Decoupling PI stage question

[Luthierwnc]

I just assembled and spark-tested a new head. It is a low-voltage, high-current amp built along the lines of a blonde bassman. My faulty calculations have the PI volts lower than I'd like. The power tube plates are 296VDC, rectified by a full-wave diode string and filtered by two 50uf caps. The screen stage runs through a choke with a 20uf cap.

The dropping stage for the PI goes through a 1k resistor and another cap to a typical Fender-style 82k/100k supply. On the 82K I get about 164VDC and on the 100K I get about 193. I'd like those to be in the low 200's but would a collective opinion on lowering the supply resistors and any problems that might come up.

I need to put in a little bit of NFB but other than that it is a pretty good sounding amp. I think I can improve the headroom and dynamics by getting the PI volts up. Thanks in advance for any comments, Skip

[LOUDthud]

Did you mean 164V across the 82K? Otherwise those voltages don't make sense. It would seem like there is only about 10V across the 1K decoupling resistor. That's not much to play with. I would think that the plate voltage on the PI is too low and would increase the biasing resistor, typically 470 ohms to something like 620 or 750.

[briane]

I'm just finishing up a (somewhat) similar build, with some similar problems.

I built a blackface AB763, with many dumble mods, and a dumbletor after the clean stage as a clean boost channel. My PT was a hammond 270dx, whose voltage is a tad low for this build (was a salvage, so near free). It is essentially a 100k plate dumble clean only amp, and is coming out wonderfully.

The key to getting my build to sound great was really in the bias circuit and the power supply (plus the coupling caps). Yes, there was not a lot of voltage to play with, but I got it very close to my goal. I have about 420 vdc available, and am running fixed bias at -57 volts. the drop is 420/413/336/264. I would like to have a tad more on the pre-amp, but I have been told the AB763 was a fairly brown amp, so I am waiting on that change till I am sure it is really needed. Its only been running for a week or so as is.

The amp is a bit brown sounding, but luckily that was part of my goal.

This is an extremely clean amp, with no distortion whatsoever, except when the player rides the strings hard, then it cuts in.

I found ayan's writeups on his 100k plate amp builds very helpful, though its funny, but I was going down the same lines as he did, and came to many of the same conclusions. You might also find these discussions helpful.

The best part of ayan's & dogears (and others input on the 100k plate) was their reflection was right in line with what I was seeing, and was after. It gave me that warm fuzzy feeling of being on the right path.

[Luthierwnc]

LOUDthud; Those were plate readings. All the tube draws are on spec. I am mostly concerned about the plate loads and bias points of the two PI triodes. If I reduce the plate resistors to raise those voltage readings, do I need to change the cathode and tail resistors to compensate?

BTW, I'm using a 12AT7 with an 820R K resistor and a 27K tail. It is still a little squirrely so I might put a 100R resistor at the tie point to ground and run a little NFB from the 4R jack into the inverter. sh

[LOUDthud]

The side with the 82K should have the highest voltage. Try another 12AT7 and/or check that there is not voltage across the grid leak resistors.