1970s Ampeg V4 Strange Input Stage Behavior

[dubulup]

16R dummy load built to mimic the players cab and initial testing complete.

TEST1
input: .333V,
-9dB, knob at the 10:30 mark,
output tubes PIN5=293V and PIN3=33V,
signal is a clean hair over 80Vpp,
speaker jack measures 29Vrms.

anymore volume clips positive and negative peaks and waveform doesn't amplify any further.

TEST2
input: .100V,
-9dB, vol. knob at the 12:00 mark,
output tubes PIN5~293V and PIN3~33V,
signal is a clean hair over 80Vpp,
speaker jack measures ~29Vrms.

anymore volume clips positive and negative peaks and waveform doesn't amplify any further; basically the same result as TEST1

TEST3
input: .015V,
-9dB, vol. knob MAX,
output tubes PIN5=100V and PIN3=13V,
signal is a clean 40Vpp,
speaker jack measures 10Vrms.

TEST4
input: .015V,
0dB, vol. knob MAX,
output tubes PIN5=243V and PIN3=31V,
signal is a clean 80Vpp,
speaker jack measures 24Vrms.

TEST5
input: .030V,
0dB, vol. knob 12:00,
output tubes PIN5=243V and PIN3=31V,
signal is a clean 80Vpp,
speaker jack measures 24Vrms.

MAXING out the channel volume knob on this test revealed signs that looked like an incorrect bias (attached picture), so I brought the bias down to schematic levels -62V and it made lower volume output signals look bad (transition notch), so I adjusted back to -50V which seems appropriate for the mods in this amp.
Other notes: B+ measures 515V and screens 440V.

The part that concerns me is...this clean output signal of 80Vpp has 1.3Vdc at the output jack; makes no sense. It is 0.0Vdc right up to just before 80Vpp and it ramps up rapidly, then falls apart at max ch. vol. Maxing the channel volume shows an unstable DC reading blinking between 0-5V, however the highest DC voltage measured was 4V. Do you think the clipped signal could appear like DC to the meter? Sort of like a PWM signal???

[Enzo]

If your output is a true sine wave, it will average out to zero volts. If your output is distorted in some way, it likely spends more time positive than negative by some amount. Your meter responds to that by averaging.

[Helmholtz]

No (output) transformer can transfer any average DCV.

Seems your DMM can't correctly average large AC in DCV mode.

If you care, check DC level with your scope by switching between AC and DC coupling.

[g1]

I think you reversed your pin 3 and 5 readings in post # 46. Pin 5 is grid.
Sounds like it is mostly doing it's thing if you still only have 2 power tubes installed. Still does not make sense that the B+ is so low, especially considering you only have 2 power tubes installed.
I think you had mentioned some mods with zeners, are they set up to drop the main B+ at node A of the supply, or in the line to the center tap of the OT primary?

[dubulup]

I think you reversed your pin 3 and 5 readings in post # 46. Pin 5 is grid.
Sounds like it is mostly doing it's thing if you still only have 2 power tubes installed. Still does not make sense that the B+ is so low, especially considering you only have 2 power tubes installed.
I think you had mentioned some mods with zeners, are they set up to drop the main B+ at node A of the supply, or in the line to the center tap of the OT primary?

You are correct, I reversed PIN3/5 (got the first one wrong and copied and pasted wrong throughout)...late at night trying to brain dump in bed. Yes, with two power tubes installed I am getting a clean 28.6Vrms output with a 0.042mA input at 3:00 ch. vol. (from 3:00 to MAX nothing changes). I've set up a 16ohm dummy load using the 8ohm switch (due to half the power tubes) so 28.6V^2/16R = ~50W.

I started thinking the PSU is holding it back...but power output seems great. Schematic shows 540V at the plates, I'm reading 530V with no signal, so <2% low...it does drop to 515V during load.

The Zeners are in-series with the screen (1st - pic attached); set up to drop B voltage to screens.

Also attached I increase vol to the point of output clipping and chased the signal to see where it distorted first and it shows up between the cathode follower and where the feedback network ties in to the second triode of the 12DW7 (pics attached; first pic is distorted C10 / V3-6 volume at 1:30; second pic is V3-7 grid).

