~450V AC across screen filter caps

[defaced]

I'm using the Peavey Deuce as the basis of an all tube power amp. So far things have gone ok, but after building a new power tube board and bias circuit, thing have been wonky. The output is significantly quieter than it should be and very distorted to boot. I think I've tracked the problem down to the power supply, but I'm honestly not sure if this is the source of the problem or not.

All of these reading were taken with a mains voltage of 125V AC and ~380V AC (this is very jumpy) from the secondary.

Part ___ DC Voltage _____ AC Voltage
C6 _____________________ 0.1
C7 _____________________ 0.1
C8 _____________________ 450
C9 _____________________ 450
IC10 ___________________ 775
500V ______ 490
490V ______ 480
300V ______ 350

The only thing I can think that could be causing the 300V DC to read 350V DC is that I have redesigned the bias circuit (the last schematic in: http://music-electronics-forum.com/s...ead.php?t=5415) and attached everything to Earth ground. I have no clue why I'm getting AC voltages across those caps.

[loudthud]

This is a misterious one. An oscilloscope would help here but we'll try to forge on blind. Is there a big AC voltage at the speaker terminals? I'm thinking the thing is oscillating. Try disconnecting the feedback. Just for a reality check, clip one side of your meter to the chassis and ohm every point that is supposed to be grounded. I noticed that on the schematic there are two different symbols used for ground. Check that they are connected.

[defaced]

I'm going to build a probe tomorrow and use some scope software to see what this looks like. From listening to what little hum I can hear from the speaker, it's 120 Hz hum, but I'll confirm that once I get a scope on it.

The feedback is already disconnected. I have tried to reconnect it and it does nothing. I'm going to deal with that battle later.

I'll check ground wires tomorrow as well.

[defaced]

I'm still fighting with these issues. I think the DC issue is from my isolating the bias circuit. Basically I think that the bias circuit was also pulling that point on the B+ down to 300V.

I thought I figured out the AC issue (something to do with the center two sockets of my output section), but that doesn't seem to hold true for all conditions. At idle, the problem is solved when those two sockets are run empty, but after playing, the problem re-appears.

I'm going to rebuild the power supply to suit the isolated bias supply and see what happens.

[Enzo]

What do you mean isolated bias supply? The only thing the original bias supply shares with the other circuits is ground. It has its own winding, not even a tap. To bias the tubes, that negative voltage has to reference to ground anyway.

What was wrong with the stock power supply?

What kind of meter are you using?

[defaced]

Attached is a schematic of how the bias supply and the power supply are wired today. These are connected in a star grounding scheme with chassis ground as opposed to the mixed grounding scheme that Peavey used.

In the initial start up of the amp a shorted wire put 500v across one of the screen filter caps. I was able to catch it before it did too much damage, but being that I know damage was done (the voltage drop across the two screen filter caps is not equal, but they both are in spec), I'd like to replace it (the amp has worked fine with this having happened, so I do not believe it's the root of the problem). I would just replace that one cap, but I know that things that have worked together for years tend to wear together too, so replacing one thing usually means a cascade of issues until everything is replaced. Being that filter caps are pretty inexpensive, and the amp has sat for who knows how long unplayed, I figured rebuilding the power supply couldn't be a bad thing. Also, if the bias supply didn't affect the power supply, then something is wrong in the power supply circuit because the DC voltages I'm getting are not what I was getting a couple of weeks ago (the changes I'm seeing are not a function of AC mains).

I'm using an inexpensive digital meter that I've used for many years: GB Instruments GDT-11.

[Enzo]

Before you blame the caps, check the dropping/equalizing resistors. If one of those opens or has a serious value shift, it will skew the voltages.

I asked about the meter, because some inexpensive meters are confused trying to read AC on a DC node. They cheat but just rectifying the AC right onto the DC scale, then a little internal math makes the reading right. But in these cases, the DC will also run through the rectifier. There is no way you have 700v on IC10.

And your readings don't make sense. How can you have 775 VAC on IC10 and have 350DC there at the same time? After all, the 300v node IS IC10.

Your 300v is high at 350v, but that is because you don't have all the preamp tubes in there or some other circuits that draw off that node, is that the case? I don't know what all you've changed. To drop the 190 volts across 44k of resistance, there must be 4ma or so flowing. You seem to have only 3ma.

SInce your readings seem to change with tubes in or out of some sockets, I'd be concerned for wiring errors, and I also would be real sure where all the grounds are. If your meter was referenced to some AC node instead of ground, that might explain the 700 volts.

Make sure your power supplies are tied to chassis and then take all meter readings referenced to chassis.

[defaced]

All of the dropping/equalizing resistors are between 250K and 270K. The two 22K resistors read 24-25K.

There may be no way to have 700v across IC10, but then why am I seeing what I'm seeing as some times (like when all of the power tubes are in their socket) and other times not (like when I remove the center two tubes, and the system hasn't been on long)? I don't know, and agree that the readings don't make since, but at the same time, there's no logical reason as to why the readings are there in the first place.

The 300v being high at 350v is not because of the absence of a load. The thing that gets me, is that when I get AC readings across the filter caps, the B+ for the plates and for the screens decrease by about 20v and the B+ for the preamp increases by about 30v.

All of the power supplies are tied to chassis ground and I am taking DC readings from that ground. When I take an AC reading, I'm taking it across the cap. In 4 our of 6 cases, that is chassis ground.

[Enzo]

Where are those other two cases grounded then? ANd what resistance is ther between those two grounds?

I didn't say the 300v had no load, I said it had insufficient load. Dropping from 4ma to 3ma would do it. The only thing that sets the voltage there is the drop across those 44k ohms.

At this point, I am not confident in your grounds. Not being nasty, I mean that there may be a lack of continuity in your grounding somewhere. Removing two tubes shouldn't affect anything other than a little B+ sag. I suspect you have some connection going open

[defaced]

Cap C6 and C8 do not connect directly to chassis ground; they connect to the + side of C7 and C9 (referring to the original Peavey schematic). Those are the two cases.

Your 300v is high at 350v, but that is because you don't have all the preamp tubes in there or some other circuits that draw off that node, is that the case?

...


I didn't say the 300v had no load, I said it had insufficient load.

Either way the difference is the same, and the answer is still the same, no, the load on that node has not changed from when I had the circuit working properly. I have three preamp tubes available to me, and two are in use in this circuit. Swapping them around does not alter the phenomena, or the sound of the output of the amp, so I don't suspect that to be the issue.

[defaced]

I went through yesterday night and checked every resistor and rechecked every ground in the entire circuit. I did not find anything of interest.