How to measure high voltage (1500V, very low current) accurately

[jety89]

Hi, I have a High Voltage supply on an integrated circuit board installed in a Pentium I computer, that (in theory) can give a 0-3200 V Bias Supply to a Germanium detector, and only 0-1000V multimeters to test with. I have a Soviet analog multimeter, a cheap new one (digital) and professional-looking Digital Multimeter. Their readings only agree up to a poin (200 V maybe), but they all seem to suggest that the High Voltage Bias Supply that I'm testing is out of order. It worked for over 20 years, but now I can't get a normal spectra with this spectral analyzer anymore.
I need help diagnosing this plate.

How can I test High Voltages with very low currents accurately?

The plate is propriatery, so the manufacturer doesn't include a scheme.

 

[berkeman]

Hi, I have a High Voltage supply on an integrated circuit board installed in a Pentium I computer, that (in theory) can give a 0-3200 V Bias Supply to a Germanium detector, and only 0-1000V multimeters to test with. I have a Soviet analog multimeter, a cheap new one (digital) and professional-looking Digital Multimeter. Their readings only agree up to a poin (200 V maybe), but they all seem to suggest that the High Voltage Bias Supply that I'm testing is out of order. It worked for over 20 years, but now I can't get a normal spectra with this spectral analyzer anymore.
I need help diagnosing this plate.

How can I test High Voltages with very low currents accurately?

The plate is propriatery, so the manufacturer doesn't include a scheme.

You can get a HV probe that works with your DVM -- Fluke 80K-40, for example:

http://www.specialized.net/Specialized/Fluke-80K40-40kV-High-Voltage-Probe-5296.aspx

 

[jety89]

That seems to be Waaay out of my budget. Bear in mind: I live in Ukraine, and this is leagacy hardware. Could I perhaps make a voltage divider myself? What would be the correct parameters for it?

 

[berkeman]

That seems to be Waaay out of my budget. Bear in mind: I live in Ukraine, and this is leagacy hardware. Could I perhaps make a voltage divider myself? What would be the correct parameters for it?

I googled How To Make A High Voltage Probe, and got lots of good hits. Maybe try that to see if you find what you want. And please be careful!

 

[jety89]

No worries, I *am* very methodical. I am at a point when I need advice from someone with senior experience with bias supplies, and with nuclear electronics with general. The HV seems to be present in a certain checkpoint in the circuit, when I raise it with the proprietary software, but it never reaches the exit point, and it seems to dissipate shortly after the voltage is raised, never to return. With the charges dissipating so quickly, I cannot diagnose this part of the circuit reliably. I might have dissipated the charge myself, with the act of measuring it. I cannot tell, if it is a fault in the circuit, or my actions that cause the problem in the circuit (No HV). My problem, it seems to me, is not Measuring HV, per se, My problem is not to disturb the circuit in the process of measuring it, or be able to tell tree from forest.

 

[skeptic2]

How many ohms per volt is your multimeter that has the range of 0 - 1000 volts? For instance if it is 20,000 ohms per volt then on the 0 - 1000 volt scale your meter has a resistance of 20 megohms. For it to show 1000 volts it requires 1000V/20M ohms = 50 uA of current. If you want your meter to read 2K volts full scale put 20 megohms in series with the meter. For 3KV full scale you'll need 40 megohms in series. You can select what voltage the fullscale meter reading corresponds to by selecting the appropriate series resistance.

This assumes of course that you know all the precautions about measuring high voltages.

 

[jim hardy]

And you don't want everyday resistors , they have a voltage limit typically 400 volts.

Look at high voltage resistors like this:
http://www.ohmite.com/cat/res_maximox.pdfMany DMM's are 10 megohm impedance. A 990meg resistor in series would make a 100::1 divider.

There is a brief appnote about making voltage dividers here:
http://www.ohmite.com/contact.php?function=form&which=divider&set=maximox

 

[jety89]

I think these answers fully cover my question. Thanks.

 

[jety89]

(The following shouldn't be taken too seriously)
I found this funny suggestion on an actual site, dealing with static electricity.
"Next, use a DC high-voltage power supply with a large current-limiting series resistor to charge your body to 5,000 volts. ( EXTREME DANGER! You MUST limit the current to below a few hundred microamps by using a series resistor chain. If you don't know how to handle high-volt DC power supplies safely, then don't mess with them. IF YOU DO THIS WRONG, IT CAN KILL YOU.)"

 

[jim hardy]

He's right. Those little carbon resistors that have a max voltage in the hundreds will arc over and become a short circuit.
Never try to measure the 2500 volts in a microwave oven with an everyday meter - it'll blow up.

Hope we were some help.

 

[the_emi_guy]

I have done this by stringing a bunch of 1/4W resistors in series, nice straight long chain, heatshrink over the whole thing except the last resistor that you measure across. Design it to limit voltage across each resistor to a conservative level.

 

[Averagesupernova]

(The following shouldn't be taken too seriously)
I found this funny suggestion on an actual site, dealing with static electricity.



"Next, use a DC high-voltage power supply with a large current-limiting series resistor to charge your body to 5,000 volts. ( EXTREME DANGER! You MUST limit the current to below a few hundred microamps by using a series resistor chain. If you don't know how to handle high-volt DC power supplies safely, then don't mess with them. IF YOU DO THIS WRONG, IT CAN KILL YOU.)"

Reminds me of when I was younger and was working for a manufacturer on CRT monitor assemblies less the CRT. The second anode terminal in this particular test setup was connected to a very heavily insulated alligator clip. About 5000 volts was on this wire and of course current limited. There was enough leakage through this alligator to slowly charge up your body over about 30 seconds provided you were not touching anything that would leak it back off. Anyone who walked by was likely to get a static shock from me.