How to Recondition (Reform) Electrolytic Capacitors and Why

[mockingbird]

Originally Posted by kc8adu

i reform to rated voltage then add 20% once they seem ok.
or use the surge voltage stated on the can.
look at the data sheets.some caps give the surge rating.
but i can tell you i have never needed to reform a motherboard cap.and i have installed around 20,000 caps in the last 10 years.and today i installed some 16v caps that are 15 years old.in a +12 supply for a cnc.of course they are rubycons and problems are rare.
cheap crap probably should be reformed if you have no choice but to use them.

Can you please post some info on your re-forming jig so we can get some more examples?

How do you achieve the rated voltage on certain caps? An LM317 or its variants will go up to 50 volts or so, how do you get to say 400v if you need to? Also, which method are you using to test the leakage?

Thanks

[Per Hansson]

Originally Posted by mockingbird

Per,

Just curious, did you find a solution? I'm scrapping my "prototype" if you can call it that... The board I got is very poor quality and soldering is just impossible with it, and I want to incorporate plug jacks like yours.

Yes I bought one like Behemot recommended.
Have not had time to install it because it's summer though
It was very big though so it might look a bit odd, but I think it will work fine...

[socketa]

A leaky cap (as in, capacitor with a high leakage current… not one that is physically leaking electrolyte) will usually trick your ESR meter to show lower ESR than what the capacitor may have. So if the cap has gone high ESR, your meter may not show it and you might end up putting a faulty cap back in service. To avoid this, check the capacitance of the cap. If it is higher than 20% of its specified capacitance, it is likely leaky and it is time to reform it.

Can i clarify this point?
So say the capacitor has 1000uf printed on it.
Does that mean that it is likely leaky, and would require reforming if its measured capacitance is over 1200uf?
What about capacitors that measure lower then their specified capacitance?
Will reforming be of any help to these?
Can capacitors loose their capacitance over time, in storage?
i got some 1500uf 16v MBZ Rubycons, that i believe are old stock, but they all measure around 1400uf
Their ESR is in the ballpark of 10 - 20 mOhms

[mockingbird]

Originally Posted by Per Hansson

Yes I bought one like Behemot recommended.
Have not had time to install it because it's summer though
It was very big though so it might look a bit odd, but I think it will work fine...

Can you post a link please?

One would think in a climate like Sweden summer would be the only time you'd want to do it, or at least that's the case here in Canada... Can't do any electronics projects in the dead middle of the winter, just way too darn cold indoors.

[Per Hansson]

There are a couple on eBay, I bought one like these but they seem popular so the price can be higher on them.
http://www.badcaps.net/forum/showpos...2&postcount=35

The winters are made comfy here in Sweden by my computers

[mockingbird]

Originally Posted by Per Hansson

There are a couple on eBay, I bought one like these but they seem popular so the price can be higher on them.
http://www.badcaps.net/forum/showpos...2&postcount=35

The winters are made comfy here in Sweden by my computers

Per,

In the description they state:

A 10A 75mV Shunt (not included, available in our store) must be added if the meter is used to measure current / amps; otherwise the meter will be damaged.

Just curious, the resolution on this 20V/10A model is 0.01 for both the voltage and the amperage. So the last digit is µA. How many microamps does the circuit need to draw for the last digit to switch from a 0 to a 1?

Also, would a current shunt be necessary for such low current draw?

[Drack]

Hi, I have close to 200 caps that are in spec but on the high side of capacitance don't have ESR meter so don't know high might be, it's kind of a big job to reform all those caps, a big part can hold voltage for days

[ruky con]

i have TONS of old caps that are so old that some do not even have vents
here are some of the old caps i have

#1 [old ventless elna 470uf 63vdc] [marcon 470uf 63vdc] and [nippon chemi-con 1000uf 16vdc]

#2 [nippon chemi-con 680uf 16vdc] [a weird ventless nichicon 470uf 35vdc] [no-name ventless 680uf 16vdc] [Illinois capacitor 2200uf 10vdc]

