20kv capacitive voltmeter

Hi
I will build a high voltage (20kV) voltmeter using an ordinary 220V voltmeter series with a 20pF capacitor. My 220V voltmeter is calibrated and works good. So I must only change it’s calibration factor for 20 kV. Capacitor impedance is about 113MΩ at 50 Hz. So I calculate calibration factor for range about 1-20 kV using measurement results. But unfortunately in this range it is not constant. How can I calibrate my 20kV voltmeter to increase it’s accuracy?
Can somebody help me?
Thanks!

Making a C-R voltage divider is a bad idea because it only works for pure sine voltages. By using a sufficient large parallel capacitor for the 10 MOhm voltmeter input, you can make a capacitive divider, e.g. 20 pF/20 nF 1000:1.

The voltmeter input capacitance must be considered, it's typically in a 100 pF range.

Hi,

I personally don't think it is a good idea...

But let's try to improve it.....
Safety first. You need to be very careful with this high voltage. You should be very experienced with high voltage and with the safety regulations for those high voltages. Especially if you build it for other people to use it.
Consider redundant safety construction.

Now to the problem.
Capacitors are not ideal. They may have leakage current causing errors.
And a mayor problem will be the voltage dependency of the dielectric material.
I don't think that ceramic capacitors are useful here. Use foil capacitors. If you find polystyrene, then try this.

Also in this range of voltage the ionisation of the air may generate additional error current. It depends on voltage, geometry, air pressure, humidity and temperature, so it about impossible to compensate. Also pcb material and other isolating material may influence accuracy. Maybe the use of teflon (PTFE) ist an improvement.
Be aware that the precision depends on waveform. You can calibrate for a perfect sine, but if waveform deviates, then the voltage reading also deviates.

********
Where do you see the benefit in using capacitors instead of high voltage resistors?
For sure here also you need a careful design.
The power dissipation in a 100M resistor is 4W. For redundancy reasons I recommend to use 10 x 10M resistors in series. Consider potting in elastic potting material.
Each resistor should be rated with 1W and a voltage of 5kV continous AC. Please carefully read the datasheet often the 'rated' voltage is a 'test voltage' with a limit to 1 minute.
Again: take care of the safety regulations for this high voltage.

If you consider potting of the capacitors, then you should know that the potting material also acts as dielectric material, increasing the overall capacitance. Especially two component potting material have has large temperature coefficient in dielectric constant. Choose the material carefully.

Klaus

I also built a 35 Kv probe back in 1979. Capacitive probes are dangerous because the capacitor may remain charged.
Additionally, HV capacitors are not available in high precision, but resistors are.

Therefore, I built a long array of 1 Meg 1% resistors. I housed them inside a 6 mm PBC tube, filled with transformer oil.

It was very accurate and very safe.

I agree with FvM that a large value cap across the 10M is better than a resistive shunt. If you need capacitive coupled measurements (reject DC) then this approach is good.

It's also a good idea to put a fast clamp (low leakage) or a simple RC to slow down transients. Your frequency is fixed but often with "probe" type HV sensors the user is moving the probe around. This can cause fast transients that are coupled into the meter. A resistive only divider is more tolerant to this as the low side shunt capacitance acts as a low pass filter.

Cap dividers are used for sensing voltage for special applications such as high voltage transformers for power distribution. They are well isolated and large.

if you do choose HV plastic caps, you must design protection ccts to discharge the voltage.

For laser printers, and other applications, where 25kVdc are routinely regulated using cascaded R feedback 3500V 10~33M 1/4W resistors in a linear string. For best protection a spray on conformal coating to prevent dust and creepage breakdown is advised. then make a 1000:1 or 2000:1 for better BDV margin.

THis is what a 100kV capacitance transformer looks like.