Precise and accurate measurement of high value resistors

[EC8010]

That looks useful; I could have solved the problem with ten resistors rather than ten 10M plus trimming resistor (making eleven). I've never used Python, mind.

[ch_scr]

All the "UI" is the command line parameters (like the batch scripts of old). All you need is for python to be installed (and make sure to have the "add to path" option checked when installing), on windows.

[PartialDischarge]

Measuring high valued resistors used in HV dividers usually does not make much sense, since it’s the ratio that matters. So having a precise high voltage source would be more interesting and allows to precisely set the division ratio, by either adjusting the lower or higher arm, and measure the VC too.

For example I have some ceramic 80kV dividers whose ratio is laser adjusted to be better than 1%, but the overall R has like 5..10% tolerance

[floobydust]

What I have done is apply a known high voltage say 100VDC to the resistor or divider, and measure (output) current as nA to GND, or divider output voltage if the voltmeter's input resistance is known.
But there is also the voltage-coefficient of resistance so at high operating voltages, resistance changes so I'm not sure how much accuracy can be achieved.

[stevebyan]

Measuring high valued resistors used in HV dividers usually does not make much sense, since it’s the ratio that matters. So having a precise high voltage source would be more interesting and allows to precisely set the division ratio, by either adjusting the lower or higher arm, and measure the VC too.

The HP-410B VTVM range divider covers seven ranges, so there’s more than one divider ratio that has to be accurate.

[ivo]

If you have two good meters, you can use voltage divider effect to measure the high resistor. Measure first meter's impedance in its volts mode, probably 10M nominal but of course you want to know that exactly. Then measure some very stable voltage exactly. Then you can put first meter underneath the 50M resistor in series connection with this voltage. With a 10V source you would get in the neighbourhood of 1.67V reading by the voltage divider formed. Work backwards to determine the exact 50M value. You can test with other smaller resistors to get a sense of the accuracy or determine viability of method.

[Alex Nikitin]

Just in case if someone needs a set of reasonably accurate and stable resistors from 10K to 10G

Or you can just use the idea and build your own  , I use such a box for many years and find it very useful.

Cheers

Alex

[MarkT]

Up in the 10G range I would worry about leakage currents - you need proper quality coax and connectors (polythene or PTFE insulation), and low humidity.  Often a faraday cage is useful to keep out stray electric fields.

[HighVoltage]

We had a similar discussion here:

https://www.eevblog.com/forum/metrology/my-triax-cable-shielded-box-project-for-keithley-instruments/

It is not that easy to measure high value resistors.

[KK6IL]

Back in the 1970's, I was measuring insulation resistances up to 10^12 by putting the unknown R in series with a 1M 0.01% R and the pair put across an accurate 1000V from a Cohu 324A voltage calibrator. Leakage current created a voltage across the 1M resistor, which was measured with a Wavetek 704 0.01% differential voltage. A lot more accurate than I needed to test insulators but that's what I had to work with.

There's a limit on what one would be willing to spend to work on an old VOM, but an old Fluke differential, even if long out of cal would likely be accurate enough to implement the above test.

BTW, mounting a copper wire in an aluminum block does not create an insulated terminal but instead, a battery.

John   KK6IL