Motivation
• Investigating worst-case transients in the
shorted stubs in E-H tuner
• B. Foster had concern about transient standing
waves that could be generated in stubs, with
possible large amplitudes
• Simulate E-H tuner behavior, investigate
amplitude of generated standing wave
• Used PSpice 9.1, Student version (free):
Available at
http://www.orcad.com/downloads/demo/default.asp
The components concerned
Isolator Shorted Stub
Klystron
Branchline Coupler Hybrid
Transmission
Line Shorted Stub
Resonant
Cavity
Klystron/Isolator Model
• Simulate Klystron as sine-wave voltage
source, with 50 Ω front-termination resistor
for isolator
SPICE Model
Branchline Hybrid
• Four quarter-wavelength transmission
lines, with impedances as shown at
left, with shorted stubs (coax, ferrite) off
of ports 2 and 3
http://www.microwaves101.com/enc
yclopedia/Quadrature_couplers.cfm
#branchline
SPICE Model
Preliminary Tests Performed
• To test setup, ran tests
with 0λ, λ/4, λ/2
difference in stub lengths,
with 50 Ω to ground
termination resistor –
power split as expected
• Added constant length to
SPICE Model
both stubs, gives pure
phase-shifted output Port 4
Electrical length
difference (λ) Voltage Output (port 4,V)
• Assume that stub impedance
0 1
doesn’t change, only electrical length
0.25 0
0.5 1
Electrical Length Difference=0
Red=Drive Signal, Blue=input, port 1, Green=output, port 4
2.0V
2V
1.0V
0V 0V
-1.0V
-2V
-2.0V
0s 2s 4s 6s 8s 10s 12s 14s 16s 18s 20s 22s 24s 26s 28s 30s
V(T5:A+) V(R3:1) V(R3:2)
Time
After a few cycles, 100% of power goes to output port 4
Electrical Length Difference=λ/4
Red=Drive Signal, Blue=input, port 1, Green=output, port 4
2.0V
2V
1.0V
0V 0V
-1.0V
-2.0V
0s 1s 2s 3s 4s 5s 6s 7s 8s 9s 10s 11s 12s
V(T5:A+) V(R3:1) V(R3:2)
Time
After a few cycles, 0% of power goes to output port 4, and all power is reflected back
-2V
to port 1
Electrical Length Difference=λ/2
Red=Drive Signal, Blue=input, port 1, Green=output, port 4
2V 2.0V
1.0V
0V
0V
-1.0V
-2.0V
0s 2s 4s 6s 8s 10s 12s 14s 16s 18s 20s 22s 24s 26s 28s 30s
V(T5:A+) V(R3:1) V(R3:2)
-2V Time
After a few cycles, 100% of power goes to output port 4
Resonant Circuit
• On resonance, cavity acts like
a 50Ω resistor, thus, lose 50% Node where cavity voltage is measured
of voltage in front-termination
resistor, no phase shift
• Use this to tune LRC circuit
(with Q=~100s to save
simulation time, cavity step-up
ratio is only about 5:1 for
plotting convenience) to act
like resonator so that output Drive
voltage at port 4 approaches point
half the driving voltage as
cavity approaches resonance
SPICE Model
Cavity Resonance Test
5.0V
5.0V
0V 0V
-5.0V
-5.0V 900s 901s 902s 903s 904s 905s 906s 907s 908s 909s 910s
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks V(R5:1) V(C1:2) V(T11:A+)
V(R5:1) V(C1:2) V(T11:A+) Time
Time
Green =2V driving voltage, Red =Voltage in Cavity Node, Blue=
Voltage at Port 4
-Thus, the resonant circuit looks like a 50Ω resistance without
phase-shift on resonance
Monitoring standing waves in stubs
• Divided shorted stub into two pieces, one
fixed at length λ/4, the other variable
• Reason: Monitor maximum of any
possible standing wave (can place probe
there)
SPICE Model
Investigation on Transients
• Sine wave source that will turn on, turn off,
or jump phase instantaneously
• Look at turn-on and turn-off transients
• Look for worst case conditions: jump
klystron phase by 180o (might be caused
by control problem?)
