DC on the input?

[hasserl]

I just fired up a new build, it is a variation on the 5E3X2 where I grafted a 6AU6 pentode into the preamp as the first gain stage for one of the channels (the chassis had an extra socket hole, so I decided to experiment a little). I'm getting serious scratching on the guitar vol/tone pots, and I found about .8vdc on the input, which seems to be coming from the heater circuit as I've got the voltage with the amp on standby. I pulled the tube and the voltage goes away but comes back when the tube is reinserted. I tried 3 different tubes and they all give the voltage on the input. Any ideas why? The grid connects to pin 1 and the filaments are on pins 3 & 4, I see no fuzzies or solder blots or anything connecting either of those pins.

I could use a cap to block the dc from the guitar, but this seems like a bandaid approach. I'd like to get rid of the dc if possible.

Note that the 12AX7 channel has no dc and no scratchiness on the guitar controls.

Besides this problem I need to do a little sorting out, but this is an interesting sounding amp, the tone and gain are very interesting. Once sorted I think I'm going to really like it.

Edit: wait a minute, filaments are ac, that couldn't be the source of the dc voltage on the input. But in Standby all the DC is cut off. hmmmm...

[Enzo]

The heaters make the cathode hot, B+ or not. When they get hot, they emit electrons - the space charge. Those electrons can accumulate on other elements in the tube, the tube can act like a rectifier even. In fact tubes ARE rectifiers, they just have extra elements in them to control the current with a signal. Without digging into it, I'd say that is why the heaters are creating this voltage, NOT because the 6VAC itself is going anywhere.

[hasserl]

I checked the B+ rail and I've got the same low voltage all across the rail when in standby. But I don't know how that voltage would get across to the grid of the 6AU6. It does not show up on the grid of the 12AX7 that serves as the input for the other channel.

[Don Moose]

Like Enzo said, the cathodes spew electrons, and they look for somewhere to go.

Do you have a DC path to GND from the grid (with nothing plugged in)? Maybe it needs to be a smaller resistor.

[hasserl]

Yes, the input is grounded thru the switched jack with nothing plugged in. With a cord inserted the ground is lifted, then there is the 1M grid to ground resistor.

So maybe the grid is just collecting the electrons from the filaments? In which case a cap would keep the DC off the input and I wouldn't get the scratchy guitar controls. I suppose a .1uf would be large enough to give me full frequency.

I'm still curious why the 6AU6 would have the DC on the grid but not the 12AX7. Is there something about those tubes that would make this condition more likely?

[hasserl]

Well, I've put a .1uf cap right at the grid of the tube and that is effective at blocking the DC, I get no more static on the guitar's controls. It still feels like a bandaid repair to me, and it doesn't look very nice either. I need to come up with a smaller low voltage cap that will fit a little nicer and not look so ungainly.

[Enzo]

It's not a band aid, it is the whole point of putting a cap there on many amp designs - DC isolation from the real world.

Right now you plug a guitar in there, but what if at some future time you plug an effect pedal in between the guitar and amp, and that pedal has an unterminated cap output? The input grid of your amp wil charge up to the voltage on the other side of that output cap. Your input cap blocks any such DC attack. It may be DC, but it can go both ways.

[cbarrow7625]

Sounds to me like you have a ground potential problem.

Where is the input jack shield terminated? Directly to the chassis? To a buss bar?

Where is the input grid resistor grounded?

Where is the tube cathode grounded?

My guess is that if you bring these three things together at the same exact ground point you may see your voltage go away.

Make sure you do not have excessive ground terminations throughout your circuit design (like the way Fender did it for so many years). Use a common signal ground bus & attach it to the chassis at one point only! Attach the shield wire from the cable directly to chassis ground as well.

It'shard to say without seeing the circuit & layout but I cannot think of another reason for DC to appear on the grid other than a ground potential difference between the cable shield and the tube signal grounds (since youu have already ruled out a bad tube).

Let us know what you find.

Chris

[hasserl]

Originally Posted by Enzo

It's not a band aid, it is the whole point of putting a cap there on many amp designs - DC isolation from the real world.

Right now you plug a guitar in there, but what if at some future time you plug an effect pedal in between the guitar and amp, and that pedal has an unterminated cap output? The input grid of your amp wil charge up to the voltage on the other side of that output cap. Your input cap blocks any such DC attack. It may be DC, but it can go both ways.

Good point Enzo, thanks.

[hasserl]

Originally Posted by cbarrow7625

Sounds to me like you have a ground potential problem.

Where is the input jack shield terminated? Directly to the chassis? To a buss bar?

Where is the input grid resistor grounded?

Where is the tube cathode grounded?

My guess is that if you bring these three things together at the same exact ground point you may see your voltage go away.

Make sure you do not have excessive ground terminations throughout your circuit design (like the way Fender did it for so many years). Use a common signal ground bus & attach it to the chassis at one point only! Attach the shield wire from the cable directly to chassis ground as well.

It'shard to say without seeing the circuit & layout but I cannot think of another reason for DC to appear on the grid other than a ground potential difference between the cable shield and the tube signal grounds (since youu have already ruled out a bad tube).

Let us know what you find.

Chris

Well, the amp is based on a 5E3 circuit, so I have followed standard 5E3 grounding schemes. I did not use a shielded input. The input grid resistor is grounded right at the jack, the jack is just grounded to the chassis via the body of the jack mounting to the brass plate. All the other ground connections are to the brass plate, as standard 5E3 grounding. The power tube cathodes are grounded thru the cathode resistors to the same PT stud as the PT secondary center tap. The filaments are grounded to the power tube cathodes.

[cbarrow7625]

I guess you could try running a wire from the ground lug on the jack directly to the copper buss where the first stage cathode resistor is grounded.

If the input jack shield is not making good contact with the chassis you could have jsut the situation I described.

Do you have the copper buss grounded directly to a lug on the chassis or are you just relying on friction between the buss & chassis to ground it for you? I would definitely connect a wire from the buss to a lug tightly screwed against the chassis.

When you duplicate a production amp you always have to remember that rarely did they ever do anything because there was some magic in it or because it worked better. It was almost always just the easiest, most repeatable, most economic way to do it when you have to make thousands of them. We have the opportunity when we make custom amps to wire things the way they ought to be....painstakingly thought out, well constructed & solid. I'm ranting now....

Chris