Tales of unexpected hum sources: Thomas Vox Amps

[R.G.]

I just verified one source of persistent hum in the bigger Thomas Vox amps, and I thought it might be a useful bit of knowledge to others. It's the indicator bulbs.

The Thomas Vox amps are known for just humming a little. I thought I could just fix that while I was on my trail of making up full replacement PCBs and such for resuscitating them, but they always came out with a little hum that I could never get rid of, using all the usual magic tricks for hum. The hum was clearly power supply ripple, sharktooth shaped and at 120Hz, not 60Hz, but I could not find how it was getting in.

Then I happened to have a broken wire on the power on and standby bulb supply. Hum gone. How can that be? The bulbs are simply resistors, not active devices, and they pull a constant current. How could they generate hum?

It turns out it was because they're resistors. The 28V rated bulbs are fed -31Vdc through a dropping resistor, and share the preamp chassis' ground return to the power chassis. They are being fed raw -31V from the power supply, including ripple, and that current holds the ripple profile of the power supply voltage; Georg Ohm was right again, V = I*R, so the I they pulled had ripple on it just like the V feeding them. This ripple-containing current was mixed with the ground return of the preamp and duly put a few millivolts of ripple voltage across the "ground" wires referencing the preamps to the power amps, resulting in that amount of hum being there always.

There are several solutions to this, of course, the simplest but hardest to accomplish being using a separate ground return wire for the indicator bulbs, but possibly involving driving the bulbs with a constant current, which would also clear up this mess.

So the received wisdom is - of course resistors can cause ground ripple! They just have to be fed ripple-y voltage to do it.

[Mike Sulzer]

Would an LM317 and one resistor make a usable constant current source?

[R.G.]

Yes, it would. And you just published my "secret solution"!!

[nickb]

Yes, it would. And you just published my "secret solution"!!

Or replace with an LED + resistor at 1/10th of the current and so 1/10th of the hum.

[Mike Sulzer]

Or replace with an LED + resistor at 1/10th of the current and so 1/10th of the hum.

I was thinking about that and concluded that one LED at much reduced current would not give as much light. I think you would need multiple LEDs in series to give the same light intensity as one tenth the current, but perhaps this is wrong.

In any case, preserving the ambience of the original bulb might be part of the fun.

R.G.: Sorry about the secret! But you can alway hide it in plain sight on the PCB you are designing and no one will be the wiser.

[Leo_Gnardo]

I was thinking about that and concluded that one LED at much reduced current would not give as much light. I think you would need multiple LEDs in series to give the same light intensity as one tenth the current, but perhaps this is wrong.

There are a couple manufacturers that offer "super bright" LEDs in a number of colors. Very efficient, with only a couple milliamp one will drill a hole thru my retinas. The hassles I find with swapping in LED's are 1: narrow viewing angle, you have to be in line with the axis of the lamp give or take 20 degrees or so, and B) a pilot jewel can mask the LED if the color isn't a good match. Nonetheless swapping in a super bright LED can be a good option, certainly worth a try. And they're not dead cheap either like regular LEDs. Generally about a buck apiece from Mouser in low quantities. I'm sure other vendors have them too.

[R.G.]

Yeah, pretty much anything that reduces the total indicator current or makes the current more constant would do it.

I've suggested LEDs for indicators to Vox users for years, but ran into the same oddity that I got when first recommending MOSFETs in tube amps for certain jobs - there is a certain set of users that feel that anything "not original" won't sound the same, or in the case of indicators won't look the same, or look "original" or "vintage". Oh, well.

In this case, having much lower hum really does make the amp sound not original, because correctly functioning original ones do hum. People are funny.

One oddity I still have to think about is this. I can make the filament bulbs have less current variation in the bulb by putting a BFC across the bulb. I think that would make the problem worse, as the AC ripple would see only the limiting resistor for an impedance, so you'd have more ripple current, not less, on the return line. First impulses are not always right. A series inductor in the indicator supply line would cut the ripple current a lot, but then you'd have to use an inductor, and that is neither small, cheap, or pleasant.

