Vox Repair PCB in test

[R.G.]

I just finished populating and am now testing the bugs out of a complete repair/replacement PCB for the "big head" models of the Thomas Organ Vox line.

The intent is to completely replace the original circuit board and breakage-prone wiring harness with a board that uses components that are available now, easy to replace, and wiring that is simple and direct.

The normal and brilliant channels are working, and I'll dig in further in the next few days.

This is built on my earlier several-small-boards approach to fixing the Thomas Vox line, so the circuits have largely already been tested. What's new is the PCB that fits where the old one did and replacement all of the wiring harness. It bolts into any of the Beatle/Guardsman/Buckingham/Viscount preamp boxes. The signal circuits are from the V1143 Beatle, and if the host amp didn't have an effect function, just leaving off the enabling wires to the footpedal socket makes it work in that model.

I think. It's still being tested.

[52 Bill]

This is built on my earlier several-small-boards approach to fixing the Thomas Vox line, so the circuits have largely already been tested.

I can understand the reason for this new board as it fills a need for some things, but have you stopped working on the individual board approach? I thought that the removal of some or all of the wires was a big plus.

[R.G.]

No, I haven't stopped the many-small-boards approach.

There's a complicated set of circumstances that set this off. Probably biggest is a Vox amp that really needed a new whole board. The schematics for the small-boards approach were already in the can, so what was needed was only some place and route. Well, OK, some document writing, too.

The two approaches are actually complementary. The small-boards approach is useful for correcting the remarkably silly "architecture" of the Thomas Vox amps, like putting the reverb assignably to only one or the other preamp channel, having no controls for the distortion, and some of the other things. It's useful for an architecture update.

Doing architectural redesigns of fifty-year-old amp designs is itself silly, except for me. The vast majority of people who have these amps will just want them to be "original" to some degree. Hence this.

As of now, all three preamps work, the mixer-limiter works, and I got one wire wrong on the tremolo modulator circuit. Fortunately, it's an easy fix; fortunately for me, because I bet a couple of hundred dollars making PCBs that I'd be able to make this work. The minimum-quantity price hump for a few prototype boards is a big one.

[R.G.]

Reverb driver and return amplifier work, at least for "fake" signals. Haven't hooked up a tank yet, but I'm guessing that will be unexciting since it really is a re-do of a design that was manufactured many times. If it passes signal at all and doesn't oscillate, it probably works just like new. Well, I can hope can't I?

Tremolo oscillates and modulates, and turns on and off from the footpedal switch. Oh, reverb switches on and off from the footpedal line too. I'm down to getting the repeat percussion running and testing the on-PCB signal switching. When that's done, there's not much left but to put it in a head, hook up a tank, and try to run it.

Encountered defects are (1) wire from a resistor went to the wrong side of a cap; I put it that way in the schematic (2) had to change three biasing resistors, one for each of the input JFETs because I can't get 2N4303s any more and used 2N5485's (3) Had to change one power supply dropping resistor to reduce the current into one section of the tremolo input amp.

The main +24V regulator runs a bit hot to the touch. The magical laser-infrared thermometer from Harbor Freight says it's running about 122F in a 79F ambient workshop. If I were feeding it +31V and that happened, I'd probably leave it, but I'm running it from a bench supply at about 28V. The power dissipation could nearly double, and that would make it HOT. It's running with no heat sink on the TO-220 package, which is only good for about 2W in free air with no sink.

+24V current is right at 100ma, so it's dissipating 0.1A* (28-24) = 400mW. Not huge, but then anything worth doing is worth overdoing, right?

I suppose I could split the +24 loads and put in another 7824, but I'm lazy and that would need a new rev of the PCBs. I did put space for a small heat sink around the TO-220 7824, so I'll just attach a heat sink, which should dispose of the heat problem. The subsidiary 18V regulators don't even get warm.

