shelf life of new electrolytics

[Jered]

I bought 6 new F&T 50+50uf 500 volt electro caps thinking I was going to need them. Plans changed. How long can these sit before they go bad/dry up? Thanks

[pdf64]

If still unused 5 years after manufacture, I think best practice may be to charge them up to their rated voltage via a current limited supply, eg 100k resistor, every year or so.

[Stevem]

Even with new filter caps that go into repairs that I do every amp goes on my veriac for 6 hours or so to form the filters slowly.

I start out with 100 volts of V+ showing up and then go up in 6 steps per hour.
I find amps done this way have less 120 hz hum and a tighter bottom end with more punch.

[FUCHSAUDIO]

I agree with charging them to form them. If you won't use them for a year or a few, that makes sense. An old engineer friend of mine showed me how using a resistor and a power supply to form a cap was very easy and worked very well.
We basically did the math and picked a resistor that allowed a 24 hour time constant on the cap value in question. Set the supply for the rated cap voltage, and after 24 hours you should have NO voltage across the resistor and the cap should be at supply value. Discharge slowly with a resistor, not a jumper, and you're golden...

[Phil_S]

...did the math and picked a resistor that allowed a 24 hour time constant on the cap value in question. Set the supply for the rated cap voltage, and after 24 hours you should have NO voltage across the resistor and the cap should be at supply value...

Would you please give an example? Or maybe just tell me if I got it right. I have a bag of old good quality electro caps that I'd like to use at some point if I can. They are 20uf rated at 500V. I am not trained in electronics and easily get out of my comfort zone with calculations. I tried to calculate, but have no frame of reference to know if I got it right. I arrived at a 259K for the resistor. My limited understanding suggests almost no current is flowing, so a 1/2 watt resistor would be more than adequate, but no harm and probably better to use a 5W sand block resistor? Many thanks.

[sluckey]

We basically did the math and picked a resistor that allowed a 24 hour time constant on the cap value in question. Set the supply for the rated cap voltage, and after 24 hours you should have NO voltage across the resistor and the cap should be at supply value. Discharge slowly with a resistor, not a jumper, and you're golden...

Hmm... There are 86,400 seconds in a 24hr period of time. So, for a 1µF to have a TC of 86,400s you would need a resistance of 86.4 BILLION ohms. But wait. It takes FIVE time constants for a cap to charge to approximately full supply voltage. Maybe y'all did some different.

[Roe]

...did the math and picked a resistor that allowed a 24 hour time constant on the cap value in question. Set the supply for the rated cap voltage, and after 24 hours you should have NO voltage across the resistor and the cap should be at supply value...

Would you please give an example? Or maybe just tell me if I got it right. I have a bag of old good quality electro caps that I'd like to use at some point if I can. They are 20uf rated at 500V. I am not trained in electronics and easily get out of my comfort zone with calculations. I tried to calculate, but have no frame of reference to know if I got it right. I arrived at a 259K for the resistor. My limited understanding suggests almost no current is flowing, so a 1/2 watt resistor would be more than adequate, but no harm and probably better to use a 5W sand block resistor? Many thanks.

Use a 100k resistor.
10v across the resistor: barely usable
5v: good
2v excellent

[sluckey]

... They are 20uf rated at 500V... 259K for the resistor.

20µF and 259K have a time constant of 5.18 seconds. Since it takes 5TC for a cap to charge to approx 100% of the supply voltage, your cap would be fully charged in 25.9 seconds.

[martin manning]

Use a 100k resistor.
10v across the resistor: barely usable
5v: good
2v excellent

The voltage across a 100k in series (indicating leakage current) should be a function of the rated capacitance and voltage, so you need to make that calculation.

F&T's data sheet https://www.tubesandmore.com/sites/defa ... -a_new.pdf indicates leakage after 5 min at rated voltage is 0.008 × C[μF] × U[V] + 4 μA, and after an hour at rated voltage 0.15 times that. So, for a 100µF 500V cap that would be 0.15 * 0.008 * 100 * 500 + 4, or 64 µA; 6.4V across a 100k. If the leakage is higher than that, you can try leaving the cap connected through the 100k and see if it comes down with time.

[Phil_S]

There are 86,400 seconds in a 24hr period of time. So, for a 1µF to have a TC of 86,400s you would need a resistance of 86.4 BILLION ohms. But wait. It takes FIVE time constants for a cap to charge to approximately full supply voltage. Maybe y'all did some different.

I was certain I'd gotten this wrong, which was the reason for asking. It is TC that I don't fully understand. I'll mull this over a bit and see if I can absorb it.
Thanks for all the answers.

[sluckey]

TC or time is really simple. Use the formula T = RC where T is in seconds, R is in ohms, and C is in Farads. So, for your example of 20µF and 259KΩ, TC = .000020F X 259000Ω = 5.18s.

Read this... https://www.electronics-tutorials.ws/rc/rc_1.html

[bepone]

essentially reforming electrolytics means charging them with some minimal current (in mili Amps ) to nominal voltage. so resistor should be in megaohms. then discharging them slowly and do again several times. it is possible slowly also on variac, even better.

[FUCHSAUDIO]

It's been a few years, but I know we left the caps on overnight. We were buying photoflash caps to use as filter and output coupling caps in output transformerless tube amps.
They were usually 330uf 350 volt I think, and we used a few hundred K. Maybe Ted wanted them to sit "at voltage" for awhile after they formed. It was in 1985 or so, so I could be remembering wrong.
When they failed, they would literally split their sides or pop their vents and toss out some acrid nasty stuff, that's for sure....

[martin manning]

It's been a few years, but I know we left the caps on overnight. ... They were usually 330uf 350 volt I think, and we used a few hundred K. Maybe Ted wanted them to sit "at voltage" for awhile after they formed.

That makes sense. The leakage current should drop significantly in the first hour, but it might take longer if the caps are old, and it may continue to decrease slowly with more time as the plates are reformed. If after 24 hr it isn't acceptably low, they should probably be tossed.

[Jered]

Great stuff, thanks guys