All things considered, an LM317 and a programming resistor is a pretty cheap solution, all right.

[Mike Sulzer]

Yeah, pretty much anything that reduces the total indicator current or makes the current more constant would do it.

I've suggested LEDs for indicators to Vox users for years, but ran into the same oddity that I got when first recommending MOSFETs in tube amps for certain jobs - there is a certain set of users that feel that anything "not original" won't sound the same, or in the case of indicators won't look the same, or look "original" or "vintage". Oh, well.

In this case, having much lower hum really does make the amp sound not original, because correctly functioning original ones do hum. People are funny.

One oddity I still have to think about is this. I can make the filament bulbs have less current variation in the bulb by putting a BFC across the bulb. I think that would make the problem worse, as the AC ripple would see only the limiting resistor for an impedance, so you'd have more ripple current, not less, on the return line. First impulses are not always right. A series inductor in the indicator supply line would cut the ripple current a lot, but then you'd have to use an inductor, and that is neither small, cheap, or pleasant.

All things considered, an LM317 and a programming resistor is a pretty cheap solution, all right.

It would seem that the LED has another advantage. It needs a large series resistor, which would act as an OK current source, reducing the hum without adding any active circuit. This is in addition to any improvement from reducing the current.

[nickb]

Yeah, pretty much anything that reduces the total indicator current or makes the current more constant would do it.

I've suggested LEDs for indicators to Vox users for years, but ran into the same oddity that I got when first recommending MOSFETs in tube amps for certain jobs - there is a certain set of users that feel that anything "not original" won't sound the same, or in the case of indicators won't look the same, or look "original" or "vintage". Oh, well.

In this case, having much lower hum really does make the amp sound not original, because correctly functioning original ones do hum. People are funny.

One oddity I still have to think about is this. I can make the filament bulbs have less current variation in the bulb by putting a BFC across the bulb. I think that would make the problem worse, as the AC ripple would see only the limiting resistor for an impedance, so you'd have more ripple current, not less, on the return line. First impulses are not always right. A series inductor in the indicator supply line would cut the ripple current a lot, but then you'd have to use an inductor, and that is neither small, cheap, or pleasant.

All things considered, an LM317 and a programming resistor is a pretty cheap solution, all right.

I such situations I've used yellow LEDS with no complaints.

The other solution would be to re-route the return path so that is does not share the low level audio signal path. I take it there is a physical problem that prevents that?

[cgoepel]

I such situations I've used yellow LEDS with no complaints.

The other solution would be to re-route the return path so that is does not share the low level audio signal path. I take it there is a physical problem that prevents that?

Yes, there's a physical limitation to adding an additional ground. The pre amp and power amp sections are connected with a multi-pin connector. The multi-pin connector does not have additional pins to run the additional ground. I am currently working on two V1141 heads that show the same problem. I first ran across the hum issue when I converted the power cord from 2 wire to 3 wire. Any tips for solving this issue would be greatly appreciated.

[R.G.]

Hmmm. One of my postings didn't make it in here.
Re-creating from backup.

@Mike Sulzer: The idea that all impedances are relative is a decent approximation. In this case, the LED resistor is indeed significantly larger than the bulb's 'on' resistance. The bulbs are (IIRC) nominally 28V, 65ma types, and that gives a nominal effective resistance of 430 ohms at operating temps. Super bright LEDS are often used with resistors in the 4K range for pedal indicators. If you want them brighter, you need to juice that up to 1/2 or 1/3 the resistance, so in the range of 2K-ish. An LED's resistor is then about 5-10 times the resistance of the bulb, and correspondingly 5-10 times less ripple passed through. This is, of course, a crude approximation kind of argument I'm making here.

An active current source worthy of the name will have a source impedance of over a meg at low frequencies like 120Hz. I'd have to go look up how effective the LM317 is at this, but it's pretty good at power line frequencies. Again, roughly, call it 1M.

So the LM317 CCS would run to about a thousand times less ripple current than the bare bulb, and several hundred times better than an LED with a higher resistor; again, all at a crude approximations kind of level. The universe seems to conspire that any decent answers need real math.