Hit some funny things that explain certain bugs that I've heard of on Vox amps. If the input wire from a pot to a section of the mixer limiter opens, the mixer-limiter oscillates at about 1MHz. Only happens with an open lead from the pot. The pot resistance damps this down and it never raises its head. It's noisier than I'm used to seeing on a scope, an artifact of the very low signal levels Thomas used to run the effects before they hit the mixer-limiter. All the gain in this thing is mostly in the mixer, about 30db. So everthing ahead of it puts in noise, then it gets amplified.

[drewl]

You should use the design of the 7 series English hybrids, they'd probably sound better!

[R.G.]

Been there, done that.

I did a UK 7-series PCB a few years back. The 7-series and the Supreme/Conqueror/Defiant are actually very similar to the Thomas Vox circuits, in that the preamp, effects circuits, and effect switching are recognizably similar circuits. I found this odd until I read some about the collaboration between Thomas and Vox in the UK. Denney actually flew out to California and worked with the Thomas engineer(s) on the design of the solid state stuff early on in the Thomas Vox design. Even the power amp designs are remarkably similar

Be that as it may, people with a dead Thomas Vox amp just want it to work again like it used to. This board does that.

[drewl]

I read that article too.

I've mentioned before that I have run Thomas preamps into tube power amps like the English ones and it sounds great.
Especially with some Bogen pa's I have.

One of these days in going to build an amp like that.

[loudthud]

Did you substitute film caps for the non-polar electrolytic coupling caps? I did that on an amp I re-capped.

[R.G.]

Yes, I did, most places. I used film where it matters for frequency determination, like in the tone/volume stacks, but left them BP electro where they're just signal coupling.

One reason I didn't do all of them is that the radical reorganization of the wiring to be simpler should eliminate the issues with flexing the board up to replace components, so they're replaceable without endangering the whole amp. Another is that - well, this is a replacement board and it can be replicated too.

For the original Vox boards, you're dead on target: if you fix it, fix it so you never have to go back in, or as close as you can get to that.

[R.G.]

Waiting on pots from Mouser.

The real ugly one is a 40K/100K asymmetrical dual for the speed control on the tremolo. I suspect it's also C/reverse log taper. Nasty if you have to replace that one. Fortunately, most people using something like this will already have one of them in the preamp chassis. But if it's busted, it gets hard to find.

I have a couple of possibilities I mention because you might run into something similar in other amps.

You can fake a series connected reverse log pot by using a 10X pot value with a 1.1 times the pot value you want as a fixed resistor in parallel with the wiper/cold. Just did the math. For instance, if you want a 100K reverse log pot set up as a series resistor, you can get that by hooking up a 1M linear with a 110K from cold to wiper, then connecting the cold and hot terminals as the variable resistor. This combination goes from about 100K with the shaft all the way CCW to nearly 0 with the shaft all the way CW, and does it in a close approximation of a reverse log pot.

In the case of the Vox amps, you'd need a 500K/1M asymmetrical dual linear pot, with 55K and 110K tapering resistors.

Right. Where am I gonna get 500K/1M asymmetrical duals?

Well, one way is with a 1M TRIPLE OR QUAD section pot. You parallel two 1M sections for a 500K and use the remaining 1M section by itself. This works for all kinds of pots if you can find a triple section. Works for single section funny-value pots if you parallel both sections of a dual too.

Another way is to use modular pots. There are outfits which will make up odd combinations of pot sections out of little 1/2" square modules and bolt them together. You can get up to quads with multiple switch sections. As you might guess, this can get expensive, but for a really important repair, it is at least possible. State Electronics and "potentiometers.com" do this. Don't know if they'll do one at a time, or what it costs, but the request for quote is in.

Another way is to ditch pots entirely and use an LED to two-or-more LDRs. You can set up photocell/LDRs to share a light source and crudely fashion a "dual pot" from the LDRs while using a single pot to control the current to the LED. I've done this, subbing out the dual pot in the LFO of a univibe clone. Works fine. Clumsy, but works.

[drewl]

Finally got a chance to read through your work.

Fantastic as always, thanks.