Electronics > Metrology

Is precision capacitor measurement with a DVM and signal gen possible?

(1/2) > >>

DAIRVINE:
Is it possible to use a DVM and sine wave source and variable resistor to get a precision capacitor/inductor measurement?

I was thinking to have the capacitor in series with the resistor powered by a sine wave. Adjust the resistor so that the voltage across the capacitor equals the voltage across the resistor. Then use trig to find the capacitor value. Use a DVM to measure the voltage from the source, across the capacitor and across the resistor. What errors are there and how do we compensate for them? What sort of accuracy for L/C/Q could we get?

For reference I have a 3456A and fluke 189.

A bridge is not the idea because that requires precision capacitors/inductors.
A DE5000 is not the idea because I hope to make precision measurements with DVM and signal source and precision resistor. And improve the precision provided by a DE500, and also the input protection.

For example: With a 1uF quality capacitor, a 0.01% resistor, a 3456A(<90pF input), a decent (possibly filtered) sine wave source with frequency to ppm, can I make capacitor measurements to 0.05%? If the frequency is varied so that the voltage across the resistor is equal to the voltage across the capacitor/inductor, then we can calculate the value. What is needed from the DVM is precision, not absolute accuracy. If we can also measure the source voltage then we should be able to calculate the Q.

Sounds good, but these ideas are always full of holes. So what is wrong with this and what accuracy can be achieved? How can I compensate the input impedance of the DVM? (or low capacitance buffer)

kripton2035:
then the next device you could buy is a lcr meter like a deree de5000...

Dave:
Yes, measuring capacitance and inductance like this is possible.

You have to bear in mind that the input impedance of the multimeter is going to significantly affect measurement of small capacitances, because impedances are going to be high (you are limited by the DMM to relatively low frequencies) and those 10Mohm and maybe tens to hundreds of picofarads on the input of the multimeter might make a difference.

The second thing you need to be aware of is that multimeters can only measure AC accurately up to tens, maybe a couple of hundred kHz. Again, not a problem with measuring fairly large capacitances (microfarads), but it does pose a problem with small capacitances, because the voltages across capacitors are going to be low.

I wouldn't know what sort of accuracy you can expect from these, because there are many sources of error in these sorts of measurements and all contribute to the final result. I think your best bet would be to get access to an accurate LCR meter to compare your results to and go from there.

Henrik_V:
Hi,
always a question of precision ;)

here : http://www.sciencedirect.com/science/article/pii/S0263224110001089
a nice method is proposed using a generator, a DMM and build a low input cap. FET driver...
Basic idea: Use one (or more)  fixed know resistance (with low inductance) and change frequency until the ac voltage equals on both legs (using a 2P2T switch)
Like that the absolute uncertainty of the AC measurements at a given freq. doesn't hurt that much ... you compare two AC voltages of almost same frequency and magnitude.

Greetings from Germany
Henrik

setq:
Yes. See: https://en.wikipedia.org/wiki/Bridge_circuit

Obviously the source is AC and you're measuring (or comparing) impedances instead of resistances.