Thomas Vox V1032 troubleshooting help?

[MWaldorf]

Whew, I thought that the NP caps would be ok, but then again, there's far more I don't know than know, so glad to hear it from an expert!

I plugged in a guitar and there's definitely distortion. Maybe not as bad as it was, but it's there at all volume levels.

I did notice a few other odd behaviors:

Turning up the reverb brings in a lot more distortion. At least there's a good amount of reverb, too! However, when I plug in the footpedal, things get weird. With the footswitch set to on, turning up the reverb actually decreases the overall volume until there's no volume at 5, and then gradually comes back as a wet signal, but quiet. With the footswitsch set to off, the dry signal gets quieter as I increase the reverb.

The tremolo works fine without the footpedal. As with the reverb, plugging in the footpedal makes it wonky. With the footswitch set to on, it behaves, but with the footswitch set to off, there's much less volume, and when the speed control is turned up past about 8, it gets into a kind of thumping oscillation.

The MRB seems to work as expected, though it's effect is much more pronounced now that it was when I first got the amp. That's a good thing in my book, but not sure if it's supposed to be a substantial volume boost or not.

Given that C35 helped, do you think adding R103 would make a difference?

[R.G.]

Whew, I thought that the NP caps would be ok, but then again, there's far more I don't know than know, so glad to hear it from an expert!

I plugged in a guitar and there's definitely distortion. Maybe not as bad as it was, but it's there at all volume levels.

I did notice a few other odd behaviors:

Just like an onion. You peel off layer after layer, and shed tears as you go.

Turning up the reverb brings in a lot more distortion. At least there's a good amount of reverb, too! However, when I plug in the footpedal, things get weird. With the footswitch set to on, turning up the reverb actually decreases the overall volume until there's no volume at 5, and then gradually comes back as a wet signal, but quiet. With the footswitsch set to off, the dry signal gets quieter as I increase the reverb.

The reverb on Thomas Vox amps is odd compared to most guitar amps. They buffer the dry signal and feed it to one side of a mix pot. They feed the other side of the mix pot from the reverb recovery amp, and take the mixed dry+wet off the wiper. That means that in fact, the dry signal does get quieter as you turn up the reverb. But the reverb signal is supposed to be about the same loudness as the dry to make up for it. Thomas Vox is one of the few amps that can be set to 100% wet reverb signal.

Did you pull out any temporary changes we did to the reverb circuit in digging out the first few layers?

The tremolo works fine without the footpedal. As with the reverb, plugging in the footpedal makes it wonky. With the footswitch set to on, it behaves, but with the footswitch set to off, there's much less volume, and when the speed control is turned up past about 8, it gets into a kind of thumping oscillation.

Don't know yet without some measurements, but I think that there may be a wiring bug on the footpedal plug or in the footswitch. Sounds like the tremo is feeding back into the reverb path.

The MRB seems to work as expected, though it's effect is much more pronounced now that it was when I first got the amp. That's a good thing in my book, but not sure if it's supposed to be a substantial volume boost or not.

What does the scope say about a sine signal's loudness with/without the MRB switch on with the sine in the middle of the MRB boost range?

Given that C35 helped, do you think adding R103 would make a difference?

Absolutely not. I think what's really wrong is that R103A is in there. R103 is the stock value that just bypasses the regulator to C35 in the old circuit. It's useless and worse unless you pull out the Vreg. Well, Ok, it might make a difference, but either very little or a bad difference.

At this point, scope your signal generator to see if it has those funny crossover notches. Then remove any other temporary circuit fixes we did along the way to the preamp and reverb. Take DC readings on the reverb transistors with no signal at the input, and with/without the footswitch engaged.

[MWaldorf]

I removed the test modifications as we went along, so the reverb is wired as it should be. Remember, the pan input and output impedance values are off - input 172r, output 183r - vs. the spec for 4FB2A1C of input 1475r, output 2250r. Given it's an order of magnitude off, could that be the issue?

Given that the reverb control is a mix between wet and dry, does it then make sense for the volume to drop when the reverb is switched off? The control would be something of a reverse volume in that circumstance, right? If so, that's an odd design, but I don't intend to have the reverb off ever anyhow.

