EECS 117

Lecture 2: Transmission Line Discontinuities

Prof. Niknejad University of California, Berkeley

University of California, Berkeley

EECS 117 Lecture 2 p. 1/2

Energy to Charge Transmission Line

Rs v+ i = Z0
+

+ Vs

+ v+

Z0

The power ow into the line is given by

+ Pline

v + (0, t) = i+ (0, t)v + (0, t) = Z0

Or in terms of the source voltage

+ Pline =

Z0 Z0 + R s

Z0 Vs2 = Vs2 Z0 (Z0 + Rs )2

University of California, Berkeley

EECS 117 Lecture 2 p. 2/2

Energy Stored in Inds and Caps (I)

But where is the power going? The line is lossless! Energy stored by a cap/ind is 1 CV 2 / 1 LI 2 2 2 At time td , a length of = vtd has been charged:
1 1 2 CV = C 2 2 1 2 1 LI = L 2 2 Z0 Z0 + R s Vs Z0 + R s
2

Vs2
2

The total energy is thus

1 2 1 1 Vs2 2 LI + CV 2 = L + C Z0 2 2 2 (Z0 + Rs )2
University of California, Berkeley

EECS 117 Lecture 2 p. 3/2

Energy Stored (II)

Recall that Z0 = L /C . The total energy stored on the line at time td = /v :
Vs2 Eline (/v) = L (Z0 + Rs )2

And the power delivered onto the line in time td :

l Z0 Vs2 Pline = v = 2 v (Z0 + Rs )

Vs2 L LC C (Z0 + Rs )2

As expected, the results match (conservation of energy).

University of California, Berkeley

EECS 117 Lecture 2 p. 4/2

Transmission Line Termination

Rs v+ i = Z0
+

+ Vs

Z0 , td

+ v+

i=

vL RL

Consider a nite transmission line with a termination resistance At the load we know that Ohms law is valid: IL = VL /RL So at time t = /v , our pulse reaches the load. Since the current on the T-line is i+ = v + /Z0 = Vs /(Z0 + Rs ) and the current at the load is VL /RL , a discontinuity is produced at the load.
University of California, Berkeley

EECS 117 Lecture 2 p. 5/2

Reections
Thus a reected wave is created at discontinuity
VL (t) = v + (, t) + v (, t) 1 1 + v (, t) v (, t) = VL (t)/RL IL (t) = Z0 Z0

Solving for the forward and reected waves

2v + (, t) = VL (t)(1 + Z0 /RL ) 2v (, t) = VL (t)(1 Z0 /RL )

University of California, Berkeley

EECS 117 Lecture 2 p. 6/2

Reection Coefcient
And therefore the reection from the load is given by
V (, t) R L Z0 L = + = V (, t) R L + Z0

Reection coefcient is a very important concept for transmisslin lines: 1 L 1

L = 1 for RL = 0 (short) L = +1 for RL = (open) L = 0 for RL = Z0 (match)

Impedance match is the proper termination if we dont want any reections

University of California, Berkeley

EECS 117 Lecture 2 p. 7/2

Propagation of Reected Wave (I)

If L = 0, a new reected wave travels toward the source and unless Rs = Z0 , another reection also occurs at source! To see this consider the wave arriving at the source. Recall that since the wave PDE is linear, a superposition of any number of solutins is also a solution. At the source end the boundary condition is as follows
+ + Vs Is Rs = v1 + v1 + v2 + The new term v2 is used to satisfy the boundary condition

University of California, Berkeley

EECS 117 Lecture 2 p. 8/2

Propagation of Reected Wave (II)

The current continuity requires Is = i+ + i + i+ 2 1 1
Vs =
+ (v1 v1 + Rs + v2 ) Z0 + + + v1 + v1 + v2

+ Solve for v2 in terms of known terms

Vs =
+ But v1 =

Rs 1+ Z0

+ (v1

+ + v2 ) +

Rs 1 Z0

v1 +

Z0 Rs +Z0 Vs v1

R s + Z0 Z0 Rs Vs = Vs + 1 Z0 R s + Z 0 Z0
University of California, Berkeley

Rs + 1+ Z0

+ v2

EECS 117 Lecture 2 p. 9/2

Propagation of Reected Wave (III)

So the source terms cancel out and
+ v2

R s Z0 = v1 = s v1 Z0 + R s

The reected wave bounces off the source impedance with a reection coefcient given by the same equation as before R Z0 (R) = R + Z0 The source appears as a short for the incoming wave
+ Invoke superposition! The term v1 took care of the + source boundary condition so our new v2 only needed to compensate for the v1 wave ... the reected wave is only a function of v1
University of California, Berkeley

EECS 117 Lecture 2 p. 10/2

Bounce Diagram
We can track the multiple reections with a bounce diagram
Space T i m e
v 1
+ v1

= L v 1

td

+ v2 = v s 1 = s L v + 1

2td

v = 2

+ L v 2

= s Lv 1

3td

+ v3 = v s 2 = 2 2 + s v L 1

4td

v = 3

+ L v 3

= s Lv 1

2 3

5td

+ v4 = v s 3 = 3 3

6td

+ s L v1

/4

/2

3/4

University of California, Berkeley

EECS 117 Lecture 2 p. 11/2

Freeze time
If we freeze time and look at the line, using the bounce diagram we can gure out how many reections have occurred For instance, at time 2.5td = 2.5/v three waves have + + + been excited (v1 ,v1 , v2 ), but v2 has only travelled a distance of /2 To the left of /2, the voltage is a summation of three + + + components: v = v1 + v1 + v2 = v1 (1 + L + L s ). To the right of /2, the voltage has only two + + components: v = v1 + v1 = v1 (1 + L ).

