# Capacitive voltage measurement

#### CCarr518

Joined Sep 26, 2013
7
I need to measure the AC voltage on the surface of a plate (plate x) without making any direct contact with it. In theory I should be able to take another metal plate (plate y), secure it some distance above the plate x, and measure the induced voltage on plate y.

Is this feasible in reality? All of the "wireless" sensing methods I have came across are fairly expensive, this seems like a simple and cheap solution if I do not require high precision measurements. I understand the dimensions of plate y and the distance between the two plates would have to be optimized.

Any thoughts? I appreciate any and all input. Thank you.

#### Hypatia's Protege

Joined Mar 1, 2015
3,226
I need to measure the AC voltage on the surface of a plate (plate x) without making any direct contact with it. In theory I should be able to take another metal plate (plate y), secure it some distance above the plate x, and measure the induced voltage on plate y.

Is this feasible in reality? All of the "wireless" sensing methods I have came across are fairly expensive, this seems like a simple and cheap solution if I do not require high precision measurements. I understand the dimensions of plate y and the distance between the two plates would have to be optimized.

Any thoughts? I appreciate any and all input. Thank you.
As I understand your inquiry, what you are contemplating is theoretically sound -- as a practical matter, however, variations in frequency, dielectric properties, local ionization -- as well as interference by adjacent fields, etc... will pose significant challenges --- If the proposed device is to be dedicated (i.e. used In situ) it may be practically achievable...

As regards non-contact electrometry - you may find a study of the Kerr effect helpful

Best regards
HP

Last edited:

#### OBW0549

Joined Mar 2, 2015
3,566
I need to measure the AC voltage on the surface of a plate (plate x) without making any direct contact with it. In theory I should be able to take another metal plate (plate y), secure it some distance above the plate x, and measure the induced voltage on plate y.

Is this feasible in reality?
It might or might not be feasible, depending on the details:

What would be the typical voltage on plate X? Millivolts? Volts? Kilovolts?
Over what range of voltages do you need to make measurements, and how accurately?
What is the frequency of the voltage on plate X? Is the frequency constant?
What are the physical dimensions of plate X, and what is its shape?
How big can plate Y be, and how close can it be placed to plate X?
Can the spacing between plates X and Y be held constant, or will the spacing have to vary?
Are there any other sources of AC electric field nearby that could interfere with the measurement?
What kind of environment is involved here? Any extreme temperatures or humidity levels?

All of these factors will influence how practical this approach would be; it could be fairly easy, or it could be extremely difficult.

#### AnalogKid

Joined Aug 1, 2013
8,806
Plate y forms a capacitor with plate x, which means that any device connected to plate y will draw AC current (redundant, I know) through this capacitor. No matter how high the input impedance of the measurement device is, it will draw a current, and that current will change the potentials of both plates.

ak

#### alfacliff

Joined Dec 13, 2013
2,458
yes it would work within limits. look up capacitive voltage devider. just two plates measurement would be poor, humidity and local air pressure would affect the measurement, as well as spacing.

#### WBahn

Joined Mar 31, 2012
26,398
One issue that is going to be a problem for you is drift and coupling to other signals.

Imagine that I sneak in and hook up a battery between the reference node your are using and the plate that you are measuring the voltage of. I've now fixed the voltage that you will see regardless of the voltage on the node you are trying to indirectly measure. But I don't have to be so blatant -- if I change the voltage on some piece of metal that is anywhere near your measurement it will couple to it and be added to your measurement. Or if there is any AC moving charge anywhere nearby, such as someone passing by that walked on a carpet recently. Which events will constitute a significant upset will depend on how sensitive your measurement has to be. Shielding it, especially from low frequency excursions, might proof very challenging.

#### CCarr518

Joined Sep 26, 2013
7
All,

Thank you very much for your input! Here is some additional information.
• Plate x is not actually a plate, but a rotating steel cylinder a few inches in diameter. The expected voltage range is anywhere from 2 to 2000 V.
• The voltage on plate x will be a harmonic rich signal with a fundamental frequency of 60 Hz and significant harmonic content in the kHz range.
• The dimensions of plate y are not defined. I plan to have the plate made at my employers machine shop.
• The distance between plate x and y is not yet defined, but will remain constant during measuring.
• Plate x will be the largest source of AC fields in the general vicinity. No other sources will be close by.
• I am not sure of humidity, but temperatures may reach 50 deg-C.
Basically, there are three operating modes of plate x; one where it is at ~2 V, one at ~200 V, and one at 2000 V. I would be using this method to differentiate between the three modes.