@everyone else
It's probably best not to get too wrapped up in what is a quirk of the Thomas Vox amps; they're a backwater, albeit my backwater , and I mentioned this here just as a pointer to others that you may someday find a situation where a resistor, including an indicator light, can induce hum by coupling power supply ripple into a ground return.

@cgoepel
Yeah, that cable and its pin connections are the limitation I was mentioning. I've been digging through the Thomas Vox amp grounding scheme for a couple of years now, finding one issue after another.

There is usually a pin position open in the earlier versions, including the V1141, that can be used to carry the third-wire ground to the power chassis from the preamp. On some amps, this works fine. On others, it induces hum, which I find bizarre. I >think< this has to do with the odd way the grounding is set up in these amps, and I've been tracking it down.

The original setup floats the preamp chassis on the local DC ground brought in on the 9-pin cable, and then runs that ground, warts and all, over to the power amp input on that RCA cable. It makes the hum be common mode(ish) in the base amps. But hard-wiring safety ground to both chassis means the floating is now ineffective, so you get hum. Sometimes.

On the preamp replacement boards I did, I ran a local star on the entire preamp circuits, and brought independent grounds out for each preamp input jack and the RCA output. That let me put a 10 ohm resistor in series with the "ground" wire to the chassis on the RCA socket, and that improved things a lot. Many times, ensuring that the input jacks have no ground wires back to the PCB helps. It's tangled.

One thing you might consider is ditching the 9-pin cable connector for a 12-pin universal mate-n-lok cable connector. These are readily available and cheap, and the additional three positions will let you have space to run an indicator-lamps return wire and eliminate this particular hum.

In the Beatles, some of the hum is magnetically transmitted from the power trans being very near the bottom of the preamp shell. That's what the two metal plates are for, M-field shielding. It takes both of them, and still some gets through. There should be the steel plate and the thick aluminum plate on the power chassis between the power trans and the preamp chassis. When assembled, these two help force the M-field back towards the transformer.

[Mike Sulzer]

OK, I will try my hand at at a barely decent answer with no math. Figure 7 in the readily available TI pdf file on the 317 shows ripple rejection (on the output voltage) of at least 63 db, approaching 70 db under some conditions. (And by the way, figure 8 shows that it is almost as good out to 10 KHz) Since the device works by holding the voltage across the adjustment resistor very nearly constant, the current through the resistor to ground is held constant to a similar degree of accuracy. Thus the ripple rejection on the current through the resistor to ground must be essentially the same as that of the output voltage. So your estimate of 1000 appears to agree with this somewhat different way of looking at it.

[R.G.]

Kewl!! I'm always looking for different ways of being accidentally right! Doesn't happen all that often, so I like it when it does.

[cgoepel]

R.G. Thanks for the update. This must be truly a labor of love for you. I see your efforts to make preamp boards available, years in the making. I too have a soft spot for these amps. I wanted one in the worst way - 50 years ago and I still do today. Also, thanks for the multi-pin connector suggestion and the E-shield and M-shield plates.

When trying to explain why I kept having broken wires on a customer's V1141, I made up a rough estimate that I share here below:
Dear customer, "In attempting to carefully troubleshoot the distortion booster, interconnecting wires inside the amp broke off causing an open circuit, meaning no sound. After several hours of attempting find the open connection, the amp was put aside. I would attempt to repair this issue several times throughout last summer with limited success."

Wire Tally To/From Main circuit board for Vox V1141:

Control Type Qty # of Wires Subtotal wires

Rotary Control 11 3 33
Input Jack 6 2 12
Rocker Switch 2 4 8
Rotary Switch 1 12 12
Footswitch Jack 1 6 6
Capacitor Bank 7 2 14
Reverb 2 2 4
Relay 2 6 12
Factry adj pots 2 3 6
Power supply 1 7 7

Total wires 114

[Ampkat]

Found this to be the case in old Executone Amps also mainly the big 6550 PA monsters and the transistor ones also that were giving us grief in the hospital overhead systems. Went in and removed all the 24 volt lamp indicators and the hum went away !!