How to Recondition (Reform) Electrolytic Capacitors and Why

[alindumitru46]

Sorry , A capacitor can have zero ESR value?! I have measured with a chinese ESR meter like this

[mockingbird]

Well, I finally got off my fat bottom and took a serious approach to this project.

What I wanted to ask is about the initial POT for reducing the 24V from my laptop brick. I can't use the wall-wart with the multi-tap transformer and the voltage selection switch because it stopped working.

The problem is that the counterfeit Bourns 1KOhm eBay pot was either garbage, or I may have broken it, and I had to throw it out.

I'm talking about this POT:



Is it safe to test these pots without no current limiting load on them? Because when I was testing it, it burned itself out inside. There's a swiper inside that touches a long, helically-wrapped metal wire. After it burned itself out, there was no constant resistance anymore betwwen one of the pins and the middle, only the variable resistance worked.

edit: momaka - is there any reason why something like an LM317 wouldn't work here? That way I can use bog-standard inexpensive POTs of which I have plenty here just to regulate the 317, and the 317 will be handling the actual current and not the POT.

[keeney123]

What would be a good idea is if the design engineers allowed this to happen in their design. By shutting down the other part of a circuit and only applying the extra needed voltage to the electrolytes. I see this would save a lot of waste.

[mockingbird]

Well, I ordered this kit for about $2... A lot better than experimenting with shoddy POTs, and it can pass a whole lot more current through.



...not to mention much more cost-effective. Some of the circuit is unecessary, as it's going to be drawing from a DC source, but I don't think I'm going to be using their PCB, but rather I'll be integrating it on the prototype PCB.

As I understand it, the included POT (Those are only rated for 0.08 watts) is only controlling the transistor gate, so the amount of current passing through it is irelevant.

[momaka]

Originally Posted by mockingbird

Is it safe to test these pots without no current limiting load on them?

Yes, but only if you connect it right and don't exceed its current/power handling capability.

By connecting it right, I mean connect the power source between the pot's fixed resistance pins. A 1 KOhm pot with a 24V power source will dissipate 0.576 Watts and 0.024 mA of current.

Originally Posted by mockingbird

Because when I was testing it, it burned itself out inside. There's a swiper inside that touches a long, helically-wrapped metal wire. After it burned itself out, there was no constant resistance anymore betwwen one of the pins and the middle, only the variable resistance worked.

Was the pot rated to handle more than 1/2 Watt. If not, that's one reason it could have burned out. Also, if you connect the power source between the pin with variable resistance and one of the fixed resistance pins, turning the pot one way will cause the resistance to decrease and thus possibly allow a lot of current to flow through it and burn out.

Originally Posted by mockingbird

edit: momaka - is there any reason why something like an LM317 wouldn't work here?

Not at all.

In fact, that is a brilliant idea! The only reason I didn't suggest it is I was trying to keep the circuit simple. Simply wasn't aware you could find a kit for that for such a cheap price. $2 is well worth it, considering a LM317 regulator from Digikey/Mouser is typically at least $1.

Originally Posted by mockingbird

Well, I ordered this kit for about $2... A lot better than experimenting with shoddy POTs, and it can pass a whole lot more current through.

http://badcaps.net/forum/attachment....1&d=1464282735

Nice! Mind posting a link where you found this kit?

Originally Posted by mockingbird

... but I don't think I'm going to be using their PCB, but rather I'll be integrating it on the prototype PCB.

I'd say DO use their PCB, but only install the components you need for your circuit.

I see 5 diodes there. 4 of them could be for a bridge rectifier (which you won't really need if powering this from a DC source), but one of them is likely for protection of the LM317 (which you should install). Not sure what the small TO-92 transistor is for or some of the other components, but I am pretty sure it is one of the application circuits in the LM317 datasheet. So overall, I'd just assemble the whole thing minus the bridge rectifier and possibly the terminals.

Really good kit for that price, though.

Originally Posted by mockingbird

As I understand it, the included POT (Those are only rated for 0.08 watts) is only controlling the transistor gate, so the amount of current passing through it is irelevant.

Yes, the pot is controlling the LM317 adjust pin, and the current for that is really small.

[mockingbird]

Originally Posted by momaka

Was the pot rated to handle more than 1/2 Watt. If not, that's one reason it could have burned out. Also, if you connect the power source between the pin with variable resistance and one of the fixed resistance pins, turning the pot one way will cause the resistance to decrease and thus possibly allow a lot of current to flow through it and burn out.

Yes, I think this is what I did. Doh! Oh well, it was only a few bucks.

In fact, that is a brilliant idea! The only reason I didn't suggest it is I was trying to keep the circuit simple. Simply wasn't aware you could find a kit for that for such a cheap price. $2 is well worth it, considering a LM317 regulator from Digikey/Mouser is typically at least $1.

Thanks

Yup, the regulators go for between 50 cents to a dollar, depending on the brand and package type. I'm aware that I'm getting a counterfeit one, but who cares, this isn't going to handle more than an amp... Nothing that some backroom Chinese semiconductor lithography couldn't handle.

Nice! Mind posting a link where you found this kit?

Sure. It was this one in particular:
http://www.ebay.ca/itm/301777278859

I'd say DO use their PCB, but only install the components you need for your circuit.