I'll try shorting the tremolo footpedal plug and see if the behavior is the same.

I'll show my ignorance and ask, what frequency would be at the middle of the MRB? Would 1kH work?

Thanks as always for your help, R.G.

Mel

[R.G.]

I removed the test modifications as we went along, so the reverb is wired as it should be. Remember, the pan input and output impedance values are off - input 172r, output 183r - vs. the spec for 4FB2A1C of input 1475r, output 2250r. Given it's an order of magnitude off, could that be the issue?

Good on you for repairing the temporary stuff as we went.

On the reverb pan, are you making the distinction between the DC resistance of the reverb coils and the AC impedance of the coils? For instance, most output coils are 2250 ohms impedance, and only a couple of hundred ohms of resistance, Accutronics/Belton has tables of these. I can't remember right now, but the 183R output may be the DCR of the 2250 coil.

Given that the reverb control is a mix between wet and dry, does it then make sense for the volume to drop when the reverb is switched off? The control would be something of a reverse volume in that circumstance, right? If so, that's an odd design, but I don't intend to have the reverb off ever anyhow.

Hmmm. The more I look at this, the confusier it gets. The dry signal comes out of an emitter follower Q3/R16 at a quite low impedance. It drives the reverb control of 3K on the cold/CCW side. The reverb recovery transistor Q5 drives the hot/CW side of the reverb control from an impedance of 22K, but lowered by the feedback from collector to base down into the few-K region (at a guess), so I may be wrong about 100% reverb. What I was missing is that the reverb recovery side is NOT a voltage source (low impedance) and the dry signal is.

When the reverb footswitch is off (i.e. open) then reverb works as if the footswitch was not plugged in. When the reverb footswitch is on (i.e. closed), it attaches the collector supply voltage for Q5 to ground through 3.9K. That lowers the collector supply voltage to Q5 dramatically, as it's coming through R21, 150K. The voltage will be lower than 0.5V, effectively turning Q5 off by starving the base of bias. So the signal goes to zero, and the impedance at the collector gets "high", on the order of the 22K collector resistor.

The sudden change doesn't cause clicks and pops because C16 charges and discharges through resistances, either R21 or R25, and this slows down the DC voltage change on C17 to below its ability to pass the pop/change.

I'll show my ignorance and ask, what frequency would be at the middle of the MRB? Would 1kH work?

The best way to do this is to apply a signal, turn on MRB and watch the signal level at the CW/hot end of the volume control. Turn the frequency knob til this is biggest. That's the peak/middle of the MRB. I get 1.9kHz from calculation, but it sure seems like it ought to be more like 300-500Hz to me from some vaguely-remembered reading from long ago. Once you get the center, it's easy to operate the footswitch and see how much it changes.

[g1]

the pan input and output impedance values are off - input 172r, output 183r - vs. the spec for 4FB2A1C of input 1475r, output 2250r. Given it's an order of magnitude off, could that be the issue?

Your numbers are close enough for a 4FB type impedance tank:

[MWaldorf]

I didn't get the chance to test the MRB or tremolo this evening, but I did confirm that the DC resistance readings I got on the reverb pan match the pan in my Buckingham Reverb, with all its glorious drip, so clearly that's not the issue.

Also, thanks G1 for the tank impedance chart!

[MWaldorf]

Got some testing done.

First, the crossover notch is coming from the signal generator. Not sure why, or what I can do about it, but it's not the amp.

The MRB has the biggest effect in the 545-550H range. Using a 50mV signal, here's the reading at the Volume CW leg by frequency, first value being with MRB off, then MRB on

400H 12mV 25mV
500H 8mV 50mV
525H 8mV 62mV
535H 8mV 70mV
540H 8mV 75mV
545H 8mV 77mV
550H 8mV 78mV
575H 8mV 75mV
600H 9mV 65mV
625H 9mV 55mV
650H 9mV 50mV
750H 14mV 35mV
1kH 20mV 32mV
2kH 45mV 50mV

I measured the DC voltages on Q3, Q4, Q5 and Q6 with the footpedal in and the reverb switched on and off.

