Leaking film caps

[dynaman1]

What's an acceptable amount of DC leakage for a plate coupling cap? I've been checking by measuring DC voltage from ground to the "out" side of the cap. The few that I've tried all seem to hover around 200-500mV. Too much? Am I measuring incorrectly?

[phsyconoodler]

You have to lift the neg side out of the circuit to measure properly.

[tubeswell]

What's an acceptable amount of DC leakage for a plate coupling cap? I've been checking by measuring DC voltage from ground to the "out" side of the cap. The few that I've tried all seem to hover around 200-500mV. Too much? Am I measuring incorrectly?

+1 what phsyconoodler said about measuring the caps by lifting the ends. I prefer zero millivolts DC through the coupling caps if I can get it.  If you haven't lifted the grid-ends of the caps to measure them, then you might be detecting/measuring grid current on the following stage which is bleeding through the grid leak resistor.

[jojokeo]

Any leakage coming from coupling caps should not be allowed and replaced immediately as they will negatively affect the biasing of the following stage. They are cheap and easily replaced and will affect your tone and amplifier's proper operation. If in doubt, switch it out. Now, if you have a couple volts from a filter cap (after it's charged, then settles for your reading) that's a different story and will not affect your tone or biasing. If bad enough, you will hear ghost notes, wierd modulation, out of tune kind feeling w/ your guitar, and at worst noise and humm.

I use a simple DIY battery cap tester along w/ my VOM to check caps with when out of circuit that works well and is all you need - besides a little common sense to know what to look for. The cap gets charged and/or continues to charge but no measurable DC is measured across it. In cicuit is easy too when access isn't an issue but be CAREFUL to get a reading up stream and then on it's output to see that it's performing like it should.

[Fresh_Start]

Why do you have to lift the end of the coupling cap?  Can't you just pull the tube?

I know that sometimes you have to lift one end of a cap if there's a network like the tone stack.

Cheers,

Chip

[sluckey]

Why do you have to lift the end of the coupling cap?

Short answer, to get a true reading of the 'leaky' voltage. Many times the circuit following the coupling cap is supposed to have a dc voltage present. Examples would be grid of a fixed bias output tube, bootstrap biased split load PI or LTP PI, fixed bias preamp tube, etc. It would be impossible to tell for sure if the coupling cap were leaking unless it's leaking a bunch.

Even coupling caps that should have zero volts on the downstream side will probably have a resistor or network of resistors to ground. This resistance forms a voltage divider with the leaky cap (leaky means the cap has a dc resistance). This voltage divider makes it hard to tell exactly how much voltage is leaking thru the cap.

So, the best practical way to measure leaky voltage thru a cap is to disconnect the side that should have the lower voltage and measure the open circuit leakage with a high input impedance (10Meg or more) meter. Here again, the meter forms a voltage divider with the leaky cap. That's why a high input impedance is necessary. A 1KΩ/volt analog VOM would be useless here, except for a shorted (verrrrry leaky) cap.

[HotBluePlates]

You'd have to know the input impedance of your meter to figure how much leakage current there is.

0.001mA (1 microampere) of leakage in a 1MΩ input impedance meter is 1mV measured on the meter; with a 10MΩ meter, it's 10mV.

And depending on your meter, the input impedance may get bigger as you set the meter to measure lower voltages.

But you almost have to unsolder the cap because the meter's resistance is in parallel with the grid reference resistor of the next stage. Because leakage is leakage current and a measurable voltage is a byproduct, the meter's parallel resistance makes the measured leakage look smaller.

If you know you had a very high impedance meter (I've got one that will manage 200MΩ input impedance on these low ranges) you could measure without removing the cap. Or if you had a a very low-current meter (again, I've got one good down to 1pA) you could lift the leg of the cap and place the meter in series with the circuit to measure the leakage directly. That won't work with meters that only get down in the 10-100mA range.

So what to do? Your meter is probably a 10MΩ impedance (check specs to know for sure). If the bias voltage is very large (like 10's of volts in output tubes) or the effective grid resistance is small (like 100-220k in output tubes), relatively high leakage is still workable. If the leakage causes a 0.1v shift in a preamp stage (where bias might be as little as 1v), that could be significant.

But you're probably using film caps of some kind, and you'd like to see the extremely low leakage these should be exhibiting. Assume a preamp stage, 1v of bias and a 1MΩ grid resistor. 0.05v is maybe as much bias-shift you'd want to tolerate due to leakage, which represents 0.05uA of current. The same current across 10MΩ would measure as 500mV. You might call that too much and replace the cap.

It's really your call; severity of leakage depends on if and how it impacts the circuit. Some guys used to repairing and calibrating test gear tolerate no measurable leakage. But make sure to measure again with the new cap, to make sure you're not detecting some other issue.