I will measure bias current later this evening, I'm starting to think this could be aging power tubes...remember the bias is -50V; which is MAXED on the trim pot installed. Running more negative made the signal look worse so that is why I think that is a good bias for the screen/plate voltages.

[Helmholtz]

I will measure bias current later this evening, I'm starting to think this could be aging power tubes...remember the bias is -50V; which is MAXED on the trim pot installed. Running more negative made the signal look worse so that is why I think that is a good bias for the screen/plate voltages.

Good idea. Never bias for more than 70% plate dissipation. Typically 55% to 60% is fine.
Biasing for zero crossover notch at clipping is likely to cause overdissipation.

[Enzo]

No (output) transformer can transfer any average DCV.

Please enlighten me. Conditions I am assuming are grounded output winding, and asymmetrical waveform. My intuition says a waveform spending more time positive should average same. What am I missing? I am willing to be wrong, I just want to know why.

[Helmholtz]

Please enlighten me. Conditions I am assuming are grounded output winding, and asymmetrical waveform. My intuition says a waveform spending more time positive should average same. What am I missing? I am willing to be wrong, I just want to know why.

Transformer physics requires average DCV at secondaries to be zero. Means that positive and negative voltage half-cycles must have equal area.
With asymmetrical signal, the "center line" will shift accordingly.

[Enzo]

OK, I learn something every day.

[Helmholtz]

Same effect as when using a coupling cap.

[g1]

If the picture in post # 46 is power tube clipping, the tube providing the negative alternation seems much weaker than the other. I would swap them around and see if you now have less swing above the crossover notch rather than below.

Agee with Helmholtz that bias should be set to run cooler than 70%. In stock condition they are often running around 50% dissipation. 60% should be a maximum for this model.

[dubulup]

You are correct, one power tube measured a mere 18mA bias current and the other about 33mA. I ordered a new matched set, hopefully they show up in a few days.

An up coming gig is next week so hopefully after new power tubes, everything checks out so I can reassemble.

I’m starting to wonder if someone has been cranking up the bias as these tubes have been degrading…as I mentioned the trim pot was maxed out when I took possession of the amp.

much appreciated for all the conversation and suggestions…it has really helped me to keep moving through the stages.

[g1]

While you are in there, check the solder on all the board mounted tube sockets. They often need re-soldering.
And one more wrench to throw in the works. It's very rare, but sometimes the faulty condition was the before, not the after. It's conceivable that the preamp had some issue that made it weak, and all of a sudden the problem cleared. Like a dirty tube socket or bad solder joint. In those kind of situations, the customer is actually complaining about something working right because they are used to it being in defective condition. We had a similar weird one doing the rounds in the shop once because no one could find the problem. It was a stereo power amp and one side was quieter than the other. Everything checked out ok for that side. It turned out the fault was actually in the supposedly 'good' side, and it had a fault giving it extra gain.

[dubulup]

Thanks for the tips…I can see a lot of the sockets have been retouched, some other solder joints still have what I’m assuming is a factory red dot on them (maybe QC??); like the ultra-hi switch, and PT points on the board.

Off topic question: how do you remove the clips for the pilot lights?

[dubulup]

If the picture in post # 46 is power tube clipping, the tube providing the negative alternation seems much weaker than the other. I would swap them around and see if you now have less swing above the crossover notch rather than below.

Agee with Helmholtz that bias should be set to run cooler than 70%. In stock condition they are often running around 50% dissipation. 60% should be a maximum for this model.

Welcome back guys! new matched set of tubes installed. measuring 31/32mA with the bias pots still maxed out at -50V. Now that -50V bias doesn't seem so hot...actually quite cool. Dropping Zeners must really be restricting flow vs. factory set up.

The same distorted plot shown at output as in post46 https://music-electronics-forum.com/...064#post957064 (pic1). Swapped neg/pos sides and even from inside to outside with no change.

DOn't ask me how I wasted an hour or so assuming my 820k rising to 1 meg didn't matter. Check those two resistors for value.

going back to the PI circuit...I pulled R32 and R33. R32 measures a dead on 1M, however R33 (820k) measures 890k; almost 10% high. This could be the reason V4-7 grid looks like this(pic2)...maybe. V4-2 grid looked way better...from memory, I'll be checking again once I decide how to get 820k back in there.