#3 [Illinois capacitor with marcon style vent 470uf 25vdc] [ventless rubycon cap 470uf 10vdc] [Illinois capacitor with older jamicon style vent 470uf 10vdc]

#4 [old bellcon cap 2200uf 25vdc]

[Drack]

40 of the ones I talked about don't have vents, they are Ruby's and nichicon

[ruky con]

Hello, i'm here to reform a old 330uf cap, the cap is rated for 35V but my PSU can't go above 28V.
Will 28V still work for a 35V cap?
thank you

[acercanto]

So I've read a fair bit on here, and a bunch of other places on reforming caps, and while a lot of it is useful knowledge, there seems to be a ton of misinformation going around.
This thread was hugely helpful in setting me straight.
My use-case has been hard to pin down. I have 11 18,000uf 19v GE electrolytics (they're enormous, like mini soda can size) that I'm working on reforming. I want to make a battery tab spot welder with them, and they've been in storage for at least 20 years, according to the surplus company I got them from. I put the first one on my bench PS at 19v and 250mA for around 13 hours over several sessions. Each afternoon when I reconnected it, I'd check the voltage before reconnecting it. It seemed to hold a bit more charge each time. After 13 hours, I thought surely it's done, and did a few "fast" charge cycles at 5A (max current on the PS) to see what would happen. I think it has developed an internal short, as the residual charge after a few hours is now only a volt or so, where before, it was holding 9 and 10 volts overnight.
The second try, I have had at 19v and 250mA for over 24 hours in multiple sessions, and I'm still hesitant to put it to work. I just put another one on, but held the current down at 10mA, and it's been cooking for about 6 hours so far.

My questions:
Is that first one dead, or can I revive it?
Is 250mA too much, even for such a big cap as this?
Without any special equipment, is there any way to test the ESR of them? I have a multimeter and a CC/CV PSU, but that's about it. Oh, I also have a subwoofer board that I could probably scrounge a few components out of. It's mostly 56v caps of varying uf, with a few resisters thrown in, and some SMD stuff.

As far as my electrical knowledge, I know there's a formula that goes like P=I/R, and it relates power, inductance, and resistance, which are somehow related to volts, amps, and current?

Any ideas? Thanks!
Acer

[momaka]

Originally Posted by acercanto

Is 250mA too much, even for such a big cap as this?

Yes, probably.

Trying to guess how much current is "too much" or "too little" can prove a little hard. This is one reason why I suggested in my first post to use a series resistor of 1 kOhms or greater - it simplifies the process. And after all, that was the recommendation in some Panasonic datasheets too.

With a 1-kOhm resistor, the initial current will be limited as follows:
Supply Voltage --------- Current Limit
10 V ------------------- 10 mA
20 V ------------------- 20 mA
30 V ------------------- 30 mA
...
100 V ------------------ 100 mA
...
etc.

However, this initial current above will decrease as the capacitor chargers up and the voltage across its terminal increases. So in the first case with 10V supply voltage: once the capacitor charges up to 5V, the 1-kOhm will limit the current to 5 mA, and once the capacitor charges up to 9V, the 1-kOhm resistor will limit the current down to 1 mA. However, if the capacitor is leaky for some reason and it want to draw 2 mA of leakage current, then the voltage across its terminals will never be able to go past 8V with that 10V supply and 1-kOhm resistor.

That said, most capacitor datasheets give a worst-case leakage current formula for each capacitor series. It is usually given as:
I = 0.01 * C * V
or
I = 0.03 * C * V
where I is the capacitor's internal leakage current in uA, C is the capacitance of the capacitor in uF, and V is the voltage across the capacitor.

I didn't want to include that formula in my first post, as I wanted to make the cap reforming process as simple as possible for people. However, I do recommend to run through this calculation, as it will give a worst-case scenario of what to expect for the leakage current from any given capacitor.

In your case with the 19V, 18000 uF capacitors, the maximum leakage current for each capacitor should be:
I = 0.03 * 18000 * 19 = 10260 uA = 10.26 mA

Thus, if any of your reformed capacitors draw more than about 11 mA after fully charging them to 19V, they may be too out-of-spec to use. As for this:

Originally Posted by acercanto

Is that first one dead, or can I revive it?

