55v ceramic caps

Interesting. I guess it's Y5V type. Can you disclose the part number?

Our derating rules, mostly copied from the JPL ones, defines max 80% allowed voltage stress

Cheers

Klaus

Ceramic cap voltage ratings seem to be more determined by the c-v curve than breakdown.

NT

This guy at NASA came to a similar conclusion. Interesting!

Either 2.5 or 100, I guess.

Really good paper!

Also by time. HV failures are cumulative where the part deteriorates with the number of exposures and total time. That is why in med tech we have a very limited time to do hipot testing in production.

If you left it at 400V for a few weeks it might have failed as well.

A capacitor manufacturer told me you could put ceramic caps in series without resistors.

Cheers

At 300 volts, the current seemed stable for hours, but some sort of electromigration might be going on. I guess I could run one for weeks. The cute little Bertan HV supply is rarely used around here.

I think these 50V caps are OK at 55 volts. I hate to add new parts to stock.

Leakage is nonlinear on voltage for this kind of hi-K dielectric, so they probably self-equalize.

Aluminum electrolytics are similar.

Den torsdag den 20. juli 2017 kl. 17.42.22 UTC+2 skrev John Larkin:

but the cap with the smallest capacitance will have the highest voltage with which will cause it's capacitance to drop, maybe

allowed voltage stress

I ran AWDs at about 120v for a research project, they were fine for a fair while but deteriorated badly over time, but they were only 30v rated. Ceram ic caps are very reliable at rated V, so I don't see you getting into troub le there. I wouldn't be so confident with 100v.

NT

On Wednesday, July 19, 2017 at 4:32:27 PM UTC-7, John Larkin wrote: ...

...

Thanks John - useful measurements.

One problem I found when measuring the value of ceramic caps under bias is that the capacitance value is bias voltage history dependent - after being under bias it may take hours or days for the value at low voltage to return to the original measured value.

kevin

allowed voltage stress

I ran AWDs at about 120v for a research project, they were fine for a fair while but deteriorated badly over time, but they were only 30v rated. Ceram ic caps are very reliable at rated V, so I don't see you getting into troub le there. I wouldn't be so confident with 100v.

NT

Yikes, more complexity.

What's wrong with some decent 100V parts in stock?

Every new line is more money & space tied up. I like to specify ranges for values rather than fixed values where it makes no significant difference.

NT

Den onsdag den 19. juli 2017 kl. 23.01.23 UTC+2 skrev John Larkin:

put two in series and be done with it

We don't have any in stock. Adding a part to our inventory is a nuisance; better to use what we already have reels of.

It makes design a little more challenging/fun.