Ping Win Hill: AoC xChapters 3x.15.4 output monitor

After thought: Yes I'm posting a recreated schematic from the book, but it is from a sample chapter on the AoE website.

I did a Pockels cell driver, made a 7 ns wide pulse, 1200 volts, 4 MHz. One of the hardest parts was picking off a good monitor output.

After a couple of interesting failures, I wound up with this:

J1-J3, the output connector, is actually a 3-position Phoenix terminal block. It was getting hot from dielectric losses; I ripped the metal parts out of the unused J2 middle section, which helped a lot.

lf quite useful!

ombination of the two- and three-switch reversible kilovolt pulser designs. While I'm at it, I might as well include the output monitor.

e, but alas I'm having trouble with it. Here's a link to an image of the sc hematic and traces of the non-inverting input of LT1363 (v_div) and the out put of BUF634 (v_out).

TI. (The BUF634a model doesn't help.) As you can see the output is quite u nstable. The most improvement in stability is by increasing C3 to 1nF, alth ough I don't know if this is a disconnect from reality. See here in another screenshot.

build the circuit? I'm questioning if the models are even perform well at n s resolution. My other question is if you have a board layout that you coul d share. I'm curious as to how to properly shield the output monitor as ind icated in the figure.

Not sure I can help you Matt. Those composite amps can be tricky. If you can throw away signal... (and after all it's only a monitor. :^) Then I've done source terminated monitors.. Just a fast buffer and 450 + 50 ohms on the output.

George H.

elf quite useful!

combination of the two- and three-switch reversible kilovolt pulser designs . While I'm at it, I might as well include the output monitor.

ce, but alas I'm having trouble with it. Here's a link to an image of the s chematic and traces of the non-inverting input of LT1363 (v_div) and the ou tput of BUF634 (v_out).

m TI. (The BUF634a model doesn't help.) As you can see the output is quite unstable. The most improvement in stability is by increasing C3 to 1nF, alt hough I don't know if this is a disconnect from reality. See here in anothe r screenshot.

build the circuit? I'm questioning if the models are even perform well at ns resolution. My other question is if you have a board layout that you cou ld share. I'm curious as to how to properly shield the output monitor as in dicated in the figure.

Thank you, John. If I understand it, you used a capacitive voltage divider to to drive the gate of a biased fet? Do the two 22nF caps go off to someth ing unrelated in the top left corner?

And interesting observation on the dielectric losses. That isn't something I have come across.

self quite useful!

combination of the two- and three-switch reversible kilovolt pulser design s. While I'm at it, I might as well include the output monitor.

ice, but alas I'm having trouble with it. Here's a link to an image of the schematic and traces of the non-inverting input of LT1363 (v_div) and the o utput of BUF634 (v_out).

om TI. (The BUF634a model doesn't help.) As you can see the output is quite unstable. The most improvement in stability is by increasing C3 to 1nF, al though I don't know if this is a disconnect from reality. See here in anoth er screenshot.

t build the circuit? I'm questioning if the models are even perform well at ns resolution. My other question is if you have a board layout that you co uld share. I'm curious as to how to properly shield the output monitor as i ndicated in the figure.

In reality I likely only need to condition the signal to play nice with sco pe probe.

That's from the 1200 volt Pockels pulser.

I had all sorts of problems. PCBs getting hot from skin loss in traces and dielectric loss in the FR4. I fried two generations of Coilcraft inductors that, superficially, should have been fine. Skin loss again. I had to wind my own, and arrange for heat transfer to keep them from smoking.

The green thing is the Phoenix with its guts ripped out.

D'oh. Not sure how I missed that. For some reason I was thinking it was coming from the Phoenix connector.

There are some very affordable high-voltage scope probes available from Amazon.