Leakage in capacitors?

[Oddvar]

Hi.

Is there an "easy" way to measure leakage in caps?  I have some old Astrons that measure reasonably ok on the Cap Meter, but I feel they might miss in performance? Allthough they work?

[trobbins]

Insulation resistance meter - often referred to as a Megger tester - usually tests at 100, 250, 500 and 1kVDC.  Often very cheap on eBay for 10-20 year old devices.

[PRR]

> an "easy" way to measure leakage in caps? 

Why do you care?

Power filter caps, leakage doesn't matter until it drops the supply voltage or warms-up the cap.

The main place we care is when we have a plate at 200V, cap-coupled to a grid, and we want the grid to be "zero" volts (or some defined bias).

Then the obvious "easy" test is to see if those grids are at "zero" volts (or whatever bias desired).

"Zero" really means "very small compared to desired bias". We usually want 1V-2V of cathode bias on our 12AX7. 1V of leakage at the grid is a big upset. 0.1V is tolerable but we usually expect less, suspect that cap. 0.050V or less is fine, the cap isn't leaking much.

Note that after turn-on, when the whole amp jumps up from nothing to 400V and then settles into normal biasing, the leakage will appear high until the grid resistors bleed the caps toward final voltage. For most guitar amp coupling, 1 second is 100 time constants. Since you can't hardly measure a voltage in a second, most C-R nets will be nearly final-state by the time you measure. A few coupling networks may be slower. If it looks leaky, watch it for a minute and see if it is settling.

[Oddvar]

I have eight old Astron yellows .022 uf 400 volt.  When I use them in different amps, the sound seems to be weaker than other ones, Sozo blue ones f x.  My regular tech says they are useless and changed to others.  I was wondering what chances there are for these to go "bad", and if, what "bad" could mean?.

[shooter]

My regular tech says they are useless

If your tech is competent and you trust him.......

these to go "bad", and if, what "bad" could mean

You don't get desired results, I spent lots of years fixing broke things, I learned early on "why" is for the real engineers, I just wanted to go rock-climbing and move on to the next broke thing.  caps are cheap, acrylic paint is cheap, once your get a sweet sounding one, paint it blue

[Willabe]

I was wondering what chances there are for these to go "bad", and if, what "bad" could mean?.

Many if not all old film coupling caps can start to leak dcv after many years because of the materials they were made from. Many/most old classic tube radio guys don't even bother to test coupling caps to see if there leaking dcv, they just replace them automatically when restoring an old tube radio.

This dcv on the tubes grid the coupling cap feeds will throws off the tubes bias which throws off the tubes gain. The more dcv leakage the more gain it will lose, eventually to the point of no output.   

[sluckey]

Check them with your ohm meter. If they show any resistance other than infinity (open circuit) they are leaky. Toss them.

Or, connect one lead to a B+ node in an amp. Measure voltage between chassis and the free end of the cap. Anything other than zero volts means they are leaky. Toss them.

Your tech has already told you they are useless and you have heard a weaker sound with your own ears. Toss them.

But, if you just can't part with them, hollow them out and stick a Mallory 150 inside.

[Oddvar]

Thanks...

[uki]

But, if you just can't part with them, hollow them out and stick a Mallory 150 inside.

What the eyes can't see the heart wont feel !!

[mresistor]

http://www.justradios.com/captips.html

[uki]

Or, connect one lead to a B+ node in an amp. Measure voltage between chassis and the free end of the cap. Anything other than zero volts means they are leaky. Toss them.

What to read DC or AC ?

[sluckey]

DC