Arcing from large electrolytic capacitors

[zupta82]

Hello, I was wondering is it possible for arcing to discharge from the body of large electrolytic capacitors as they get old. I don’t smell any fish or see any leaking paste. I do notice that there are two large bleeder resistors. One of them gets really warm in comparison to the other. It’s a high voltage application.

[bob91343]

Show a diagram.

[zupta82]

Here is a a section of the circuit schematic where the large caps are

[Terry Bites]

Assuming your rails read correctly, you might have a leaky cap(s). If so, the chances are that the others will go the same way soon. (Who am I to Contradict Lord Carlson?) Replace it or them.

[zupta82]

I desoldered some caps out of circuit and one of the power bleeder resistors. I tried checking them with a MeterMan 37XR dmm. The caps read OL and with an audible tone no matter what range I turn the meter to. I hear this meter is not the greatest due to its high burden voltage. Also I’m guessing these cap functions on some dmm’s  are not really accurate in the first place compared to an esr meter. The resistor reads about 10k, I did notice that the black coating on this resistor looks melted. Here is a pic

[bdunham7]

I desoldered some caps out of circuit and one of the power bleeder resistors. I tried checking them with a MeterMan 37XR dmm. The caps read OL and with an audible tone no matter what range I turn the meter to. I hear this meter is not the greatest due to its high burden voltage. Also I’m guessing these cap functions on some dmm’s  are not really accurate in the first place compared to an esr meter. The resistor reads about 10k, I did notice that the black coating on this resistor looks melted. Here is a pic

I'm not really familiar with its details, but that meter should be fine if you use it appropriately.  ESR is not the issue here.  Are you using the capacitance or resistance range to test?  OL + tone seems to indicate you are using capacitance and OL is what you typically see when you measure a short in capacitance range.  Try using the resistance in autorange and being very patient--it may take several minutes for a reading to settle down.

[Jeff1946]

Take desoldered caps and charge with 9V battery (plus to plus).  Measure discharge rate with your dmm.

[zupta82]

I am definitely using the cap testing function. I did go back and use the resistance function on auto  I am now measuring a slow climb to about 10meg ohm resistance where it settled after several minutes. Here is a pic

[bdunham7]

I am definitely using the cap testing function. I did go back and use the resistance function on auto  I am now measuring a slow climb to about 10meg ohm resistance where it settled after several minutes. Here is a pic

So the maximum capacitance your meter can read is 400 uF, so those caps are too big for it, thus the OL.

If all the caps work their way up to 10M or so, then they have passed that test.  You would then have to test them at full voltage somehow or else put them back in the circuit and measure the voltage across and current through each one.  That would be quite hazardous, b/t/w, so if you do that be very careful and do everything hands-off or power off.

[bob91343]

If you want to measure leakage of an electrolytic capacitor, a DMM won't do it.  You need to apply rated voltage.  Very few devices today will apply rated voltage.  The old school capacitance bridges or teraohmmeters can do it.  Once you know the leakage you can see if that creates enough power dissipation to raise the capacitor temperature enough to raise the leakage enough to increase the power dissipation enough...

Enough already.  You get the idea.  If the leakage current stabilizes all is well.  If it keeps climbing, all is not well.

[bdunham7]

If you want to measure leakage of an electrolytic capacitor, a DMM won't do it. 

A DMM can prove a leaky capacitor bad.  It can't prove that it is good.

[Terry Bites]

You need to test them at their operational voltage or leakage problems my go un-noticed.
Have you measured the currents flowing in the bleeders?

[rsjsouza]

Hello, I was wondering is it possible for arcing to discharge from the body of large electrolytic capacitors as they get old. I don’t smell any fish or see any leaking paste. I do notice that there are two large bleeder resistors. One of them gets really warm in comparison to the other. It’s a high voltage application.

It occurred me to ask: why do you think that arcing is happening? Is there a specific symptom that is leading you to believe that could be a possibility. Do you hear any arcing sound?

I can tell that anything can arc given it is close enough.

If arcing ia happening inside any device (capacitor, resistor), it will be probably hotter than the others. As you mentioned, one of the bleeder resistors is hotter than the other - can you confirm if it is the resistor itself and not the two capacitors immediately underneath it that are getting hot and increasing the resistor temperatuve via its terminals?

If the resistor itself is getting warmer, it may have somehow reduced its resistance, thus increasing its current. This is rare, but I have seen it happen.

[zupta82]

Overall the arcing is very intermittent  I only managed to get the arcing to occur twice. First I noticed it to occur by turning the equipment on and off and it looked like it emanated out from the top of the vents of one of the caps. The second time it looked the arc came out of one of the brass terminals on the board.  Could the arcing could of conducted to the metal body of the cap.

[zupta82]

Oh I forgot to mention. Could I possibly take a current measurement with a meter in series with one of the resistors. I don’t think my meter will handle that current. I guess as long as I don’t press the pilot arc trigger on the torch and allow a large secondary capacitive discharge to occur it would basically be like checking the circuit at idle current, if that makes any sense?? Also I wonder if these bleeder resisters are for setting the time constant discharge of the capacitors??

[zupta82]

I guess these bleeders are for safety purposes. They are 10W 10kohm 1% I would imagine if the measured value is not within tolerance the capacitors would be disabled. I measured about 9.8kohm if my meter is even accurate??

[bdunham7]

I guess these bleeders are for safety purposes. They are 10W 10kohm 1% I would imagine if the measured value is not within tolerance the capacitors would be disabled. I measured about 9.8kohm if my meter is even accurate??

Close enough, it's not critical.  Did you measure all six of the caps with your DMM on resistance? 

Can you just measure the voltage between (+) and (N), and then between (-) and (N)?  Carefully of course--use alligator clips and set the meter up, then turn the power on. 

[zupta82]

I went back and checked the voltages with my dmm

Red to (+) and black(N), which measured +168V DC

Red to (N) and black(-), which measured +168V DC

I just powered the unit up and did not push the torch button, I would imagine if I reversed the test leads I would get a negative voltage

[zupta82]

Also, I forgot to mention when I switched to AC on my meter I got 28mV

[bdunham7]

Well, that all looks good and balanced.  I've no idea why your one resistor is getting hot unless the arcing you observed is causing those voltages to radically change.  You must have an intermittent issue that just isn't showing up at the moment.  Does the arcing happen at idle or just when you are using it?

[zupta82]

The arcing only occurred when I tried cycling the on/off switch. I did tighten the terminals a little more. I think I eliminated the arcing.

[james_s]

200V can't even arc in the beginning. In air, Paschen's Law tells us the minimum breakdown voltage is 327V, regardless of air gap distance (unless they are so close to have some funny quantum effects).

If the gap is zero and then opens up then 200V can certainly arc in air. An arc welder is typically 30-40V open circuit. I've had cracked solder joints and switch contacts arc too.