About capacitance measurement with DVMs

[The Electrician]

Many modern DVMs can measure capacitance.  They typically don't make the measurement in the same way a good LCR meter does.

The LCR meter applies a sine wave excitation of some selectable frequency to the capacitor, measures the voltage across the capacitor and the current through it.  From these, the capacitance can be calculated.

DVMs, on the other hand, apply a known current to the capacitor and measure how long the voltage across the capacitor takes to ramp up.  The voltage across the capacitor can be viewed with a scope, and when this is done one can see a series of repetitive ramps, as the capacitor is charged, a measurement is made, the capacitor is discharged, and the process is repeated continuously.

The ramp up is slower when more capacitance is being measured.  With hundreds or thousands of microfarads, the repetition rate of the ramping process is typically on the order of several times per second, down to maybe once per second.

Because the applied excitation is a ramp rather than a sine wave, we can't directly compare the two methods when measuring electrolytic capacitors, because electrolytic caps nearly always have a capacitance vs. frequency characteristic that isn't constant with frequency.

But, after making a number of measurements I find that the rep rate of the measurements on a DVM (the rep rate is faster with smaller caps) is somewhat analogous to the measurement frequency as applied by a true LCR meter.  What I find is that if the rep rate for a .1 uF is about 40 repetitions per second, the measured value of the capacitance is close to the value measured with an LCR meter at 40 Hz.  Similarly for other capacitor sizes and rep rates.

For example, here is a sweep of the capacitance of a nominal 1000 uF electrolytic capacitor showing capacitance as a function of frequency from 4 Hz to 1 MHz:



The capacitance at 4 Hz is 1223 uF, and at 100 Hz it's 1173 uF.  I measured this capacitor with a Fluke 187 DVM and got a reading of 1240 uF.  If we extrapolated the nearly straight line capacitance vs. frequency curve in the image to frequencies below 4 Hz (the left edge of the sweep), somewhere down around 1 Hz would be the frequency where the measured capacitance would be 1240 uF.

I looked at the waveform across the cap when it was being measured by the Fluke 187, and the repetition of the ramps was just about 1 Hz.  So, apparently the equivalent measurement frequency of the Fluke for this capacitor was about 1 Hz.  Not many LCR meters can make a measurement at such a low frequency.

This fact that this capacitor's capacitance varies rather substantially with frequency is responsible for the difference in the value measured by a DVM, and the value measured at 100 Hz by an LCR meter, such as the DE-5000.  This difference is to be expected because the DVM is measuring the cap at a much lower "frequency" than 100 Hz, and the capacitance of the cap varies a lot with frequency.

To further show that the above explanation is correct, I measured a 20 uF film capacitor with the impedance analyzer to show that its capacitance doesn't vary with frequency:



The capacitance of this capacitor is quite constant from 4 Hz (and presumably below) up to several kHz.  The Fluke 187 gave a reading for this cap of 20.0 uF, and the DE-5000 at 100 Hz reads 20.01 uF.  You can see in the image that at 4 Hz, the measured capacitance is 19.997 uF, and at 100 Hz, it's about 20.02 uF.

The Fluke 187 DVM gives the same value as the LCR meter, because the capacitance doesn't have the large variation with frequency that the electrolytic did.

The Fluke 187 has an accuracy spec of 1% when measuring capacitance in this range, yet with many electrolytics there is a difference greater than than when compared to the reading an LCR meter gives at 100 Hz.  This doesn't mean that there is something wrong with either meter.  The equivalent frequency of the DVM reading is down around 1 Hz, and the value of the capacitor at that frequency is just what the DVM reads.

The different value read by the DVM is not due to the ESR of the cap, or any other parasitics; it's entirely due to the fact that the capacitance of the electrolytic cap varies substantially with frequency.

Now, I did an interesting test to show that ESR is not responsible for the different value read by the DVM.  I connected a resistance in series with the 1000 uF electrolytic and took a reading with the Fluke 187, and with the LCR meter.  I took several readings with increasing values of resistance in series with the cap, and here are the readings I got:

Added R     Fluke reading     LCR reading
zero ohms      1240 uF           1170 uF
10 ohms         1240 uF           1170 Uf
100 ohms       1240 uF           1170 uF
400 ohms       1240 uF           1170 uF
500 ohms       1250 uF           1180 uF
600 ohms       1280 uF           1180 uF
1000 ohms     1400 uF           1180 uF

Notice how the Fluke DVM continues to read the correct value even with 400 ohms in series with the cap.  This resistance is MUCH larger than the ESR of a non-defective capacitor, and shows that ESR isn't the cause of the different value read by a DVM compared to an LCR meter.

I tried this with a couple of other DVMs and found that they weren't nearly as good as the Fluke.  A Yokogawa DVM read OK with 10 ohms in series, but with 100 ohms in series, the display became unstable and just flickered; it couldn't get a reading at all.

A Gossen Metrahit also couldn't cope with much additional resistance; the reading became unstable.

It would be an interesting test for people with different DVMs to try.  If anyone does try it, please report your experience here.

[BravoV]

Great thread  , missed and didn't realize this after you pointed out at other thread, will report mine once I had the chance.