volume drop when negative feedback is doisconnected

[Robert M. Martinelli]

Alex,
As I and Enzo were saying, you most likely have a "wrong phase" problem ( PFB loop ), and I think it has been that way from the beginning, as Steve says. As long as you have to check the signal's phase only, and not its amplitude, you can "ignore" your scope's accuracy problems IMHO. You can either use the two channels of a scope, even if not perfectly calibrated, or a single channel scope taking the two readings separately. Your goal is to obtain a perfectly still image of the waveforms, it will be simple then to check whether the second waveform is "in phase" ( bad ) or not ( good ) with respect to the first one.

Also, as per Enzo' s advice, check the waveforms for the presence of oscillations ( in your case they're above the audible frequencies' range ), caused by the PFB. They're so nasty that, though we don't hear them, they can be of a severe hindrance to an amp, and a scope is the only way to physically "see" them.

HTH

Best regards

Bob

[Steve Conner]

I'd try switching the phase around, and see if the amp now behaves normally (volume high without NFB, lower with NFB, no weird squeals, hoots, or red plates)

If it doesn't behave, but acts up even worse, then the phase was probably right to start with and the original problem was something else: other posters have given you lots of leads to try here.

When using a scope to check NFB, I like to open the loop up and check that the amplitudes and phases at the two "loose ends" make sense. In the case of a series-feedback setup like the LTPI, not to mention almost every other valve power amp, the fed-back signal should have the same polarity as the drive signal (the LTPI takes the difference between them and amplifies that)

[Gaz]

strange PI voltages with feedback disconnected...

before trying my hand at troubleshooting with the scope, i thought i'd at least takes some voltages readings on the LTP PI with and without the feedback connected. here's what i got:

FB connected:

pin 1: 250 vdc
2: 33 vdc
3: 59 vdc
4: H
5: H
6: 254 vdc
7: 38 vdc
8: 59 vdc
9: H

FB disconnected:

1: 395 vdc
2: 333 vdc
3: 371 vdc
4: H
5: H
6: 395 vdc
7: 330 vdc
8: 372 vdc
9: H

so WTF is going on here?! why are the voltages shooting up to near B+ when the feedback is disconnected. something must be miswired... honestly i'm a bit relieved because i don't think i'm dealing with PB or oscillations.

thanks in advance for any suggestions.

-alex

[Gaz]

SOLVED

i quickly realized that something must be wrong with the grounding of the PI... sure enough, without the NFB connected the whole PI wasn't grounded. This is because I decided to switch to the 'scratchless' style presence control (ala JCM 800), forgetting to ground the PI elsewhere. with the feedback connected, the PI was grounded via the output jack ground, and without, just floating in space. thanks all for the suggestions.

-alex