University of California, Berkeley

EECS 117 Lecture 2 p. 12/2

Freeze Space
We can also pick at arbitrary point on the line and plot the evolution of voltage as a function of time For instance, at the load, assuming RL > Z0 and RS > Z0 , so that s,L > 0, the voltage at the load will will increase with each new arrival of a reection
vL (t) Rs = 75 RL = 150 s = 0.2 L = 0.5

+ v1 = .4

.6
v1 = .2

.64
+ v2 = .04

.66
v2 = .02

.664
+ v3 = .004

.666 vss = 2/3V

v3 = .002

td

2td

3td

4td

5td

6td

University of California, Berkeley

EECS 117 Lecture 2 p. 13/2

Steady-State Voltage on Line (I)

To nd steady-state voltage on the line, we sum over all reected waves:
+ + + + vss = v1 + v1 + v2 + v2 + v3 + v3 + v4 + v4 +

Or in terms of the rst wave on the line

+ vss = v1 (1 + L + L s + 2 s + 2 2 + 3 2 + 3 3 + L L s L s L s

Notice geometric sums of terms like k k and k+1 k . s L s L Let x = L s :

+ vss = v1 (1 + x + x2 + + L (1 + x + x2 + ))

University of California, Berkeley

EECS 117 Lecture 2 p. 14/2

Steady-State Voltage on Line (II)

The sums converge since x < 1
vss =
+ v1

L 1 + 1 L s 1 L s

Or more compactly
vss =
+ v1

1 + L 1 L s

Substituting for L and s gives

vss RL = Vs RL + Rs

University of California, Berkeley

EECS 117 Lecture 2 p. 15/2

What Happend to the T-Line?

For steady state, the equivalent circuit shows that the transmission line has disappeared. This happens because if we wait long enough, the effects of propagation delay do not matter Conversly, if the propagation speed were innite, then the T-line would not matter But the presence of the T-line will be felt if we disconnect the source or load! Thats because the T-line stores reactive energy in the capaciance and inductance Every real circuit behaves this way! Circuit theory is an abstraction
University of California, Berkeley

EECS 117 Lecture 2 p. 16/2

PCB Interconnect
Suppose = 3cm, v = 3 108 m/s, so that tp = /v = 1010 s = 100ps On a time scale t < 100ps, the voltages on interconnect act like transmission lines! Fast digital circuits need to consider T-line effects
ground conductor

PCB substrate

dielectric logic gate

University of California, Berkeley

EECS 117 Lecture 2 p. 17/2

Example: Open Line (I)

Source impedance is Z0 /4, so s = 0.6, load is open so L = 1
+ As before a positive going wave is launched v1

Upon reaching the load, a reected wave of of equal amplitude is generated and the load voltage overshoots + vL = v1 + v1 = 1.6V Note that the current reection is negative of the voltage
i v i = + = + = v i v

This means that the sum of the currents at load is zero (open)
University of California, Berkeley

EECS 117 Lecture 2 p. 18/2

Example: Open Line (II)

+ + At source a new reection is created v2 = L s v1 , and + note s < 0, so v2 = .6 0.8 = 0.48. + At a time 3tp , the line charged initially to v1 + v1 drops in value + + vL = v1 + v1 + v2 + v2 = 1.6 2 .48 = .64

So the voltage on the line undershoots < 1 And on the next cycle 5tp the load voltage again overshoots We observe ringing with frequency 2tp

University of California, Berkeley

EECS 117 Lecture 2 p. 19/2

Example: Open Line Ringing

Observed waveform as a function of time.

University of California, Berkeley

EECS 117 Lecture 2 p. 20/2

Physical Intuition: Shorted Line (I)

The intitial step charges the rst capacitor through the rst inductor since the line is uncharged There is a delay since on the rising edge of the step, the inductor is an open Each successive capacitor is charged by its inductor + in a uniform fashion ... this is the forward wave v1
i+
L

i+
L

i+
L

i+
L

v+

v+

v+

v+

University of California, Berkeley

EECS 117 Lecture 2 p. 21/2

Physical Intuition: Shorted Line (II)

The volage on the line goes up from left to right due to the delay in charging each inductor through the inductors The last inductor, though, does not have a capacitor to charge Thus the last inductor is discharged ... the extra charge comes by discharging the last capacitor As this capacitor discharges, so does its neighboring capacitor to the left Again there is a delay in discharging the caps due to the inductors
This discharging represents the backward wave v1

University of California, Berkeley

EECS 117 Lecture 2 p. 22/2