Electronics > Beginners
Can electrolytic capacitors freeze and burst?
<< < (5/5)
Xena E:
Can they freeze? I'd say yes, however not at -20°C if they're good quality. I'm aware there are different specifications from commercial grade to mil spec.

Storage and operating parameters are also quoted at seperate figures but you would think that a respected manufacturer wouldn't use a poor component in a design that they claim will work above –30°C

It would be an interesting and infomative experiment to test a capacitor, freeze it down to the minus 20, then re-test it, I have some electrolytic capacitors here, I'll seal one on a container now, put it in the freezer, and report later.

Just out of interest what was the value and voltage rating of the burst capacitor?

X
cybermaus:
As per image of similar model device linked above: 6 pieces of 220uF 160V 105C (presumed low ESR as well)
Which means quite a wide spec margin from the 12V battery and 36Vmpp solar that was attached to it.

I was planning a similar test with the remaining non-ruptured ones if/when they would reject warranty
Out of interest, not because it would be worth trying to change their mind

CaptDon seems to think they more likely have failed due to other reasons, and the frost merely prevented me from smelling/discovering the electrolyte.
But, with a 36V to 160V voltage margin, what on earth could have made them fail. Must be a similar bad batch, and if we assume a bad batch, then all would be possible again
Siwastaja:

--- Quote from: cybermaus on Today at 07:32:32 pm ---But, with a 36V to 160V voltage margin, what on earth could have made them fail.

--- End quote ---

Internal heating from ripple current is what often kills electrolytic capacitors, not as much overvoltage (they usually sustain 10-20% overvoltage just fine).

In cold temperatures the ESR is much MUCH higher, meaning the capacitors dissipate more power than usual - but then again this is a self-rectifying situation, as they heat up then the power dissipation goes down again. But I'm not sure if there can be some sudden failure simply from localized sudden heating (high thermal power density) and temperature difference it causes, maybe it's a viable idea?

Given it's Victor and not some crappy Ebay special, the capacitor is probably not seriously underdimensioned either. So maybe quality control problem with the capacitor itself combined with the temperature-related hotspotting acting as a trigger?

Note though that if panels are rated to 36Vmpp, then Voc is much higher, and at low temperatures get even higher. Still not anywhere close to 160V, though, so voltage is unlikely cause.
CaptDon:
Why such a voltage margin?? That makes no sense at all since a 220uf@160vdc would be much more expensive than one rated for 50 or 63vdc. Are you sure that capacitor bank only had up to 36 volts on it? That hits me as very strange from a manufacturing / cost benefit point of view. The circuit boards loaded with capacitors and even super capacitors that I speak of are going into locomotives and we have some thermal chambers which use dry nitrogen which eliminates the buildup of frost and other chambers with dual stage cooling that do tend to accumulate frost on the boards which can have its own merits or drawbacks. Testing to -40C is always done, the +85C or +100C spec is set by the customer. On a side note, we have learned that ethernet phy magnetics seem to stop working above 105C and switched mode magnetics when the supply is at 100C and the magnetics are running around 130C start to also not work properly.
cybermaus:
Well, those are the models Victron sells.

75V by 15 Amp
75V by 20 Amp
100V by 30 Amp
100V by 50 Amp

It's kind of logical for Victron to not choose 100V caps for 100V device, and 160V is the next one up.

It is not Victron's fault I choose to only hook 2x 18V panels in series to it.
Actually, 4 100W panels in a 2x2, giving 400W or 27Amps theoretical max on 12V)

So I (theoretically) needed the 100V model, because 20Amp was too low.
I could have done all in 4 series but that would have surpassed the 75V model anyway, so still stuck to the 100V model.

Had they made a 75V by 30Amp, it would have been a better theoretical fit, that is true.
In retrospect, I never seen it charge more then 12 Amp (flat roof, and also usually the battery is already into absorption before we reach the middle of the day) so probably a 75x20 would have been plenty.
Navigation
Message Index
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod