Cap C-V test

I used an old Rohde & Schwarz KARU capacitance meter to match capacitors. It was really good at that, but sometimes it just wouldn't work. At some point, I decided to finally figure out why. It turned out the KARU applied over 300Vpp of AC at tens of kHz to my little capacitors.

Oh.

Jeroen Belleman

The fungen+scope can tell you a lot more about a cap than the single number reported by a c-meter (that you maybe can't trust anyhow.)

or in series, I believe there are some automotive standards that call for that on high current lines because ceramics tend to fail short with mechanical stress

Two of those caps in series would make about half the capacitance of two in parallel, but ESR and ESL would suffer by 4:1.

We don't see many ceramic cap failures. Our worst parts are "sealed" relays that actually aren't. Some switches, too.

We don't use surface-mount film caps. They are awful.

Not really - according to your measurements, at 25V, the caps should have about 1.46uF, so the series combination has 0.73uF instead of the 0.83uF of the parallel combination.

cu Michael

Even the PPS type?

John

One cap is .83 at 50 volts. Two in parallel would be 1.66.

The series pair is about .73 as you note. That assumes that the DC voltage divides equally.

Ceramic cap nonlinearity is weird. To get the most C, is it always better to pick the cap with the highest nameplate capacitance?

Some caps are 1/5 their rated C at their rated voltage. Bad for timing circuits.

I haven't tried all the dielectrics, but the surface-mount film caps that we have used tend to delaminate and trap wash water. I use the radial leaded epoxy-glopped parts. We actually don't use many film caps lately.

Upps, yes. Must have been lack of coffee.

Not sure. In one case where I needed about 9uF at 12V DC, standard X5R/X7R

IIRC, X7S performed much better.

Murata Simsurfing has typical Capacitance/DC bias curves for all their parts.

cu Michael

Essential to test 'em then! Are you really getting SRFs as low as just a few K?

Maybe most ceramics have similar volumetric energy storage?

That's unusual. The Venkel data sheet has gobs of temperature curves and doesn't even mention voltage effects.

I have found them to be better than any of the alternatives. In particular, they don't melt during soldering and are quite stable. Have a look. John

I wasn't measuring SRF. I eyeballed the -3 dB point where the cap impedance was 50 ohms.

Maybe 1 nH and 2.2 uF would have its SRF null around 3 MHz.

So what's the "Fo" in your diagram refer to?

The -3 dB frequency. Capacitance in uF is 3100/Fo.

Do inductors vary with current as well? I would guess only a little bit.

If they have magnetic cores, yes! In some conditions even hugely.

Jeroen Belleman

Yes.

With saturation, yes, but what about the usable range?