--r ~~~~f9

4
Microwave Memos Memo 22 Insertion of High Impedance at Output Switch to Reduce Loading on Oscillator Q (extension of Microwave Memo 20) Carl E. Baum 16 February 2004 on t -.Zc, say 100 n OSCI ~~~~f9 Zn,m 1-2---------- --r s I switch capacitance before closure Think: of a 2 port network, reactive, which puts no voltage on terminal 2 (loaded by Zc) at fo ° When switch closes => C, is replaced by a short circuit. Then power is delivered to Zc ° Is this possible? First minimize the switch capacitance by moving dielectrics away from arc position gas - 0 0 ~ - - I••.•.•. m s ~ In etal iris to make most of the witch capacitance to ground stead of the other electrode ''---'~~dielectric Note that the iris capacitance needs to be included when designing a particular resonant frequency for the oscillator. 1

Transcript of --r ~~~~f9

Page 1: --r ~~~~f9

Microwave MemosMemo 22

Insertion of High Impedance at Output Switch to Reduce Loading on Oscillator Q(extension of Microwave Memo 20)

Carl E. Baum16 February 2004

on t -.Zc, say 100n

OSCI ~~~~f9Zn,m 1-2------------rs

Iswitch capacitancebefore closure

Think: of a 2 port network, reactive, which puts no voltage on terminal 2 (loaded by Zc) at fo °

When switch closes => C, is replaced by a short circuit. Then power is delivered to Zc ° Is thispossible?

First minimize the switch capacitance by moving dielectrics away from arc position

gas

- 0 0~

- -I••.•.•.

ms

~ In

etal iris to make most of thewitch capacitance to ground

stead of the other electrode

''---'~~dielectric

Note that the iris capacitance needs to be included when designing a particular resonantfrequency for the oscillator.

1

Page 2: --r ~~~~f9

Depending on how small one makes the feedthrough via the switch capacitance, onemight actively cancel this by feeding some power from the lower-power oscillator to the outputside of the switch.

Note for, say, a 100 Q transmission line the switch inductance L has a time constant

Which is smaller than that for a 4 Q transmission line by a factor of 25.

So a 2nH gap has a time constant of 20 ps, or risetime of 40 ps. So one might get signalsthrough it up to a few GHz. The pulse width or number of cycles is governed by the length ofthe oscillator. The lesser effect of the switch inductance may give some advantage over theswitched oscillator.

Another approach would use 2 switches

high-voltageswitch

low-voltagesharpening switch

network (perhapsjust a resistor (large)to establish zeropotential beforeswitch fires

2

Page 3: --r ~~~~f9

Another approach would take the energy out of the coax at a position of an appropriatevoltage null. Then by quickly shorting the line at some new position by a switch, the originalvoltage null position is moved away from the output and power is delivered to the load.

voltage null,current max

--------~~l·~----------n2 n22 2

short?

••I shorting switcpower in at

short

appropriateposition

.toload

hcircumferential

Sliding contacts at ends can tune up the null position to allow for perturbation due to switchcapacitance.

Note now there is no switch in series with the load. Any inductance in the shortingswitch can be used to fine tune the switch position.

3

Page 4: --r ~~~~f9

A dual approach might use a series shorting switch to change an open to a short circuit

nA2

A+-4

voltage null,current max-- ..·t ---- nA A

- + - ------.~2 4

I ~V'

r shortinpower m atappropriateposition

..,to load

gswitch,series

When the switch closes the open circuit (+1 reflection) goes to a short (-1 reflection) changingthe phase of the reflection and making the center go to a voltage maximum, thereby sendingpower to the load. For maximum transfer of power to the load we need Zc = 2 ZL .

The left end can also be made a short by shifting its position a quarter wavelength.

4