Capacitive Voltage divider, for O-scopes (10,000:1) / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Capacitive Voltage divider, for O-scopes (10,000:1)

first previous 4 5 6 7 next last

Move Thread

LAN_403

Patrick

Fri Jul 02 2010, 05:26AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

oops! antonio i meant to change that R2 (10k) from node 4 to node 9 that was an oversite, on my part.

and if R4 (50 ohms) does nothing i will delete it all together, though i wonder why goldwasser included it?

also antonio said:
"Note that you are attenuating the signal and amplifying it again. You can use this to obtain lower cutoff frequency, by increasing C2 while keeping R6, at expense of some noise and lower upper cutoff frequency."

so what do you mean? get rid of R5 (10k)?

the foloowing is sam goldwasser's original diagram:

IntraWinding

Fri Jul 02 2010, 08:31AM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

I was wondering what the 10K at the capacitor junction was for. In Sam's original circuit he's included a pair of diodes to the op amp supply rails to prevent excessive voltage excursions at the capacitor junction from damaging the op amp. In that case the 10K drops excessive voltage and limits current when a diode clamps. Are you going to add the diodes? They connect with polarity such that they don't conduct under normal circumstances. They need to be low reverse leakage and fast switching. You could just leave them out, and remove the 10K too. Put them back if op amps blow while measuring EHT. I think the main cause of an overvoltage at the capacitor junction would be if you forgot to earth the lower capacitor, but perhaps there are other ways. The diodes will increase reliability.
You could leave the 50R output resistor (with the op amp feedback point moved to its left) and use it to match into 50R coax to the scope. I'm not sure if the scope should be set to 50R input impedance though. It depends on whether that type of op amp can handle a load as low as 100R - most can't - a 10V output would have to drive 100mA into 100R, which is 'Power Op Amp' territory!

Proud Mary

Fri Jul 02 2010, 08:59AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Goldwasser's diagram has no range switch because he appears to have 30kV as the ceiling voltage of his 3000:1 divider, putting 10V max
on the op amp input.

Don't you think TL082 might take it personally if you had 33V on the input from your 100kV max measuring range? What supply voltage will you need to make sure the input voltage doesn't fly over the rails?

Have I missed something obvious here?

IntraWinding

Fri Jul 02 2010, 11:40AM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

Patrick has change the capacitor values from Sam Goldwassers (3.3pf & 0.01uF) to (700fF & 8.5nF)

The a.c. voltage division factor is given by (C1+C2)/C1 where C1 is the input capacitor & C2 is the capacitor to earth:

Division Factors
Sams: 3031
Patricks: 12144

Sam's is presumably considered close enough to 3000 (~1%).
I'm guessing Patrick hasn't chosen the final values of his capacitors yet as I presume he wants a division factor of 10,000 to get 10V out for 100,000V in.

I'm wondering if a capacitance as low as 700fF a practical value to work with?
I'm also worried about shielding, which I see a mention of in the link Patrick gives to Sam's original diagram. I don't understand the shielding issues.

Proud Mary

Fri Jul 02 2010, 12:43PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

IntraWinding wrote ...

I'm wondering if a capacitance as low as 700fF a practical value to work with?
I'm also worried about shielding, which I see a mention of in the link Patrick gives to Sam's original diagram. I don't understand the shielding issues.

Won't stray capacitance swamp this 0.7pF? Even so-called 'hand capacitance' will be significant at such a low value.

The lower the ratio of desired C to strays, the greater the chance of error.

Still, it's not something I know very much about, so perhaps I should stick to my own simple DC divider the theory of which which hasn't changed much since Christ was a corporal.

Patrick

Fri Jul 02 2010, 05:47PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

yes Proud mary i worry about the 0.7-1.2 pF being to little. i like your DC resistive one but i have to deal with Ac.
and i haved roof flashing 0.012" thick steel-zinc, which encloses the whole hv cap, lv cap,and op-amp board too. will sheilding prevent my body capacitence from being a factor?

also i chose a 12,000 divison ratio because an inst -amp has a minium gain of 1 or greater, so i use the amp to calibrate (the caps), and isolate (the o-scope impedence) this means a gain of 1.2 would turn actual ratio of 12,000 to an effective ratio of 10,000. thus 1 volt to the scope represents 10,000 volts in the circuit.

however, since i only neeed a single amp now, i wonder if a could change the proportions for noise immunity and stray cap immunity gains?

one last point, a single prototype may answer these questions, then refine/modify it then produce the final 8 i need. its possible that the ballbearing rods will need to be replaced with rod plate-round for the HV cap.

Fabio

Fri Jul 02 2010, 06:18PM

Registered Member #122 Joined: Fri Feb 10 2006, 12:55PM
Location: Milano Italy
Posts: 148

i think that these picture can self explain the problem:

Ciao!
Fabio.

Proud Mary

Fri Jul 02 2010, 06:20PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Patrick wrote ...

yes Proud mary i worry about the 0.7-1.2 pF being to little. i like your DC resistive one but i have to deal with Ac.
and i haved roof flashing 0.012" thick steel-zinc, which encloses the whole hv cap, lv cap,and op-amp board too. will sheilding prevent my body capacitence from being a factor?

Shielding should keep 'hand capacity' at bay, so long as it is thoroughly connected to ground. With 0.7pF you definitely can't afford to have any moving conductors anywhere near the capacitor either, and microphony might try to sneak in as well.

The measurement of such tiny capacitances is outside of my experience, Patrick, and I'd guess that specially designed instruments would be needed to get anywhere near an accurate result because strays from the test leads, or the components own leads, are likely to be bigger
than the capacitance under test.

Is there some overwhelming reason why you can't use a more real-world value like 10pF, or 22pF?

Edited afterthought: I've just seen Fabio's post, which must have been written whilst I was writing mine, and he makes it look more plausible, without however explaining how 0.7pF is to be accurated measured.

Patrick

Fri Jul 02 2010, 06:49PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

i have a special set-up to make ~1pF cap and measure it with a BK precision 875B meter witch can "see" the cap if carefully calibrated but only two sig figs are available, thus the op-amp is used to refine the division caps, from lets say 12,000 - 8,000 effective, to 10,000 exactly with 5 sig figs, instead of just 2.

i suppose i could increase the HV cap to 22.0 pF.

Fabio's pic is the model i developed 8 months ago as far as shield (roof flashing).

as per Proud mary's afterthought: i cant measure 1.0000 pF with 5 sig figs, nor can i measure 10.000 nF thus the 1.0000 / 10.000 cannot equal 10,000:1 at 5 sig figs with out calibration from the amp.

Proud Mary

Fri Jul 02 2010, 06:59PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Patrick wrote ...

i have a special set-up to make ~1pF cap and measure it with a BK precision 875B meter witch can "see" the cap if carefully calibrated but only two sig figs are available, thus the op-amp is used to refine the division caps, from lets say 12,000 - 8,000 effective, to 10,000 exactly with 5 sig figs, instead of just 2.

i suppose i could increase the HV cap to 22.0 pF.

Do those extra sig figs actually mean anything, I wonder? I often suspect that all those definite-looking numbers after the point in DVMs give an impression of accuracy which is not always there.

first previous 4 5 6 7 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob