Thomas Vox V1032 troubleshooting help?

[J M Fahey]

The output transistors are modern replacements - MJ15016G which are PNP.

Sorry but those are NOT replacements.
Original circuit calls for PNP *germanium* transistors, those are PNP *silicon* ones.

Bias will be absolutely insufficient and amp will distort badly, because of lots of crossover distortion.

There are no transistors or anything else for that matter on the heat sink (see photos).

Donīt understand how to read this ... I see 2 power transistors and the driver transformer mounted on the heat sink.

I looked at the datasheet for this transistor and it says that leg 1 is the base, leg 2 is the emitter and the case is the collector. However, I don't see anything about orientation to determine which is leg 1 and which is leg 2.

Notice emitter and base legs are not exactly halfway between both mounting holes but closer to one of them, so you know how to look at them:

shown seen from legs side..
IF you turn them around , they will NOT fit in the socket.

[R.G.]

@Mel:
I'll go ruminate on this for a while to see if I can come up with a simple way to get the wiring right. Good that you have an o'scope. It may be needed. This reinforces my suspicion that I need to get adopters to either label wires carefully before removing them from the original output stage or be able to figure out which was which after not labeling, so they can be put back where they belong.

@JM:
You're right, the originals were germanium. But the silicons were replacements based on my recommendations. I stopped chasing and pricing germanium replacements for Vox amps when the price got over US$80 a pair. Instead I figured out how to pop in PNP silicons and rebias it. I believe that's what Mel has done, based on my recommendations.

The new silicons need one resistor changed per device to bias it up and remove crossover.

The original Vox amps used a terminal strip on the heat sink plate to hold some power resistors for the biasing. With modern power resistors and PCB technology, I was able to absorb those bias resistors back onto the single PCB, removing the parts from the terminal strip. My question to Mel was a coded "did you remember to get rid of the old resistors?" and his reply was "yes, I removed any of that stuff and only the transistors are on the sink".

That also led to his question about which pin to hook up to which wire. The silicon PNPs drop right into the sockets for the older germaniums if you're just replacing the output trannies in an amp without a new PCB, but with the new PCB, you have to remove all the old wires and take new wires to the transistors, so it gets tricky.

What's really going on here is that my first several PCB testers were able to just go figure out stuff that I hadn't thought to include in the PCB manual, and it just worked. Mel is technically capable, but is missing a few steps that the first adopters just filled in on their own and didn't tell me ought to be in the instructions for other people.

I'm really happy this kind of review is getting done. My secret mission is to make available boards to keep the Thomas Vox amps out of dumpsters and landfills. The more people that can cope with them, the better.

[drewl]

Your excellent work RG has helped me keep TV amps out of the dumpster and I will recommend replacement boards to the owners of every one I get in for repair.

[MWaldorf]

R.G.,

Thanks for your patience and allowing me to be a tier 2 guinea pig. I took a number of photos before I started disconnecting things, so while I didn't directly label the wires, I know where they were before. I also have your V1031 Master Repair and Board Replacement documents which I followed and reviewed to make as much sense as I could. So I matched the wires from the heat sink based on these. Not saying I haven't made a mistake at this, but that was my intention.

The silicon output transistors were already in the amp when I got it, and there was nothing else on the heat sink. On page 8 of the Board Replacement doc, you say to change R32 and R34 to 13 ohms for silicon PNPs, which I did. However, I inspected the original board and I see that whoever replaced the output transistors used 20 ohms. My amp has MJ15016 transistors, which aren't on your list of suggested replacements (closest I see is MJ15004). Could it be that the MJ15016 requires the higher resistance that was in my amp?

FWIW, I'm learning a lot about troubleshooting here, which I greatly appreciate. Also may be handy in the future because the normal channel on my Buckingham has developed a bad noise problem!

JM, thanks for the info on the transistor orientation. That should have been obvious to me. Of course, typing transistors when I meant resistors in my last post is a demonstration of my attention to detail at the end of a long day!

Mel

[R.G.]

I took a number of photos before I started disconnecting things, so while I didn't directly label the wires, I know where they were before. I also have your V1031 Master Repair and Board Replacement documents which I followed and reviewed to make as much sense as I could. So I matched the wires from the heat sink based on these. Not saying I haven't made a mistake at this, but that was my intention.

And you may not have got them switched. It's just something to check. If disconnecting the output stage made the oscillation go away, then something about the output stage is wired in a way that makes it scream. I picked the transformer connections as something that's easy to confuse (well, OK, it was for me) that would explain this all by itself. I think it needs checking, especially since your note on color coding wires made me suspicious of another color-code issue.