• Timergali’s suggestion: Investigate effect
of variable-length transmission line
between hybrid and cavity
Complete Circuit Diagram
Used two
sine
sources
to jump Adjustable electrical
phase, lengths on stubs
etc.
Adjustable
electrical
length on
cable to cavity
SPICE Model
Stub Transients during Cavity Filling
15V 15V
Transient during
filling of ~2x
steady-state Stub voltages
10V 10V
5V
Drive and Cavity voltages 5V
0V
0V
-5V
-5V 900s 901s 902s 903s 904s 905s 906s 907s 908s 909s 910s
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks V(T19:A+) V(T17:A+)+10 V(T18:A+)+10 V(C2:2) V(R8:2)
V(T19:A+) V(T17:A+)+10 V(T18:A+)+10 V(C2:2) V(R8:2) Time
Time
Yellow = Voltage in resonant cavity, Green = Driving signal, Purple=output voltage
(port 4), Red= Voltage in top stub Blue= Voltage in bottom stub
Transients during Cavity shut-off
15V
Transient during
Stub voltages shut-off not larger
than steady-state
10V
5V Drive and Cavity voltages
0V
-5V
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks
V(T19:A+) V(T17:A+)+10 V(T18:A+)+10 V(C2:2) V(R8:2)
Time
Yellow = Voltage in resonant cavity, Green = Driving signal, Purple=output voltage
(port 4), Red= Voltage in top stub Blue= Voltage in bottom stub
Transient following 180o phase-
jump of Klystron
15V 15V
Transient from
Stub voltages phase-jump ~2.5x
steady-state
10V 10V
5V Drive and Cavity voltages 5V
0V
0V
-5V
-5V 490s 495s 500s 505s 510s 515s 520s
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks V(T19:A+) V(T17:A+)+10 V(T18:A+)+10 V(C2:2) V(R8:2)
V(T19:A+) V(T17:A+)+10 V(T18:A+)+10 V(C2:2) V(R8:2) Time
Time
Yellow = Voltage in resonant cavity, Green = Driving signal, Purple=output voltage
(port 4), Red= Voltage in top stub Blue= Voltage in bottom stub
Include Transmission Line to Cavity
of length λ/4 (worst case)
20V 20V
Transient during
Stub voltages filling of ~3.3x
steady-state
10V
10V
Drive and Cavity voltages
0V
0V
-10V
-10V 490s 495s 500s 505s 510s 515s 520s
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks V(T11:A+)+10 V(C1:2) VA1(T10)+10 V(R3:1) VA1(T6)
V(T11:A+)+10 V(C1:2) VA1(T10)+10 V(R3:1) VA1(T6) Time
Time
Yellow = Voltage in resonant cavity, Light Blue = Driving signal, Purple=output voltage
(port 4), Red= Voltage in top stub Blue= Voltage in bottom stub
Include Transmission Line to Cavity of
length λ/2 (no effect on worst case)
15V 15V
Transient during
filling of ~2.5x
Stub voltages steady-state
10V 10V
5V Drive and Cavity voltages 5V
0V
0V
-5V
-5V 490s 495s 500s 505s 510s 515s 520s
0s 0.1Ks 0.2Ks 0.3Ks 0.4Ks 0.5Ks 0.6Ks 0.7Ks 0.8Ks 0.9Ks 1.0Ks V(R3:1) V(T10:A+)+10 V(T11:A+)+10 V(C1:2) VA1(T6)
V(R3:1) V(T10:A+)+10 V(T11:A+)+10 V(C1:2) VA1(T6) Time
Time
Purple= Voltage in resonant cavity, Green= Driving signal, Red=output voltage (port
4), Blue= Voltage in top stub Yellow= Voltage in bottom stub
Conclusions
• PSpice model correctly reproduces expected steady-
state behavior of E-H tuner
• Significant voltage transients observed in stubs
• Worst case so far seems to be a 180o Klystron phase-
jump with the λ/4 transmission line length between
hybrid and cavity (factor of 3.3x steady-state voltage,
factor of ~10 in peak power)
• Transients are important for specifications for power-
handling in stubs
• Still need to do parameter sweep on transmission line
length to see if there are worse cases