reforming electrolytic capacitors

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
problem in reforming capacitors

hello,
i have a bag of new nichicon 470µf 25 volts Muse capacitors for about 7 years on the shelf.
Yesterday i decided to try the reforming process.
i use a 230/40 volts AC transformer,connect a bridge and a 39 khoms resistor and see what voltage i get.
It was 12 volts.
Then i connect the capacitor and see with my multimeter what's going on..
After around a minute voltage increase to 27 volts!!!
I power off as i was afraid to damage the capacitor.. and do not understand why with a 12 volts initial i see that increase in voltage
Any explanation is welcome...
 
A 40Vac transformer has a nominal output voltage of 40Vac.
Passing that into a rectifier and smoothing capacitor will give ~56Vdc at the capacitor.
If you then make allowance for transformer regulation and variations in mains voltage the capacitor voltage could be anywhere from 55Vdc to 70Vdc.

DO NOT apply that to a 25V capacitor !!!!!!!!

Let's take an example.
assume the smoothing capacitor voltage is 67Vdc.
The reformed capacitor has risen to 25Vdc.
The voltage difference across the 39k current limiting resistor is 67-25 = 42Vdrop
The current is limited to 42/39000 = 1.08mA

That is the current you are forcing into the reforming capacitor.
Some is reform current (a chemical reaction), some is charging current that increases the capacitor voltage and some is leakage current.

That 1.08mA will charge the capacitor to a higher voltage until the charging portion drops to zero. Then the reform current will eventually drop to zero and all that is left is the leakage current.
This leakage current should be much less than 0.03CV
Expect it to be less than 1% of that specification figure (that's when you throw the cap in the bin)
i.e. your leakage current could/should be < 0.0003VC for a good condition electrolytic.
I regularly see ~0.0001CV as leakage of freshly reformed good condition large electrolytics, that's ~63uA for a 10mF 63V electro.
 
Last edited:
A couple of hours will get most of the reforming completed.
But the leakage can still be high.

I have found that a slow reform gets to a lower leakage than a faster reform.
I have found that 24hours gets down to a very low leakage.
Increasing above 24hours does not seem to give much if any leakage improvement.

But I did find that two 24hours slow reform periods with a resistor discharge back to zero volts, does improve the leakage significantly.

I have reported these findings a few times.
 
what do you call a slow reform? means a very small current? i have used a 37 kohms resistor in serie....
after one hour reforming i stop ,disconect the capacitor and test the capacitor's voltage.
i get 21,6 volts on the capacitor (this is thesame as charging voltage) and i loose around one volt after a couple of hours.( seems high leakage for me)
 
when i reform capacitors, i use a voltage source higher than
the dc working voltage of the cap, say 100 volts for an 80 volt cap,
then i use a 10k/10 watt series resistor....

why so? a shorted cap will just draw 10mA of current,
hardly enough to cause a faulty cap to explode....

then a dmm set to monitor voyage across cap terminal....
once the voltage reaches the rated dc working voltage, then
discharge the cap and put a shorting link accross cap terminals...

reforming takes at least 30 minutes...

i have found a 63v/56,000uf cap forms to more than 71 volts and leakage is in the uA range...
 
TonyTecson said:
why so? a shorted cap will just draw 10mA of current,
hardly enough to cause a faulty cap to explode....
Why is 10mA insufficient to cause an explosion? Surely a low (but still too high) leakage current just means that the cap takes longer to reach the required gas pressure for an explosion?

reforming takes at least 30 minutes...
My experience with reforming caps is at higher voltages, where several hours seems more normal.
 
Hello people,

I am currently thinking of reforming two electrolytic capacitors for use in a power amp.

The capacitors are two Nippon KMH 10000uF, 63V. The only serial number I can read on them, is 45aP4M and 54WP1M. I read somewhere that this may indicate they were made in 2014 and 2015 respectively, but I cannot be quite sure.

So, since I paid some money to get them and they have been sitting in my shelf for 1.5 year now (probably way more if they were manufactured in 2014), I was thinking of reforming them for one day to see if they can be used.

After all, they are not 20 years caps - in which case, I wouldn't quite think of using them.

Reading your posts etc, I plan to use the following: a 63 Vdc stable power supply, that will connect to both capacitors using a 1k resistor in series with each one.

Then, I will wait for one day, and measure the voltage drop across the resistors. The caps are specified for 3mA leakage current (0.02CV equals 12.6 mA and the datasheet specs the smaller one between 0.02CV or 3mA). So, if the caps are good leaky-wise, after 24 hours I should get a maximum 3V reading on each resistor. Is this correct?
 
