AC/DC voltages on a ceramic / polyester capacitor

[Simon]

I'm trying to buy a new capacitor for my blown inverter. The one that failed has 400V on it so I went on eBay and had a look (I don't need enough stuff for another order on farnell) and there are capacitors rated at 1000V DC but only 400V AC. Whats the game ? I've never heard of this one before.

[tecman]

Remember, you will see the peak of the AC voltage (RMS x 1.4) which is about 600 volts.  But more important is the fact that caps that are designed  for AC operation are built to handle greater dissipation (thermal) than "standard" caps.  This is likely the couse of the derating.

paul

[saturation]

Roughly AC is the RMS value; so 400VAC = 1131VAC peak to peak.  see correction above for where the AC rating comes from. 

This should be fine about line voltage but in power AC applications and higher frequency the caps are preferably derated.

http://www.amccaps.com/technical-notes/ac-power-ratings-for-dc-rated-capacitors.html

[Simon]

Roughly AC is the RMS value; so 400VAC = 1131VAC peak to peak.  This should be fine about line voltage but in power AC applications and higher frequency the caps are preferably derated.

http://www.amccaps.com/technical-notes/ac-power-ratings-for-dc-rated-capacitors.html

that will explain why a 400V cap has failed in a mains inverter

[Simon]

so a capacitor rated for 400VAC will be ok then, the old one "leaked"

[ejeffrey]

Roughly AC is the RMS value; so 400VAC = 1131VAC peak to peak.  This should be fine about line voltage but in power AC applications and higher frequency the caps are preferably derated.

http://www.amccaps.com/technical-notes/ac-power-ratings-for-dc-rated-capacitors.html

But the capacitor doesn't see the peak-to-peak voltage, only the amplitude, which is ~560 V peak.  The reason caps have much lower AC ratings that DC is because of the power dissipation due to the ESR, and also some possible overhead for expected transients on utility AC.  It is also important to keep in mind that the AC voltage rating is at 50/60 Hz.  If you put 400 VAC @ 400 Hz or even faster you will likely see fireworks.

[Zero999]

The losses due to the ESR are probably negligible because the current won't be very high. There will also be dielectric losses but at 50Hz I can't see them being significant, at higher frequencies yes.

I think it's probably got more to do with the surge rating and voltage spikes mains capacitors have to endure.

[Simon]

Well as this is out of an inverter so there is a possibility that it was being used on a higher frequency which could explain it's failure which was obviously due to heat as it burst with a melted mix of dielectric and aluminium plate.

it is also located in an area where no air flows

[saturation]

Thanks, my bad, hero999, ejeffery.  I 'recall' a 2x estimate for converting VDC to VAC on capacitors as a rule of thumb, but wasn't sure why.

Here's a summary of failure modes in caps used in PSU.  See Table 4 p.2.

http://www.apcdistributors.com/white-papers/Power/WP-60%20Avoiding%20AC%20Capacitor%20Failures%20in%20Large%20Systems.pdf