Drifting Bias Readings

[leadtower]

Hi All; Got a problem. On my ST 70 with the DIY board I am experiencing a drifting bias problem on one tube. Originally one channell died. I replaced the CCS transistor. Now upon power up; the amp sounds good. I put alligator clips on each bias test point; pins 8 of the EL 34's; then watch the readings on the meter. Both channels play. As a ten minute interval goes by I can see the left rear EL 34 begin to change voltage readings. I re-set the pot to 500 mv then null both power tubes to 0 and check again. over a period of about 2-3 minutes slowly the left rear tube begins to change again with the voltage increasing and the null point going from 0 to -50/60 mv. The sound from that speaker begins to get distorted. checking each tube for bias the rear tube drops to 440 while the front tube rises to 525-540. The right channel holds bias perfectly.
I have checked all voltages coming from the power supply and they are all on the mark. The 10 ohm and 1K resistors on the tube itself are correct. I swapped tubes but no help. I swapped driver tubes the 12AU7's; no help.
I'm baffled. Any test I can do that might help determine the problem anyone might know would be greatly appreciated.
Thanks Much. I also changed out the 12AX7 splitter tube. I re-set the bias after each change.

[lynxx]

Have you swapped the output tubes from left to right channel to see if the drift follows the tubes?

[leadtower]

Yes; swapped both power tubes but the drift stays put.

[TomMcNally]

I'd suspect a dirty tube socket or loose pins. Some of the newer tubes with "skinny pins" don't fit into the sockets tightly. Something to think about.

[leadtower]

You know the 12AU7 on that side of the board is a little shakey; I tried an older tube but I think the socket might be a little loose. It feels loose. Why I didn't change that out is because I got pretty good Volt readings on that tube. And why would it work initially then slowly go bad?
The power tube sockets are very good ceramic and very tight fitting. Just the 12AU7 socket feels loose.

[Shannon Parks]

1) For reference, is this a Rev A or Rev B diytube driver build?

2) Also, can you monitor the negative grid bias voltage? Maybe just set that balance pot dead center when the unit is off (use a DMM to measure equal resistance) and don't try to balance it.

3) Also, is your long tail voltage back in spec on the cathode since you replaced the CCS?

[leadtower]

This is the Rev A board.
Yes; I will monitor the -grid volts on test. Thats Pin 5 on the power tubes?
Pins 3 and 8 on the 12au7's look good on. Do you have a voltage test sheet BTW?
Thanks.

[Shannon Parks]

This is the Rev A board.

The board will be Rev A, but is it a Rev B build? For example, does it have 220pF caps at C3 & C4?

Yes; I will monitor the -grid volts on test. Thats Pin 5 on the power tubes?

You can also test at pin 6 - the tie-in - or at J1, J2, J22 & J23 on the board.

Do you have a voltage test sheet BTW?

Sure - checkout page 7 of the current manual.

Of course double check your voltages with the sheet and make sure a screen and plate tap aren't swapped on that side or some other wiring error.

Couple of easy tests:

1) Don't rebias and jumper the EL34 cathodes together with some alligator clips and see if they settle down.

2) You could do a test by swapping out the 1k grid stoppers at pins 5 & 6 and put in a larger value - maybe 10K-ish. If there was some freaky oscillation, that should squash it.

[leadtower]

Shannon; this is the modification of the Rev A is that what you mean by Rev B? It has the new caps and resistor. Regarding the oscillation remmember this amp was in service for a couple four months or so before the channel died. And the both channels work well for the first five or ten minutes of play. Then they begin to distort and the bias shifts. Does that mean anything?
Meanwhile I'm going over all the voltages.

[EWBrown]

My own SWAGS:

1: Swap the tubes between channels and see if this follows, then one more swap can determine the guilty tube.

2: Grid stoppers: not only a 1 to 10K for the G1 grid (standard ST70 is 1K, located on the octal sockets between pins 5 and 6), but another resistor, 100 to 1K , 1 or 2 watts, between the G2 grid and the UL tap on the OPTs, can eliminate any high frequency "sprogglies" (parasitic oscillations)which can cause drifting bias and distortion (been there, done that). Yellow_Light_Colorz_PDT_09

3: One last simple test is to pull out the 12AU7 PI tubes, and check / set the "idle" bias, and see if it drifts. (Without any signal input to the G1 grids, the tubes should be static current sinks) That would indicate an HF oscillation on the driver board, which is unlikely, but could occur, if a tube or component is going bad.

On most tube amp projects, I'll do the initial bias setting and check with just the (rectifier and) power tubes plugged in, as a start-up sanity test.

HTH

/ed B in NH

[leadtower]

Done some voltage readings and it seems on pins 1 of both the right and left 12au7's I get 130 volts while on pins 6 of the 12au7's I get 330 volts. Also On pins 8 where it calls for 13 v I get 25 and 33 respectively. The phase splitter tube looks good as well as the power tubes; all good voltages.
Thanks for the tips E.W.; I plan on adding the resistors to the grids and screen pins.

[Shannon Parks]

Done some voltage readings and it seems on pins 1 of both the right and left 12au7's I get 130 volts while on pins 6 of the 12au7's I get 330 volts. Also On pins 8 where it calls for 13 v I get 25 and 33 respectively. The phase splitter tube looks good as well as the power tubes; all good voltages.
Thanks for the tips E.W.; I plan on adding the resistors to the grids and screen pins.

Something very strange is going on. Could you measure these with no inputs or shorted inputs? Also you can try removing the negative feedback to see if the voltages go back in line (that would help narrow it down). Maybe email me some close up shots if you can. Double-check that the RCA ground and inputs aren't swapped (easy to do, really).

[leadtower]

I'll get right back to you.