The silicon output transistors were already in the amp when I got it, and there was nothing else on the heat sink. On page 8 of the Board Replacement doc, you say to change R32 and R34 to 13 ohms for silicon PNPs, which I did. However, I inspected the original board and I see that whoever replaced the output transistors used 20 ohms. My amp has MJ15016 transistors, which aren't on your list of suggested replacements (closest I see is MJ15004). Could it be that the MJ15016 requires the higher resistance that was in my amp?

The MJ15016 is a fine replacement. It's more expensive, but will work and you didn't have to pay for them! The fact that they were replaced may mean that someone before you got the output stage connections differently correct.

I would leave them at 13 ohms until it's proven you need otherwise. Get those resistors too big and the output stage will overheat.

I have a prelim write up on how to figure the phase of the driver trannie that I'll try to get posted. That will let you nail down the "dotted-ness" of your transformer for certain, and let us do a slam-dunk on the output stage connections.

[R.G.]

Transformer dot-ed-ness test
Finding out transformer winding polarity can be done with a signal generator and a DMM. It relies on the fact that if you power one winding, then "stack" a second winding on the first winding, the two AC voltages will add in both magnitude and polarity. The biggest voltage will be obtained when both windings are in phase. If they're out of phase, they will partly or completely cancel.

1. Label the leads or list their colors so you can write down results. Disconnect all of the transformer leads from the circuit if it's connected, making good notes and/or taking pictures so you can go back to what you had if you have to. But all leads must be disconnected for the following to work.

2. Use the DMM set to low-ohms resistance to separate out which winding is which. Write down which are pairs. For instance, many of the Thomas Vox drivers have red, blue, white, gray, orange and yellow leads. The resistance test shows (on the one transformer I tested here for writing this up, yours may be different) that red & blue, white & gray, and orange & yellow are winding ends, so there are three separate windings.

3. Set your DMM to read AC volts and connect it to one winding. Just pick one. Record on your list which leads you're using.

4. Set your signal generator to a frequency of between 500 and 1000Hz, turn its amplitude to zero, and connect it to the chosen measurement winding. Record on your list which wire is "ground" from the signal generator and which is signal. On my test transformer, I used the white/gray winding, and hooked signal to gray, "ground" to white.

5. Turn up the amplitude until you can read about 1V ac on your meter across this winding.

6. Pick another winding. Hook either lead (just pick one) to the "signal" lead of the previous winding. I hooked up yellow from the yellow/orange winding to gray.

7. Verify that you still read about 1Vac on the first winding ( mine was white/gray) and then move your DMM's positive/red lead to the unconnected end of the second winding. I moved the meter red lead from the junction of gray and yellow to the free orange lead.

8. Read the AC volts and record the results for these two windings "stacked" one on top of the other. I read about 0Vac.

9. Reverse the second "stacked on top" winding and measure again. I connected orange to gray, then read voltage on yellow. In this case I read about 2Vac with signal ground on white, signal generator on gray, and orange connected to gray, reading from yellow.

10. The boltage was biggest with ground to white, signal generator to gray, and orange connected to gray. To mark which was which, I drew up a schemo of the transformer, and placed a polarity dot on gray and yellow.

It happened that with this transformer, I was testing on the two equal-voltage windings first. To test the primary's polarity, I left the signal generator connected to white and gray and connected blue to gray, then reversed and connected red to gray. The AC voltage was bigger with red to gray.

From this, I marked blue with a dot, and I now know that gray goes positive compared to white, that at the same time yellow goes positive compared to orange, and blue goes positive compared to red.

No problem.

[MWaldorf]

OK, here's what I got.

My transformer is color coded differently than yours. I've got red, blue, green, green w/yellow dots, black, and black w/yellow dots.

The initial wiring was:
red - W33
blue - W32
green - Q9 base
green/yellow - W35
black - W38
black/yellow - Q8 base

I read the following resistance on the windings:
Red-Blue: 42 ohms
G-G/Y: 1.8 ohms
B-B/Y: 1.8 ohms

My test signal was 750Hz at 2V, and I got these readings:
Individual windings:
G-G/Y: 0.459 Vac
B-B/Y: 0.459 Vac
Red-Blue: 2.5 Vac

Black in combo with Green:
B-B/Y-G-G/Y: 0.923 Vac
B/Y-B-G-G/Y: 0 Vac

Red/Blue in combo with Green:
Red-Blue-G-G/Y: 2.5 Vac
Blue-Red-G-G/Y: 1.6 Vac

So for my transformer, the matched polarity seems to be Red-Blue, Green-Green/Yellow, Black-Black/Yellow.

Given this, I guessing the polarity of the green and black winds were reversed in my amp? I may be missing something on the schematic, but it looks like for both windings the yellow dot wire should be connecting to board or base, depending on which direction makes sense with the red-blue primary winding?

