Ampeg SVT3 Pro Bias Setting

[Mars Amp Repair]

Hello all,
Does anyone recall if the bias setting on the SVT3pro is extremely touchy? It's been a while since I've worked on one of these & my notes don't indicate anything about this.
ton iThis particular amp, just moving the pot a minuscule amount from proper bias of about 20mv, results in the bias jumping to over 100mv and if you continued it would allow a setting of over 200mv (across one of the source resistors).
I'm going to check Q2 next, but wondered if this effect is normal for this model.
BTW, the pot tests very smooth with no jumps on the bar meter of my Fluke 87V, so it's not worn or noisy.
Thanx, glen

[Jazz P Bass]

What gate voltage do you read on Q3/ Q4 collectors?

[nevetslab]

Perhaps you should replace the bias pot if it's that touchy. I do recall seeing that behavior from time to time working on the SVTx-Pro series in adjusting source current. Though if there are replacement MosFET's in the batch that were replaced with unmatched parts, you could have that happening too.

[diydidi]

Biasing calls for an average of 25mV across all 0,47ohm resistors. Obviously the mosfets need to be closely matched to within 0,1V for Vgs.

[Mars Amp Repair]

I decided that the wide adjusting range of the 10K bias pot might be to accommodate the also wide range of Rd as the possible ratings of MosFets this amp can end up with. Might be complete bullshit, but that's my story...
So, I ended up replacing the 10K bias pot with a 5K and the setting is now spread over about 45deg of rotation instead of like 2deg! I think that a 2K trimmer might have actually been better for this particular situation.
I set bias for 20mv & drove output under load to get very hot. Upon no signal, the bias was still stable at 15mv proving tracking of Q2 on heatsink.
The amp has been out in the field for a couple months, so I guess that's some proof of 'My Theory', by Glen Whatley (Monty Python reference for those really into the vintage Flying Circus stuff (;- ).
Thanx for the replies.