Yup, a discussion with Peter Schmelcher set me off on this path (C'mon Peter, post more, we don't bite...much).
Check my assumptions here.
Let's assume a standard load of 1 meg and 25 pf. (Which will actually be a 10x 10m scope probe across a 1.1111 meg R and a variable cap in the probe base).
So, for an additional 10x, you'd use a 10 meg (well, really 9 meg but we're going to get to scale where that one won't matter) - with 2.5 pf across it in series.
For 100x, you'd want .25 pf total across the 100 megs, which works out to 2.5 pf per 10 meg in the series string (there'd be 10 in series), right?
The idea would then be to have a bunch of resistors seriesed with plates on each end of each, immersed in some dielectric to get to capacity, all encased in something like pyrex or quartz pipe.
Capacity to things alongside a probe is unsolved at this point in my idea-space. I also don't want to load the signal too hard...I calculate with 350::1 I'm getting close enough to the 1/4 watt ratings of my 10m resistors....so I'd probably go 400::1.
However, with a "gimmick" wire from the extreme HV end alongside the divider of an old B&K probe, I got it pretty flat. Thing is, that's a long skinny wire at full HV and all the horrible issues that brings...I already have all the neon bulbs going nuts feet from this at pretty wimpy volts and Hz compared to the plan....and I want to put this in my shield pipe soonest (8" ID PVC with copper screen around the outside - just fits clean over the 6" CF flange on the tank and bolts on)
All thoughts appreciated. I don't see how I can document what I'm doing if I don't know myself, and there's a lotta slips 'twixt the driver and the actual output.
Let's go over here and discuss:
viewtopic.php?f=9&t=1092
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.