when the load impedance is infinite or zero. These limits can be
understood physically by realizing that the current or voltage, respectively, of the incident and
reflected pulses must exactly cancel at the end. The special case ZL = Zc is also of interest, since
then r = 0 and there is no reflection at all. This situation is called a matched termination or just
a termination, and is often a desirable way to minimize interfering reflections in practical
situations.
If we now imagine that our signal generator produces sine waves, we can set up standing
waves on the transmission line, much as we did on the string. Here, however, we can change the
termination, and will in fact evaluate the voltage at the input, v(0,t), for both infinite and zero
load impedances. The general harmonic solution is
1/ 2
= k ( c) (13)
Inserting v(x,t) from Eq. 12 into the loop equation 9 we eventually find B = rAexp(-2ikd), leading
to the specific solution
v(0, t) = A(1+ re )e
2ikd i t
(14)
When ZL is infinite, r = +1 and therefore v(0,t) = 0 whenever exp(-2ikd) = -1. This means that we
will measure zero voltage at the input when d is an odd multiple of a quarter-wavelength, that is
when d = /4, 3/4.... Similarly, when ZL is zero, r = -1, and there will be zero voltage at the
input when d is a multiple of a half wavelength, d = /2, , ....
The last topic we will consider is the power dissipation in a terminating resistor, RL.
When there is a time-dependent voltage v across a resistor RL, the power dissipated in the resistor
is just
2
v
P= (15)
RL
where the brackets denote an RMS average. For the harmonically driven transmission line,
which we can compute from our general solutions once we have an expression for A. This is
found by solving the loop equation for the input loop of Fig. 2. The algebra is slightly messy, so
we only quote the result when Zs = Zc, which is a good approximation for our apparatus:
vs (17)
A=
2
This function approaches zero as RL approaches zero or infinity and has a broad maximum at
RL = Zc. So, to achieve maximum power transfer from source to load, the load (and source) must
be matched to the characteristic impedance of the line.
EXPERIMENTAL PROCEDURE
The basic experimental set-up is shown in Fig. 3. The transmission line under study is a
piece of 72 coaxial cable 4.3 meters long. Two signal sources are available, one for pulses and
one for sine waves. Both have output impedances of about 70, so there should be very little
reflection at the input end. The measuring device is an oscilloscope, schematically shown as a
voltmeter and input impedance Zm (which includes a contribution from the connecting cable).
All components are fitted with coaxial BNC connectors. In addition to the obvious complication
of dealing with Zs and Zm, the experimenter needs to be aware of the fact that both impedances
and the source voltage vs may be functions of frequency.
1. Pulse measurements
The first set of data will use pulse signals to determine the speed of propagation and the
reflection coefficient. When supplied with 10 V DC from the low-voltage power supply the
pulser unit produces very short voltage pulses at a repetition rate of a few kilohertz.
Using a short cable, connect the pulser to a BNC tee at the scope input, and connect the
long cable to the other arm of the tee. If the pulser is running you can now display the
transmitted and reflected pulses on the scope screen. Be sure the scope is set for a x1 probe and
for its full 200 MHz bandwidth so you have maximum time resolution. Use the Cursor function
to measure the pulse transit time and hence the propagation velocity for the cable.
Zs
vs
V ZL
Zm
Short Long
Source Cable Scope Cable
Fig. 3 Schematic of the experimental setup.
3. Power transfer
To measure the power dissipation in the load we need to move to the other end of the
transmission line. Connect one end of the long cable directly to the RF generator and the other
end to a tee on one of the terminating resistors. Use the short cable to go from the termination to
the scope. There is no obvious way to compensate for Zm, so the measurement is best done at a
fairly low frequency, say 1 or 2 MHz, where the input impedance of the scope will be much
larger than the terminating impedance.
Use the Measure function to read the RMS voltage across each of the BNC terminating
resistors on the component board and calculate the power dissipation from Eq. 15. At these low
frequencies you can get a good estimate of the source voltage vs by disconnecting the load
resistor and letting the function generator drive the scope directly. Check to see if the variation in
power with load follows Eq. 18 when you use the known values of RL and Zc.
REPORT
Your report should include answers to the questions in the text, the data and plots
requested and a physical explanation of your results where appropriate.