#### OBW0549

Joined Mar 2, 2015
3,566
Basically, there are three operating modes of plate x; one where it is at ~2 V, one at ~200 V, and one at 2000 V. I would be using this method to differentiate between the three modes.
If that's the case, then I would say yes, sensing the voltage via capacitive coupling to a sense plate would be practical and all you would need would be the sense plate connected to an appropriate amplifier which would feed its output to some sort of level detection circuit. (This assumes that you DO NOT need to differentiate between the ~2 V case and zero volts; if you do need to, that complicates things drastically.)

#### CCarr518

Joined Sep 26, 2013
7
Thank you again for the input, this has really helped my understanding of the basic physics behind this. Here are two follow up questions.
1. One member implied this configuration is essentially a capacitive voltage divider. Can someone explain this? I understand the concept of a voltage divider, but I don't see how this applies to the measurement system I described. Wouldn't the system described be modeled as a single capacitor (each plate is a conductor and the air between them is the dielectric)? How is this similar to two capacitors in series?
2. Would the frequency content of the voltage on plate x be retained in the measured voltage on plate y? If I knew a certain high frequency was present in the plate x waveform, would an FFT of the measured wavefrom from plate y show this same frequency with a proportional amplitude?

#### CCarr518

Joined Sep 26, 2013
7
Please disregard the first question. I think I understand it now, the second series impedance would be the known input impedance of whatever measuring device is connected to the plate.

#### DickCappels

Joined Aug 21, 2008
6,791
Do you care about preserving the waveform?

#### OBW0549

Joined Mar 2, 2015
3,566
1. One member implied this configuration is essentially a capacitive voltage divider. Can someone explain this? I understand the concept of a voltage divider, but I don't see how this applies to the measurement system I described. Wouldn't the system described be modeled as a single capacitor (each plate is a conductor and the air between them is the dielectric)? How is this similar to two capacitors in series?
The "top" impedance of this voltage divider will be the capacitive reactance of the capacitor formed by plate X and plate Y, while the "bottom" impedance will be the impedance of whatever you're using to make your measurement (oscilloscope, voltmeter, buffer amplifier, etc.) in parallel with the capacitive reactance of the capacitor formed from plate Y and any objects in its immediate vicinity, and the capacitance of any wiring used to connect plate Y to your measurement device, as well as any extra capacitance added to achieve a specific division ratio.

2. Would the frequency content of the voltage on plate x be retained in the measured voltage on plate y? If I knew a certain high frequency was present in the plate x waveform, would an FFT of the measured waveform from plate y show this same frequency with a proportional amplitude?
Only if the input impedance of the measuring device is purely capacitive. If there is any resistive component to the measuring device's input impedance (and there will be), you will in effect be forming a high-pass filter that could attenuate the 60 Hz fundamental frequency relative to any higher frequency signal also present, depending on the actual resistance and capacitance values.

#### CCarr518

Joined Sep 26, 2013
7
I know it has been a few weeks, but I wanted to close the loop for anyone that had been following this.

I tested this general concept during the last week of May. A steel plate was suspended 1" above the rotor of a generator. During generator operation, I was able to confirm that a signal was induced onto the plate via capacitive coupling. The signal was highly attenuated (80% or so) but the frequency spectrum was preserved very well. Overall, this was definitely successful!

Thank you all that contributed to this thread, it was a huge help in this proof of concept.

#### cmartinez

Joined Jan 17, 2007
7,220
A picture, if possible, would come in handy...

#### Hypatia's Protege

Joined Mar 1, 2015
3,226
I know it has been a few weeks, but I wanted to close the loop for anyone that had been following this.

I tested this general concept during the last week of May. A steel plate was suspended 1" above the rotor of a generator. During generator operation, I was able to confirm that a signal was induced onto the plate via capacitive coupling. The signal was highly attenuated (80% or so) but the frequency spectrum was preserved very well. Overall, this was definitely successful!...
Thanks for the followup! A rare courtesy indeed - and much appreciated!

Best regards
HP