50V ceramic cap exposed to 160V.....OK?

[grizedale]

hello,

We have a 10pF ceramic cap in our circuit which is exposed to a square wave of duty cycle 0.2 and its peak voltage is 160V.

The frequency is 20KHz.

Will the ceramic capacitor last only a very short time?, and if it will fail, will it fail short or open?

 

[BradtheRad]

Capacitors sometimes have WV after the volt rating, which stands for working volts.
Supposedly the peaks of waveforms are allowed to exceed this value, as long as their average value stays below the rated figure.

You are applying 160V at .2 duty cycle. This averages out to 32 V.

But I don't think I would trust a 50V capacitor to survive that indefinitely.

I think it will perforate internally and allow the 160 V to arc across.

 

[keith1200rs]

If you exceed the rated voltage the dielectric can break down and the capacitor could fail. There will be some margin in the ratings - a 50V capacitor will not fail at 51V but it is unlikely to be happy at 100V.



While you will see AC and DC ratings for capacitors, the AC ratings are aimed at low frequency sinusoidal voltages and not pulses.

Keith

 

[FvM]

An old Phycomp application note related to the topic.

Generally speaking, the said scenario is beyond safe operation conditions. Capacitors with 100 or 200V rating are available in most popular package sizes.

 

[Darktrax]

You are applying 160V at .2 duty cycle. This averages out to 32 V.

If the capacitor was across these pulses, and the impedance of the source with any series components delivering them was high enough that the pulses of energy could only charge the capacitor to an average of 32V with a super-imposed waveform of part-smoothed pulses where the peaks did not exceed 50V - then maybe so.

BUT ..

If the pulses do exceed the ability of the thin dielectric layer to hold them back, even briefly in a 20kHz waveform, then the dielectric will break down! This capacitor is only 10pF. It's impedance at 20kHz is more than 795 K*ohm. It is not going to have much of a chance of levelling a 160V waveform.

You can't only "not trust" it. This scenario is guaranteed to end badly!

 

[FvM]

This scenario is guaranteed to end badly!

I think there's no guarantee for anything. Neither safe operation nor failure.

To guarantee safe operation, the pulse peak voltage must not exceed the DC voltage rating. The typical breakdown voltage levels are however a multiple of the ratings.

 

[mrflibble]

If you exceed the rated voltage the dielectric can break down and the capacitor could fail. There will be some margin in the ratings - a 50V capacitor will not fail at 51V but it is unlikely to be happy at 100V.

**broken link removed**

Useful doc, thanks!


pssst ... it lives at (without the www. bit in the url).

 

[keith1200rs]

Sorry about the bum link - I have fixed it. My mobile phone browser never copies the http://www bit so I have to add it manually. I sometimes don't notice that the www isn't there in the first place.

Keith

 

[mrflibble]

Yeah, I figured it might have been a case of autocompleted link. Organic in nature in this particular case it turns out.