Timing-accurate HV probe

This is bound to get mixed up with things in Electronics, check both. Physics-specific stuff here, mostly.

Timing-accurate HV probe

Postby Doug Coulter » Sat Sep 23, 2017 2:29 pm

He shoots, scores, sort of. No nothing but net here.

I wanted a time/phase accurate HV probe to look at the RF (well, it's a little supersonic if not really HF) on my ion generator. No such thing existed (s) commercially that would fit.
After one failed attempt where the silicone I used to seal the quartz tube the high value R was put into evidently contaminated all the insulators and/or the mineral oil I filled it with, this was try #2.

This time i strung together 10 10 meg resistors inside a rather larger 1/2" diameter quartz tube that happens to run parallel to a long "jumper' I have for when I wanted access to both the direct grid and the 50kohm/200w ballast resistor that used to be there... Both are insulated and also inside schedule 80 PVC pipe that has grounded brass screen wire around the outside.

Seems I got "lucky" here. The waveform ain't bad for 27khz square waves - here I used 40v of good old WWII-era GR pulse generator to drive the thing, and a rather nice new scope to look at the results. FWIW, I also used a B&K HV probe, which also looked somewhat OK but had far higher noise and was also WAY out of divide spec (the wrong load impedance?).

So, this one's off by around 5::1 on the high side - it's 100megs and 100k for the divider and ?? capacities but doesn't look too shabby here re timing = pure luck. It also has an unexplained DC offset.
DS0007.PNG
Scope screen capture


My real signal isn't a square wave either, so the phase shift implied by the not-flat tops on the result might matter, I'll have to do more testing with this rig and of course in real life....
Here's some pix looking at the setup:
100_3208.JPG
Overview

100_3209.JPG
Looking down the pipe

100_3210.JPG
Seen from outside, which never really looks that good

This last is trying to show the innards from the side...which never seems to work out that well due to the screen...I tried.

Now if corona or other meanies don't bite me badly, I can know just when things are happening on the ion grid, vs what I had which was a faraday probe not that close, with some differentiation no doubt along with the DC...and I can get real transit time measurements, better than the ones shown in eye candy here.

I think (hope?) I really lucked out here. With the official probe the noise was 100x greater and accuracy no better. But on the bench...waving a hand a foot away from this thing moved the attenuation at AC around way over 10::1....here at least it's inside a pretty controlled environment (other than maybe corona/ozone). Further tests will tell.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 2956
Joined: Wed Jul 14, 2010 8:05 pm
Location: Floyd county, VA, USA

Re: Timing-accurate HV probe - testing

Postby Doug Coulter » Sat Sep 23, 2017 3:53 pm

Didn't turn out too shabby, though I didn't quite hit all the bases (eg measure the shift from the old way of knowing "start" to the new way - because now I know the new way is RIGHT).

I moved the scope trigger to the actual input to the HV transformer. You can still see where the drive square wave switched anyway - the reason for some phase difference here is that I have some series inductance between the square wave driver and the input of the HV transformer, and am operating the transformer above its internal resonance a little bit - the series L makes the whole thing resonant at this drive frequency, at least under a plasma load.

So, with basically no gas in there, the scope looks like this:
DS0008.PNG
Yellow is drive, purple is HV probe, green is current via a 1k R to ground on the HV winding ground end.

For various reasons (like core losses) the driver draws more current with no load like this, and current limits the supply I''m using (for safety) at a lower voltage than I really run things at when there's gas. If I had a grid cam picture, it'd be a black square for this condition...nothing's going on in the tank other than probably some capacity coupling to other junk in the tank.

Now, with some gas, lots, as it takes a lot to light this off - around 3.5 e-2 millibar - and note I changed the gains on some of the scope channels (ain't screen caps great?) and in fact I ran the input supply to the driver up somewhat, as it was no longer in current limit when under this load...the power was going to the gas, not the losses in the other stuff.
DS0009.PNG
Scope with big gas load

im_0000_20170923_152821.jpg
gridcam picture of this set of conditions


So, just in case changing the gas load would affect this a bunch, I tried that and went down to 1e-2 mbar or so...FWIW, my ion gage reads high on D (about 2x) and even higher when this ion source is running and making a crap-ton of ions. This isn't a complete test, a more complete one would include running with DC on the main grid (or some waveform TBD), somewhat less gas (main grid activity helps keep the ion generator lit at lower pressures, and so far the lower the pressure, the better the Q and predictability).
So, here it is with less gas:
DS0010.PNG
Scope, less gas

im_0001_20170923_152957.jpg
Not as much light with less gas


More later, no doubt, but for now I'm pretty happy with how this turned out and will now put the WWII stuff away again. "Nothing was too good for our boys" it seems, all this stuff still works after a slight refurb in the '60s (the last century, you punks).
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 2956
Joined: Wed Jul 14, 2010 8:05 pm
Location: Floyd county, VA, USA

Re: Timing-accurate HV probe

Postby Doug Coulter » Sat Sep 23, 2017 4:22 pm

Never being wise enough the let sleeping dogs lie, while re-qualifying my AR-10 (.308, filling the clips with more nasty ammo than the target stuff) in case the bear family tries to knock my house down AGAIN tonight (a 500 lb animal slamming itself against a door or wall is real noticeable around here), I did yet another run with the Faraday probe I'd been using before to set timing...and it's interesting. Looks like it starts seeing stuff more or less instantly (at least on this time scale) when the HV probe indicates that a glitch due likely to current draw is happening - they don't "start" at the same times. This again was with no HV on the mains...so there's more left to learn here. More analysis later I suppose...
DS0011.PNG
ch2 is the faraday probe (front) Lots of gas

DS0012.PNG
Lots less gas, still no main HV
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 2956
Joined: Wed Jul 14, 2010 8:05 pm
Location: Floyd county, VA, USA


Return to Metrology

Who is online

Users browsing this forum: No registered users and 1 guest

cron