100uF Ceramics - Capacitance at higher voltage

[ACharnley]

I can't explain it, but it's happy with an 330uF or 470uF electrolytic, even if the SRF and ESR say it should be otherwise. A smaller capacitor which theoretically would give a better SRF for the 380Khz Buck doesn't work well at all.

The SRF of such a capacitor will be much lower than 1 MHz or even 500 kHz. Here's a sweep of the impedance and ESR of a similar 470 uF aluminum electrolytic. The measurement is made with essentially zero lead length:

View attachment 100316

The SRF is at about 80 kHz. With an additional 1 inch of lead length this is the result:

View attachment 100317

Now the SRF is about 40 kHz.

Also, the ESR is a lot higher than what you're getting with your MLCCs. The rated maximum impedance at 100 kHz for the 470 uF cap in your link is shown as 75 milliohms. The ESR at the SRF shown in the first image I've posted here is about 30 milliohms. And if you look carefully, you can see the slight increase due to the extra 1 inch of lead length in the second image.

I'm unable to explain why it works better (though since I don't have proper 100uF/25v MLCC's it's hard to compare). Possibly because an MLCC only has 1/4-8 of it's capacitance close to rated value.

Doesn't the SRF assume a full discharge/charge cycle?

 

[spec]

The Series Resonant Frequency (SRF) will be fine: it is nowhere near the frequency of interest ie 100Hz.

I haven't worked out the impact of the effective series resistance (ESR) but feel sure that will be fine too.

If what I say is incorrect please show calculations as they apply to the OP's application in hand

(see post #52)

spec

 

[ACharnley]

The new capacitors arrived. The ESR is higher and doesn't allow for 1A at 50Hz on the testbed, but in real-life 1A will only be available at a higher frequency so it's not an issue.

For a future revision I will investigate using a second PCB full of 100uF 16v ceramics, two in series, 400uF total, if I can secure the spec and authenticity. I don't get burnt twice!

Thanks, A.