are my coupling caps leaking?

[koen]

I've been trying to figure out if the two coupling caps before the power tubes in my 5E3 are leaking. When I just start up the amp, they both read about 0.05VDC (on the power tube side, relative to ground). However, the voltage slowly goes up (volumes at zero, not guitar plugged in) and after a few minutes they are about 0.25V, and climbing. I replaced both with two unused caps, but they same happens. The cap that is connect to pin8 of V2 increases faster than the one connected to pin6

Should I be worried about this, what could cause this, I thought they both need to read 0VDC. The other two coupling caps after V1 read 0VDC. Maybe another component is failing that could cause this?

[Alex R]

Those are low voltages. It is very rare for new caps to leak. Don't worry. Pull the power tubes out and see what happens, probably the voltage will go away.

...but just check that you do have those two 220K resistors going to ground from those nodes.

[koen]

Thanks Alex. I remember reading somewhere that any voltage over 1/4 V indicates leaking caps. Were they exaggerating?

I'll measure them again without the power tubes, and will make sure the connection between the two 220k resistors and ground (one of those under-the-eyeletboard-wires) is in place.

[Alex R]

Some caps do leak a little teensy bit, or appear to, and 0.25vDC is not very much to worry about. Trouble is, as soon as you set a limit, someone will be just over it and worrying, but if one were to say, for instance 1vDC, that might be a leaky cap and feeding a preamp stage it might just affect the stage's bias. Feeding the power tubes it's not going to affect a bias voltage around -40vDC too much, which is anyway adjustable, so as usual, it depends.

I think what you're looking at might be the grids rising up a little bit and putting a tiny voltage drop across the 220K grid ref resistors. Anyhow, it's not a problem! So try it without the tubes in and see for sure.

[koen]

I measured the voltage without the powertubes, and they were indeed zero. So nothing to worry about, or is one of the 220k and/or 1k5 resistors (on the socket) failing causing this small voltage? I measured the resistance between the 220k resistors and ground, and that was zero, so that's a good connection.

I actually didn't know that I could measure voltages without the power tubes in the socket (always though it could damage other parts). While I was at it, I measured some other voltages, and saw that the B+ was higher, about 450Vdc instead of 370 Vdc with the power tubes in place, is that expected?


Edit: with the power tubes back in, the B+ is still too high (400 Vdc) - what could have caused this, only thing I did in the last few days (it was fine then) is checking all connections, and reflowing some joints.

[koen]

One more observation, I measure about 3Vdc on the grid of the PI, so right after the 0.022 uF capacitor. Is that too much? When I touch the lead with the DMM probe at that site, I also clearly pick up a radio station. This is with the chassis on the bench.

[koen]

Ok, the battery of my DMM was getting low, which explains the high B+ I was measuring.

The grid of the PI still measures about 3 Vdc, though.

[loudthud]

Do you mean the half of the tube that is right before the power tubes? That tubes' grid will never read zero volts because the 1Meg grid leak resistor does not go to ground. You'll get a different reading depending on the meter you use. When you touch you meter's probe to the grid, it changes the operating point of the tube because of your meter's input impedance to ground. You can measure the cathode voltage to ground, then measure from the cathode (black lead) to the grid and you'll see about 1 volt negative.

Apparently you are close to a powerfull radio station. My shop is about a mile from a 50KW AM radio station and I pick it up any time I touch a meter probe to a live audio circuit. There is enough signal that sometimes it creates a small DC error of a couple of tenths of a volt. Interestingly, there is a TV station only about 300 yards from the shop that doesn't seem to affect anything. It's 5 Megawatts at about 500MHz.

[koen]

Thanks for the explanation, both on the grid voltage and the radio reception. Indeed there is a (AM) radio station close by, no idea how strong though.

[koen]

Ok, one more question (I hope) ...

Just played the amp and it just sounds great. I then did a final check using a chopstick while the amp was still open. I noticed that when I tapped on the coupling caps between V1 and the pots hat I can hear a clear 'thumb' through the speaker, which gets louder when I open the volume pots. These are the only two components that do this. Neither do the solderjoints that these caps are connected to. The tubes are quiet as well.

Is this something to worry about or am I getting paranoid?

[loudthud]

Capacitors will act like microphones if the plates move while there is a charge on the cap. This is how a condensor microphone works. Some caps are better than others, but it's generally not enough to worry about unless your amp has tons of gain. Carbon composition resistors can also be microphonic, they act like carbon microphones simply changing resistance in response to sound.

[koen]

Thanks again for the explanation. What do you mean by 'if the plates move while there is a charge on the cap' ?

[loudthud]

Imagine you have a pair of metal plates separated by air. The plates form a capacitor and the air is the dielectric. If a voltage is applied to the plates, electrons tend to crowd together on the minus side and they will push some of the electrons away on the positive side. A small current flows until the electrons reach a state of equalibrium based on the forces that they feel (the voltage). If you now move the plates closer together, more electrons will be pushed from the positive side. If there was no charge on the capacitor, no current would flow in response to the movement of the plates.

[koen]

So which plates are physically moving in a tube amp causing the cap to be microphonic? I guess that's what's confusing to me.

[defaced]

The foil that makes up the cap under the (typically) epoxy or plastic wrapping is vibrating ever so slightly when you hit it with a chopstick. Because they're at the front end of the amp, that vibration and subsequent change in current gets amplified many many times and you hear it coming out of the speaker.