See how much current it draws after charging up. If it's 10x more than the leakage current above, I definitely wouldn't use it.

Originally Posted by acercanto

Without any special equipment, is there any way to test the ESR of them?

You can do a "spark" test. Charge the cap to at least 5V and short its leads on a metal surface. Do beware that for a large 18000 uF cap like that, you *will* get a lot of sparks and possibly pit whatever metal you touch. So keep away from flammable surfaces and substances. And probably a good idea to wear safety glasses/goggles too, just in case some hot piece of metal try to go anywhere.

[Behemot]

With such energy, I think ESR meter will be a better choice

[acercanto]

Wow, thanks, Momaka. I didn't realize the current through the resistor would diminish as the voltage went up. It makes sense, but I hadn't thought about it.
The power supply shows the amp draw, and when it is fully charged, it shows less than 1mA on all of them, even the potentially dead one, but I haven't put my multimeter inline yet.
What will the spark test tell me? Also, is it bad for them to do it before it is fully reformed?
Sorry for the dumb questions, I really need to learn more about electricity. For someone that's been building computers and taking things apart since he was a kid, you'd think I would have picked up more.

Acer

[momaka]

Originally Posted by acercanto

The power supply shows the amp draw, and when it is fully charged, it shows less than 1mA on all of them, even the potentially dead one, but I haven't put my multimeter inline yet.

Sounds good. If they all show low leakage like that, then they are probably okay.

Originally Posted by acercanto

What will the spark test tell me?

Sometimes, an electrically leaky cap can fool an ESR meter into showing that the ESR is fine. With the spark test, you can verify if that is true. If the cap is electrically leaky, it won't spark too well (if at all). Likewise, if a cap has high ESR but you don't have an ESR meter to check it, the spark test can be used to determine only if the ESR is really bad (but it won't be able to determine marginal or just slightly out-of-spec ESR).

Don't get me wrong, the spark test is hardly useful and it takes a ton of experimenting with different caps to know what is considered "good" spark and what isn't. So overall, you don't really need to do it. But in a situation without much equipment, it can come handy.

That said, you probably don't need to do it, as you will be building a tab welder, which will essentially do the same thing as the spark test does: short-circuit the capacitor as it is charged.

Originally Posted by acercanto

Also, is it bad for them to do it before it is fully reformed?

No, you can discharge a cap regardless if it's been reformed or not. If
Only when the caps have been sitting for too long and you try to apply full voltage is when you really need to do the reform procedure.

Originally Posted by acercanto

Sorry for the dumb questions, I really need to learn more about electricity. For someone that's been building computers and taking things apart since he was a kid, you'd think I would have picked up more.

Nah, no worries. There are no dumb questions. Only dumb answers sometimes.

[budm]

Shorting the two terminal of charged capacitor is also bad for the cap due to high discharging current.

[budm]

RC Time constant:
http://www.electronics-tutorials.ws/rc/rc_1.html
So when you are using the resistor you can see how long it will take to charge up the capacitor.

[momaka]

Originally Posted by budm

Shorting the two terminal of charged capacitor is also bad for the cap due to high discharging current.

And why exactly is that bad?

[Behemot]

Because you are overstepping the rated ripple current by several magnitudes?

[momaka]

Originally Posted by Behemot

Because you are overstepping the rated ripple current by several magnitudes?

That doesn't matter. You can apply ripple currents much larger than the cap is rated for, as long as you don't do it at such a rate that it causes the internal temperature of the cap to rise above the allowable limits.

Think of it like pushing 5 Watts of power through a resistor rated for only 1 Watt. Can you do it? The answer is YES - however, only if you limit the duration of that power level so that the AVERAGE power dissipation does not exceed the 1 Watt rating. In other words, you can dissipate 1 Watt of power through the 1-Watt resistor continuously. But if you want to dissipate 5 Watts, then you can only do it for 1/5 of a second every second... or 1/10 of a second twice a second... or 1/20 of a second four times in a second.... and etc.