What is the mas allowable leakage from an electrolytic capacitor

[smackoj]

I am mostly building pedal type preamps and effects lately so the caps I'm using are small compared to power filter caps used in a tube amp. I recently picked up a 9v powered 'transistor/capacitor tester that measures ESR. I tested some of the smaller caps e.g. 2.2 4,7 10 uf at 35 or 50 volts and got ESR values from 0.23 ohm to 5.00 ohms.

Is there one max amount of leakage in capacitors that a guy would stay at or below? Or do different applications allow for an actual range of leakage that is acceptable? I'm only discussing SS effects and tube amps in this regard i.e. musical things that sound better without hiss and hum.

thanks, JD

[Willabe]

There are leakage #'s on the manufacturers spec sheet. You could see if your measurements line up with theirs?

I don't have a # to give you but this is something to consider;

I built this cap forming rig and did some experiments with it. 

http://www.electrojumble.org/reforming.htm

I used a variac and started with ~50dcv on the 450v cap in the rig. I would measure the leakage with the 1st dcv applied and write it down. Then after an hour or so I'd re-measure, write it down and then up the dcv again. I kept this up for 4 to 6 to 8 hours depending on the caps uF value and it's leakage #'s. These were all new caps.

Each time I upped the dcv, say another 50dcv, I'd remeasure and the leakage had gone way back up again, not all the way but still a lot, maybe 50% to 75%. It would jump the most at the lower dcv's and each time geting closer to the max dcv rating it would jump back up slightly less % wise. After a while you get a feel for when you need to up the dcv and when the caps leakage is as low as it will get.

I could get it down to a very low amount of leakage compared to where they started and the leakage chart on the link. I seem to remember that I got the leakage down to at least ~1/2 and maybe as little as 1/10th of what was listed on the chart?

But they won't stay that way as they sit unused, no dcv applied. Caps reform as you use them with the applied dcv.

So you'll get a different leakage measurement from any cap new or old that hasn't had a dcv applied to it in a while, the longer it has been the more leakage will probably be measured. (And older caps are drying out with age, heat and use so they will leak more and more as time goes on.)

Because of improvements in materials and manufacturing techniques, (most) new caps are made much better then the cap of the 50's, 60's, 70's, etc. So your way ahead of the game to start with. 

[smackoj]

I see what you mean. Just coming out of a pkg and testing is not really a completely accurate measurement. I think I will try to buy the 'better' name caps when I have the option.

thanks, jack

[Willabe]

I think on cap spec sheets they list the allowable leakage (>) as after 5 minutes with full dcv applied or some such length of time?   

[smackoj]

GOOD INFO THX

[HotBluePlates]

... I tested some of the smaller caps e.g. 2.2 4,7 10 uf at 35 or 50 volts and got ESR values from 0.23 ohm to 5.00 ohms.

Is there one max amount of leakage in capacitors that a guy would stay at or below? ...

No, you should measure some new caps to get a feel for it.  And the measured ESR of a "good cap" varies with capacitance and voltage rating.  See the chart below, which is glued to the face of a commercial ESR meter.

In general, more µF = less ESR, and higher rated voltage = higher ESR.

[PRR]

> leakage from an electrolytic capacitor
> tester that measures ESR.

Leakage and ESR are two DIFFERENT things. 

Take a perfect cap and replace the leads with wet spaghetti. It will be zero Leakage and high ESR.

Take perfect cap and blob wet spaghetti around both leads. It will have zero ESR and high Leakage.

Many cap functions are a lot like a water surge tank. Your town gives you variable water pressure. You add a surge-tank to even-out the variations. If your tank leaks, water is lost, and may upset something else (fine wood floor). If the pipe to the tank is small and crusty, water won't flow in/out easily, surges not controlled so well.

Real caps can have either fault. Or both. Like a surge-tank which can be leaky -OR- restricted, the two faults may have no relation.

[PRR]

> 2.2 4,7 10 uf at 35 or 50 volts and got ESR values from 0.23 ohm to 5.00 ohms.

Pick one, get both values. (Not clear here which uF goes to which Ohms.)

I happen to know that 10uFd is 1,000 Ohms reactance around 16Hz. So for almost any audio purpose, an added say 5 Ohms of ESR is no-difference. If this happens to be the 0.23r reading, it is even more who-cares.