Ok, will do. Thanks again. Can't wait to get started on these caps and start measuring their leakage current.

[watchmaker]

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.

If you see the capacitance much higher than the nominal one, immediately throw it out, it will pop in several hours or days of operation. Especially if its production date is between mid-2004 and 2010.

[eccerr0r]

Also if you get an amp of leakage current, or hundreds of milliamps of leakage current, throw it out... make sure you design your LM317T to be current limiting, and for the most part you shouldn't need a heatsink, I'd imagine you don't want to go over 10mA or so when reforming...

...If that even.

[momaka]

Originally Posted by eccerr0r

Also if you get an amp of leakage current, or hundreds of milliamps of leakage current, throw it out... make sure you design your LM317T to be current limiting, and for the most part you shouldn't need a heatsink, I'd imagine you don't want to go over 10mA or so when reforming...

...If that even.

See first page.

You absolutely DO need a series resistor for each capacitor. Recommended resistance is 1000 Ohms minimum, as per Panasonic datasheets. I typically use 1 - 15 KOhms, depending on the capacity (lower capacity -> higher series resistance and vice versa).

The 317 regulator should only be used as a constant voltage source and nothing more.

Originally Posted by watchmaker

If you see the capacitance much higher than the nominal one, immediately throw it out

Agreed. But we need to quantify that.

From my experiments here, as long as the cap's capacitance is NOT over 30-40% of the rated, then you can try to reform it to see if that restores or improves its specs. And if the cap's capacitance is NOT over 20%, then the cap should still be acceptable to use/reform, provided ESR and leakage current are not reading out of spec too.

[eccerr0r]

Originally Posted by momaka

See first page.

You absolutely DO need a series resistor for each capacitor. Recommended resistance is 1000 Ohms minimum, as per Panasonic datasheets. I typically use 1 - 15 KOhms, depending on the capacity (lower capacity -> higher series resistance and vice versa).

A 10V supply at 10mA (worst case, as said 10mA is already too high) = 1K ohms.

Just commenting that the heatsink isn't even needed on the design and you could possibly get away with a LM317L or something like that if a constant voltage is even needed ... it just needs to be close enough.

[Per Hansson]

Hi, posting a question here as it's related to the topic.
I built my reformer and used two nice voltage meter LCD's that I got for almost nothing on eBay.
The first one sits before the pot and checks the input voltage from my variable transformer, this works fine.
The second one measures the voltage across the caps being charged, so if I have 6v input it will slowly rise so I know when they are reformed.
The problem is that the input impedance is listed as just 100K, this puts a too big load on the circuit so the caps can't actually be brought to full voltage.
So I'm curious if anyone has seen any voltage monitors like this but with 1M or even ideally 10M input impedance like a real multimeter has?

[Behemot]

If its fed from the same line, of course it has high current draw. You need some which has independent power supply and only measures. I only know about digital panel voltmeters.

[mockingbird]

Originally Posted by Per Hansson

Hi, posting a question here as it's related to the topic.
I built my reformer and used two nice voltage meter LCD's that I got for almost nothing on eBay.
The first one sits before the pot and checks the input voltage from my variable transformer, this works fine.
The second one measures the voltage across the caps being charged, so if I have 6v input it will slowly rise so I know when they are reformed.
The problem is that the input impedance is listed as just 100K, this puts a too big load on the circuit so the caps can't actually be brought to full voltage.
So I'm curious if anyone has seen any voltage monitors like this but with 1M or even ideally 10M input impedance like a real multimeter has?

Per, what a coincidence, was just working on this this week.

Great idea, those digital readouts...

[Per Hansson]

Originally Posted by Behemot

If its fed from the same line, of course it has high current draw. You need some which has independent power supply and only measures. I only know about digital panel voltmeters.

It's not fed from the same line, I said input impedance for a reason.

[Behemot]

Only full-sized than http://www.ebay.com/itm/12V-10A-RED-...wAAOSwd4tT1uDE

[mockingbird]

Originally Posted by Per Hansson

Hi, posting a question here as it's related to the topic.
I built my reformer and used two nice voltage meter LCD's that I got for almost nothing on eBay.
The first one sits before the pot and checks the input voltage from my variable transformer, this works fine.
The second one measures the voltage across the caps being charged, so if I have 6v input it will slowly rise so I know when they are reformed.
The problem is that the input impedance is listed as just 100K, this puts a too big load on the circuit so the caps can't actually be brought to full voltage.
So I'm curious if anyone has seen any voltage monitors like this but with 1M or even ideally 10M input impedance like a real multimeter has?

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.

[BigTroll]

Best Advice: Old capacitors run on hope and dreams to keep them alive, believe in your device and it will last forever! follow your heart, you can never go wrong!

[kaboom]

Originally Posted by BigTroll

Best Advice: Old capacitors run on hope and dreams to keep them alive, believe in your device and it will last forever! follow your heart, you can never go wrong!

Don't forget, "if they're free, they're for me!"

[jiroy]

According to my very long experience , and my best hobby also , you cannot revive zombies . But no one can disregard the fact that many capacitors and all kind of other components didn't serve its life terms initially , due to fast and continuous upgrading we seen in last decades , so it's always worth the try , specially when you own huge amounts you don't know their status .

[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.