Tony, many thanks for your reply.

I totally agree with you, but the only thing I would like to know to be 100% certain is whether the caps were made in 2014, and not 2004! I am trying to contact Nippon on this.

Even if they are 2014 though, I might give it a try since I have never done this before. :)
 
Why is 10mA insufficient to cause an explosion? Surely a low (but still too high) leakage current just means that the cap takes longer to reach the required gas pressure for an explosion?


My experience with reforming caps is at higher voltages, where several hours seems more normal.

a shorted cap is a shorted cap, nothing you can do about it, it will not explode, since your source has a high resistance in series with the cap, the resistor will limit current, i.e. 100v/10k is just 10ma...if there were no series resistor used, then that is surely a cause for the psu to burn....

i am surprised why you would even ask....

when reforming a 63v cap say, then a voltage source higher, say 100 volts is needed and a 10k 5 watt resistor is added....
 
Tony, many thanks for your reply.

I totally agree with you, but the only thing I would like to know to be 100% certain is whether the caps were made in 2014, and not 2004! I am trying to contact Nippon on this.

Even if they are 2014 though, I might give it a try since I have never done this before. :)

yes, it can't hurt, and will give you the first hand experience....
don't be surprised if your cap reforms to a much higher voltage,
i have seen those before in some of my caps...

after reforming, discharge using an incandescent lamp,
then short out the terminals with a small piece of wire to prevent
the cap from reacquiring a charge if left open terminals,
do this if not using the caps any time soon...
 
TonyTecson said:
a shorted cap is a shorted cap, nothing you can do about it, it will not explode, since your source has a high resistance in series with the cap, the resistor will limit current, i.e. 100v/10k is just 10ma...if there were no series resistor used, then that is surely a cause for the psu to burn....

i am surprised why you would even ask....
A shorted cap cannot explode, as no electrolysis can take place in a shorted cap. However, a cap with a thin oxide layer - which is why reforming is needed - can generate gas pressure. Low speed electrolysis regenerate the oxide; high speed electrolysis generate gas. I prefer to stay at or below a few mA.
 
So Tony, since my caps are 63 V, you are advising me to reform them using 100 V and a 10k 5 W in series?

If 3mA leakage is to be achieved, then cap voltage would be around 70V.. Is this to be wished?

yes, in general, reforming voltage is much higher than 63 volts......

the 3mA spec is the maximum leakage current that is acceptable, i find this a bit high, in my reforming caps, i found that leakage current is much lower than 1mA and reforming voltage is more than 63 volts....yes, in a 56000ufd/63 volt computer grade that i reformed years ago, a united chemicon cap...

i will toss that cap if the leakage is 3mA...or higher after several hours under voltage soak and will not go any lower anymore...
 
Many thanks for the answer. But it still bothers me, why is it allowed to charge a capacitor way beyond (maybe) its rated voltage, but usually caps have a specified surge voltage?

Is it because the surge is specified during a switching load condition? So that the combination of heat + surge would destroy the cap, whereas in a slow, static environment (like a 10k in series) the capacitor can take bigger voltages?
 
It is best to follow the manufacturer's recommendation regarding applied voltage and reforming capacitors. Nippon Chemi-Con has published guidelines, attached below.

The relevant paragraphs from the document are (from page 1 "Operating Voltage"):

Do not apply a DC voltage which exceeds the full rated voltage.
The peak voltage of a superimposed AC voltage (ripple voltage)
on the DC volt age must not exceed the full rated voltage.
A surge voltage value, which exceeds the full rated voltage, is
prescribed in the catalogs, but it is a restricted condition, for
especially short periods of time.

And from page 2, "Installing Capacitors":

If capacitors are stored at a temperature of 35C or more and
more than 75%RH, the leakage current may increase. In this
case, they can be reformed by applying the rated voltage
through a resistor of approximately 1kO.

Nichicon has a nearly identical set of recommendations. They also note that if stored below 35C for up to 2 years, capacitors generally do not need reforming.
 

Attachments

  • Chemicon Al Electrolytic Application Info.pdf
    105.8 KB · Views: 170
that confirms my earlier posts.....

i had a 63 volt cap that i reformed and the voltage was up to 75 volts while leakage currents were below the 1mA levels...very good caps indeed...

fyi, initial condition for an uncharged cap is a short circuit...
charging the caps the current goes down while voltage across the caps rises,
this happens in a few time constants, and depending on the cap value can be very short indeed, but the energy stored is huge....
discharge the caps with a lamp to prevent big bangs.....
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.