For Q3 the voltages were E 0.06V, B 0.66V, C 1.5V both with the reverb switched on and off
For Q4 the voltages were E 0.8V, B 1.5V, C 25.6V both with the reverb switched on and off
For Q6 the voltages were E 4.4V, B 4.95V, C 12.2V both with the reverb switched on and off
For Q5 the voltages where E 0.2mV, B 0.39V, C 0.41V with the reverb off and E 0.06V, B 0.61V, C 0.70V with the reverb on. With the footpedal disconnected, the voltages were the same as having the pedal connected and switch on.

Q3 and Q4 are at least in the vicinity of voltages you've got in the V1031 repair supplement. My Q5 collector reading is off (2.5-4.0V vs the 0.70V I get). Q6 is off across the board (E 1-2V, B 1.4-2.4V, C 15-16V vs my readings of E4.4V, B 4.95V, C 12.2V)

Finally, I ran a jumped from pin 1 to pin 6 of the footpedal socket, and the tremolo behavior is the same as with the footpedal connected and tremolo switched off.

You had mentioned that the problem might be the tremolo interacting with the reverb. I took DCV readings on Q11, and they don't match what you suggest in the V1031 repair supplement - where you've got E 3.8-3.9V, B 4.5V, C 27V, I get E 10.9V, B 11.6V, C 26.7V. The Vox schematic shows 11V on Q11 emitter, but I know other voltages they show are incorrect, so maybe here too?

[R.G.]

Got some testing done.
First, the crossover notch is coming from the signal generator. Not sure why, or what I can do about it, but it's not the amp.

OK. One more layer of the onion peeled.

The MRB has the biggest effect in the 545-550H range. Using a 50mV signal, here's the reading at the Volume CW leg by frequency, first value being with MRB off, then MRB on

OK. The change in level is big, OK, but 500Hz is a reasonable value.

I measured the DC voltages on Q3, Q4, Q5 and Q6 with the footpedal in and the reverb switched on and off.
[...]
You had mentioned that the problem might be the tremolo interacting with the reverb. I took DCV readings on Q11, and they don't match what you suggest in the V1031 repair supplement - where you've got E 3.8-3.9V, B 4.5V, C 27V, I get E 10.9V, B 11.6V, C 26.7V. The Vox schematic shows 11V on Q11 emitter, but I know other voltages they show are incorrect, so maybe here too?

I'll go do some thinking here. To get the nominal voltlages in the manual, I measured a live amp and did simulation runs on the "ideal" circuit from the schematic. What I need to do here is to mentally think about what could cause the voltages you're showing.

[R.G.]

After some thought and some simulation to try stuff out -

We may have hit another situation where transistors are too good. The failing of the simple bias circuit used on Q5 and Q6 is no longer used because it's much more device dependent than some more modern (c. 1970, that is) biasing schemes. Even heavy emitter degeneration doesn't produce results that are all that predictable.

For Q5's case, I changed the gain on the model transistor from about 100 to 2000 and watched the gain. Gain stays about 25-28db over the whole range, as you'd expect from the 22K collector resistor and the 680R emitter resistor. What changes is the collector bias point. A lot. The higher the gain for Q5, the lower the operating voltage at Q5's collector. This **may** be causing some distortion.

It's fixable by two different techniques. One is to swap out Q5 from a nice, low-noise, high-gain 2N5088 for a modest 2N3904. 3904s have gains in the 100-200 range these days, and are moderately quiet, so this is one possibility if you have a 2N3904 handy. Another is to change out R20 for a 2.2M to 3.3M value. The higher gain and quietness of the 2N5088 are kept and the bias point moved up because the DC bias to the base is changed a lot, but now you get the additional thermal noise of the higher bias resistor. Might be OK, might not. Another way to do this is to modify R20 into a T network instead of a single resistor. This can cut the "excess" base current from the collector-fed bias down, and might be a better and more device tolerant solution.

If I were you, I'd try replacing Q5 with a 3904 or similar low-gain device or changing R20 to 2.2M or so.