[R.G.]

My transformer is color coded differently than yours. I've got red, blue, green, green w/yellow dots, black, and black w/yellow dots.

The initial wiring was:

I read the following resistance on the windings:

My test signal was 750Hz at 2V, and I got these readings:
Individual windings:

Black in combo with Green:

Red/Blue in combo with Green:

So for my transformer, the matched polarity seems to be Red-Blue, Green-Green/Yellow, Black-Black/Yellow.

Given this, I guessing the polarity of the ... on which direction makes sense with the red-blue primary winding?

I just got to this. I'd like to go through it in some detail. Getting the two secondaries opposed to each other can kill the outputs or the power supply , So I'd like to be sure I'm right, and it's late here.

[R.G.]

I got some time to look at this, and compare the original versus the new schemo I drew, as well as your colors and numeric results.

I think the path to fixing it is to swap your red and blue leads and leave the black and green windings alone. That is, connect red to W32 and blue to W33.

Reversing which transistors black/black-yellow and green/green-yellow go to doesn't do anything.
Swapping black for black-yellow with no other changes probably melts down the output or the power supply.
Likewise, only swapping green/green-yellow probably kills something. So leave those two windings alone.
Swapping red for blue inverts the phase of the entire output stage, which ought to get you back to negative feedback not postive feedback, which is what I think was going on.

I ran into a Royal Guardsman that had all the right colors on leads, but had the polarity of the red/blue primary winding matched compared to several others I've looked at. Swapping the primary wires fixed it.

Let me know how this works.

[MWaldorf]

Judging from the insanely loud screeching coming out of the speaker as soon as I flipped the amp on, I'm guessing this wasn't the issue.

After, I checked to make sure nothing was shorting, and flipped the wires back as they had been. Same behavior as before.

Then, to be certain I hadn't messed anything up, I tried the swap again. Same screeching.

[R.G.]

OUCH!!! Sorry. That may not have been it.

But it sure looked like it from what I saw. OK, more thought is needed.

[MWaldorf]

Well, I had been thinking I needed a cup of coffee for a bit of a perk up. After that screeching, I found I was wide awake again!

[R.G.]

Sorry for the jolt.

How... exactly... is your speaker jack wired? Does the speaker return go back to the speaker return wire pad on the PCB, or through the chassis? Is there a speaker jack with its bushing connected to chassis?

Edit: and how about the input jacks - connected to chassis?

[MWaldorf]

I've been testing using the external speaker jack. The speaker and the external jack wires are connected on the tagboard to the right of the pcb, and from there to the pcb. However, the external speaker jack itself is mounted directly to the chassis with no insulation so it's grounded there as well. All the input jacks are also mounted this way.

[R.G.]

Here's the ghost I'm chasing.

You volume pot experiments showed that when the wiper of the volume pot was at some position above ground electrically, the signal from the preamp and the signal from [somewhere else] cancelled mostly. Moving it up or down increased the signal. So something is interfering with the "groundedness" of the volume pot.

My first guess was a miswiring - I do this all the time, so it's often my first guess - but that didn't pan out.

Second guess was based on the oscillation that you eventually got when volume control was turned high. That led me down the path of disconnecting things that use a lot of current that might be sending an overwhelming and cancelling signal back to the volume control.

In the Cambridge/Berkeley the normal signal passes through the reverb driver setup, so I thought it might have issues with the reverb. Disconnecting the reverb trannie didn't change things. I also suspected the output stage of oscillating at high gains, and wandered down the path of being sure it's not self oscillating. We checked that out - well, OK, you did, and have the ringing ears to prove it - and found that the output stage was working fine as is. So the output stage is working OK(ish) on its own.

But disconnecting the output stage instantly stopped the volume control funnies, as shown on the o'scope. So somehow the output stage is getting back at the input stage.

One way this can happen is if the high currents pulled by the output stage shares a conductor with the input ground. The common way for this to happen is with the jacks' bushings tied to the chassis, and the chassis conducting the speaker current, and wobbling the input jacks' ground around. This is actually what was happening on a Royal Guardsman I'm (slowly) working on. Eliminating chassis ground feedback stopped it.

Hence my questions about which jack is grounded where.

Ideally, the speaker output will have a speaker drive wire and a speaker return wire. The return wire would ideally be connected to the wire pad for speaker return on the PCB, from where it has its own trace to the main filter cap star ground. But attaching the speaker jack to the chassis means the chassis is "live" with any voltage that is dropped across that return wire, as the input circuits on the PCB get their own internal star ground wire on the PCB.

In this theoretical model, removing the external speaker jack return/ground bushing from the chassis, leaving only the speaker return wire to the PCB, would affect the funny volume control behavior.

Does it?