Brief polarity reversal of electrolytic capacitors?

[Artlav]

Take a look at the nice Epcos capacitor in the attachment.
450V, 3300uF, said to be of high quality.

Notice anything strange?
If you do, lucky you.
If not, you might think that the band means negative terminal, as on every other electrolytic capacitor i've ever seen in my life.

Turns out it's a wrong assumption.
This one have the band on the positive side.

So, it was under about 150V for like 10 seconds while i stared at the dials doing - about 8A appear to just vanish into nowhere.
A few pokes shown that the capacitor was not holding the charge, and i swapped it out, preparing hot words for the seller about quality of parts they sell

Some time later i took a closer look, and noticed the unusual polarity

So, the question is - is that capacitor still fine, after a brief reversal (10-20 seconds total under 150V)?

I can't quite find any definite info - some places state immediate decomposition, some state short-term survival, some state reversible damage if the voltage wasn't too high.
It's somewhat expensive, so i wouldn't want to get a new one, but the rest of the circuit is way more expensive, so i don't want to take too much chances either.

[T3sl4co1l]

Lesson learned, use your eyes. Dipped tantalum are usually marked with a lone "+" only; molded (SMT) are usually denoted with a line, like you'd use a diode: it's the positive!

FWIW, I have a Sprague PowerLytic here within reach; the vent plug is in the same position (looking at the top face, vent up: positive right, negative left).  Not sure if that's industry standard (there's no reason for it to be -- like LED polarity based on the flatted side, or the fatter electrode..).

Electrolytics are pretty forgiving (assuming they don't explode..).  It probably still works.  But I wouldn't trust it to work at spec unless I specifically measured it.

Bring it back up to rated voltage slowly too, if you're going to keep using it.

Tim

[Fsck]

Lesson learned, use your eyes.

Electrolytics are pretty forgiving (assuming they don't explode..).  It probably still works.  But I wouldn't trust it to work at spec unless I specifically measured it.

Bring it back up to rated voltage slowly too, if you're going to keep using it.

Tim

look up IEC 384-4.

technically you're not within the 15% of Temp-derated max voltage bound, but since the time spent under reversal is extremely short, the spec should more or less work.

I agree with Tim but with the additional caveat of measuring it's properties right away and then testing a bit more after you get it to rated voltage (derated for temp + little margin in case temp changes), you should leave it charged for >125 hours to make sure it still functions correctly (and if it doesn't fail and it's properties are within spec), then it should be fine but you can expect to have expended some of its rated lifetime.

however, if it's a critical, very expensive project, or something to be sold (where reputation matters): I'd just replace it with a new one, and use the cap in a different project which does not fulfill the listed criteria.

[TrickyNekro]

(I have a very very good question for you that might save you from a lot of trouble.)

(You connected that beasty to a 150V power supply or whatever you have there, but you haven´t specified the current capabilities of the supply. ) Oops, sorry saw that 8A later...
In reverse polarity, just like Tim said, they work as diodes pretty much. If the capacitor didn´t heat up I guess it´s pretty ok  .
Still it really depends on the temperature the electrolyte in that capacitor reached. Electrolytic capacitors are pretty tough. Only imagine that even when normally charged, the
electrolyte wants to escape from the capacitor. In reverse polarity, it heats up. If it didn´t reach decomposition temperature, or didn´t boil, (which would lead to cap explosion
probably) I guess it´s ok. It depends really on the chemistry to find out what happened. Still 8A is not that much for this beast but I insist it depends on the chemistry.
You should probably find the exact material of the electrolyte and then try to find what happens to that chemistry in the event of reverse polarity.

In the end, maybe you´ll get some less capacitance out of it, maybe not, nobody can really know. What everyone agrees with is that you should test it.

Btw, I´ve read your comment with a fat russian accent. Sounded epic  !!!

Best Regards, Lefteris
Greece

[c4757p]

Wet electrolytics are indeed somewhat forgiving. Reverse-polarizing it can begin to degrade the insulating oxide layer, but putting it back around the right way can often repair this layer. Power it back up properly, but under a current limit of around 50mA, and leave it like that for an hour or so. Come back and measure the leakage current and verify that it's back to what it should be under full working voltage.

[Artlav]

It probably still works.

What everyone agrees with is that you should test it.

Well, charged it to 200V, it dropped to 150V within first minute, but was at 83V an hour later.
Does not look like there are any serious shorts.

however, if it's a critical, very expensive project, or something to be sold (where reputation matters): I'd just replace it with a new one, and use the cap in a different project which does not fulfill the listed criteria.

It's a hobby project that can fail in many other ways.
And as an input capacitor, it's not likely to kill anything else if it shorts out, will just make a mess and blow the fuses.

[Fsck]

I'd still follow the IEC's recommendations for duration and voltage to test on. (T-derated V & 125 hours)

[NiHaoMike]

That's a very big input cap. What application do you have in mind?

[Artlav]

I'd still follow the IEC's recommendations for duration and voltage to test on. (T-derated V & 125 hours)

Sure.
Meanwhile, charged it up to 360V (as high as it could possibly go, with rectified mains and variac at full), let it settle, and started measuring leakage current.
Per datasheet it should be under 6.2mA over 5 minutes.
I'm getting 0.1mA.
Seems fine so far.
Let's see how it survives till weekend.

That's a very big input cap. What application do you have in mind?

An induction heater for a few KW.

[T3sl4co1l]

Ooh, nice.  Got a project thread/website for it?

I've done a bit of that before, you could say http://webpages.charter.net/dawill/tmoranwms/Elec_IndHeat6.html

Tim

[Holograph]

In general, never assume that a white mark is negative. In many applications it is the positive. This capacitor is not the only thing like this.

If you damaged it, the change would most likely be noticeable with an ESR meter (though if you don't have a baseline measurement, you won't be able to tell how damaged it was, but should be able to tell if it's still good enough)

[tszaboo]

I have a water kettle, it is about 2000W boils a cup of water in 20 seconds. Your capacitor just had 1200W for  20 seconds, but it is lot smaller.
Although I've seen components desoldering themselves from a board because of overheat, and still working afterwards, so it might just work...

[Artlav]

Ooh, nice.  Got a project thread/website for it?

Not really.
There is a thread at 4hv with me asking stupid questions, and a few pictures/numbers - http://4hv.org/e107_plugins/forum/forum_viewtopic.php?161958.30
I've made a couple ZVS heaters up to 300W before, but this is my first attempt at a serious and mains-powered one.

I've done a bit of that before, you could say

Yeah, i recognize that site.

I have a water kettle, it is about 2000W boils a cup of water in 20 seconds. Your capacitor just had 1200W for  20 seconds, but it is lot smaller.

It was slightly warm in the end.
But it didn't have all of that for the whole 20 seconds, it was more like flip,  , flip off,    , poke, poke, flip,  , flip off,  , poke, poke,  , 

If you damaged it, the change would most likely be noticeable with an ESR meter (though if you don't have a baseline measurement, you won't be able to tell how damaged it was, but should be able to tell if it's still good enough)

LCR meter shows proper capacitance, but i can't get it to show ESR on a cap this big.
So, no idea.

[Artlav]

In the end, the capacitor appears to be fine.
It holds the charge for days well within the datasheet.
Plugged it back in (right way around), and i see no problems.

Thanks all for advice.