[MWaldorf]

I got a 2N3904 at Radio Shack and swapped it in at Q5 in place of the 2N5088. The distortion is still there, the amp output is much weaker, and the reverb starts to howl when it's turned up beyond 5. But other than that, it sounds great! The voltage readings on Q5 are now E 0.63V (unchanged) B 0.67V (previously 0.63V) C 0.96V (previously 0.70).

I'll try reinstalling the 2N5088 and swapping R20 tomorrow.

[R.G.]

Hmmm. Another perfectly good theory all shot to blazes by reality.

More thinking to follow.

Is there any chance that the reverb transformer primary or wires to it are open?

[MWaldorf]

I reinstalled the 2N5088 in Q5 and took measurements across the two sides of the reverb transformer. From the primary (W17) to the supply (W18) it reads 50.2 ohms. The reverb input pin to ground is 119 ohms. With the amp powered up, I get 25.3V on the primary and 25.9 on the supply side of the transformer. Also, not sure if it matters but with the output rca disconnected from the pan, the pan chassis reads 307 ohms to ground. It's 0 with the output rca connected.

I took a video of the current sonic output of a guitar through the amp. You'll hear the distortion on the high and low E, and how the amp starts to oscillate in low frequencies as the volume is turned up. Rolling off the bass control makes this go away. Also, the reverb goes into a squealing feedback when it's turned up past five.

https://www.youtube.com/watch?v=j0YFHP_kmf4

[R.G.]

Very useful.

Is it possible for you to re-do that video, but with the camera trained on a scope picture of the (a) speaker output and (b) volume control CW lug, or both at the same time if you have two probes?

The low frequency oscillation is a variant of motorboating, of course. That nearly always means that two stages are interacting through power and/or ground. Tube amps solve this with the resistor/cap nets to locally decouple the stages through the power supply. It us usually a much smaller problem in transistor circuits. But clearly not extinct.

After putting C35 back in, did you also leave R103A (68 ohms) in circuit?

[MWaldorf]

R103A is still in place. Should I have removed it?

I made a video of the signal and motorboating with the scope on speaker out and volume CW:
https://youtu.be/qQ3_yPegvSo

One thing I find curious - if others have built this amp without experiencing these problems, wouldn't something external to the PCB be the likely source? Either my wiring or external parts of the amp (transformers, power transistors, e-tuner)? We know the e-tuner cap shorting on the reverb transformer was part of the problem. Maybe I'm the problem here, or disastified minor dieties...

[R.G.]

R103A is still in place. Should I have removed it?

Actually, solder a wire across it. There's a whole string of thinking behind that, but the test is quick. One possibility is that even with C5 inserted, R103a is not only faking the voltage drop of the original resistor-only power, but also introducing issues with the voltage regulator. So short it. Let the regulator do its job and give up the voltage sag emulation.

It's a theory...

I made a video of the signal and motorboating with the scope on speaker out and volume CW:
https://youtu.be/qQ3_yPegvSo

That's actually very helpful. By freeze-framing it, I was able to see the start of motorboating before it got back to the top of the volume control. When the top of the volume control gets involved, it goes to full-on growling. I notice that there's a phase inversion that happens as well. The non-oscillating traces are in phase, the motorboating is out of phase. I don't yet know where that happens, but I will, eventually.

One thing I find curious - if others have built this amp without experiencing these problems, wouldn't something external to the PCB be the likely source? Either my wiring or external parts of the amp (transformers, power transistors, e-tuner)? We know the e-tuner cap shorting on the reverb transformer was part of the problem. Maybe I'm the problem here, or disastified minor dieties...

Me too. Especially since I built #1 and it just worked. I know that the Thomas Vox designs have oddities lurking down inside them. That's one thing the replacement boards were intended to pick up. Apparently there are quirks I haven't found yet.

For instance, mine didn't need that odd cap to ground from the second adopter's board.

I keep coming back to Mother Nature being ... well, a mother, and insisting that every rule all the way down to quantum physics be strictly followed. Clearly there is something I haven't found out yet.