Problem with vibrato in Silverface Twin Reverb

[bobshbob1]

Hello, I am looking at a friend's Twin Reverb that is having an issue with the vibrato/tremolo effect. Schematic attached.

The effect works very intermittently at the moment; looking into the "roach" you can see the bulb light up occasionally, not with any sort of regularity.

However, when I hook my meter to the leg of the roach that connects to the junction of the 10meg resistor and plate of V5B, set to measure volts, the tremolo starts working again, and you can see the little light flashing regularly.

I guess this is like grounding that point through the internal resistance of the meter, but I'm wondering why it makes the tremolo work? Does this sound like the optocoupler going bad? I'd like a second opinion or two before I spend $10 on a new one.

[Enzo]

Try resoldering the bulb leads then.

[bobshbob1]

I'll give that a try, however it also worked when I measured from pin 1 of V5. Same point from an electrical perspective, but a different physical location, which leads me to believe it's not a loose connection issue...

[g1]

Did you check the 10meg with 1 end lifted?
Try a different tube?

[bobshbob1]

Forgot to mention in my first post, it acts the same way with a different 12ax7.

The 10 meg measures fine with one leg lifted, and I touched up the related solder joints while I was measuring it. Still no change, only works consistently when referenced to ground through the meter. I'm thinking I should just go ahead and order a new optocoupler, or wire a 1-2 meg resistor in place of my meter...

[g1]

First you should determine if the oscillator is running and sending pulsing DC to the opto.

[Mick Bailey]

There are two checks I use when troubleshooting the optocoupler;

The first is to establish whether the neon side is functioning - lift one leg and clip a regular panel neon in place to see if it flashes. If not then there's usually a problem with the oscillator circuit (assuming the tube is good). The neon side can go blackened over time, with reduced light output, eventually not responding at all. You should be able to scope the grid of the triode and see the modulation voltage. A neon has a critical strike voltage and if this is borderline then your meter probe may be providing just enough of a path to ground to allow the neon to regularly strike. You may have insufficient voltage on one side of the neon (check the voltage to ground at the 100k junction), or the tube isn't conducting sufficiently - maybe a weak signal on pin 7 or a high cathode resistance.

The other check is to see if the resistor side is responding by measuring the resistance when the neon is pulsing. This is where an analogue meter is better than a DMM, unless you have a bar graph.

Gut feeling is you don't have a bad optocoupler. usually they work to some degree, or don't. I wouldn't replace it until you've established what the problem is. Neither would I strap a resistor in there - it gets it working but isn't a fix. I suppose there could be the possibility of the neon strike threshold increasing to the extent that it's just outside the design tolerance of the circuit.

[bobshbob1]

Thanks for the tips. Unfortunately I don't have a scope available to verify that the oscillator is working. Is there a way to check that without one? I can clearly see the bulb lighting up inside the heat shrink, and it lights regularly if I have my meter on the side connected to the 10 meg resistor/V5B plate. It still lights up, only very intermittently, if I do not (or if I have it connected to the 100k side of the bulb).

A neon has a critical strike voltage and if this is borderline then your meter probe may be providing just enough of a path to ground to allow the neon to regularly strike. You may have insufficient voltage on one side of the neon (check the voltage to ground at the 100k junction), or the tube isn't conducting sufficiently - maybe a weak signal on pin 7 or a high cathode resistance.

This is along the lines of what I was thinking, like the path to ground through the meter causes a larger-than-normal voltage drop across the bulb, allowing it to light, whereas the normal circuitry does not provide a big enough drop. Voltage to ground on the 100k leg is a constant 442V, and on the 10 meg side it jumps around between 280-ish and 330-ish, which I'm guessing is caused by the oscillator working.

Also, I've read that the neon bulb can wear out over time, especially in a DC circuit like this. It looks like the amp's owner used a shorting plug on the vibrato footswitch jack, and just turned the intensity to 0 when not using the effect. This would mean the bulb was being lit and dimmed constantly whenever the amp was on, which I can see causing it to, possibly, have an increased strike threshold, and thus not function properly in this circuit.

I think the resistor side of the opto is fine. I'll have to actually measure it to verify, but the volume definitely tracks with the neon bulb lighting, whether it's happening regularly or intermittently, so I'm thinking that's not the problem.

As a next step, I'd like to try Mick's suggestion of clipping in another neon, which could rule out either the bulb or the oscillator circuit depending on what happens. Would a typical 120V pilot lamp work for this test? or does it have to be something closer to what is actually in the optocoupler?

[g1]

I don't think your meter can make a bad or weak neon work. But it can make a stalled oscillator run.
Caps in oscillator are probably the best suspects.

[Mick Bailey]

Would a typical 120V pilot lamp work for this test? or does it have to be something closer to what is actually in the optocoupler?

It needs to be a neon pilot lamp complete with its series resistor. I've just used this method to check a Traynor YGL-3A.

[Leo_Gnardo]

Caps in oscillator are probably the best suspects.

Absolutely right. Old disc caps, feh. Swap in a trio of film caps & watch 'em wobble. Once the osc is running, you can sub a LED for your suspect neon. Read here and save us a lot of typing:

http://music-electronics-forum.com/t38440/

[bobshbob1]

Looks like the neon is it. I clipped another pilot lamp in its place and it is blinking away as it should. So the oscillator is working, just the old lamp doesn't want to light. I'm going to order an optocoupler and will be back if still having problems. Might replace the old caps as a precaution (both cathode bypass caps have been replaced at least once already) but it seems to work fine with them in place, so I'll probably just let it be. Thanks as usual for the guidance!

[Tom Phillips]

Looks like the neon is it. I clipped another pilot lamp in its place and it is blinking away as it should. So the oscillator is working, just the old lamp doesn't want to light. I'm going to order an optocoupler and will be back if still having problems...

Also consider that you can cut the heat shrink off the bug and just replace the neon bulb. Then re-heat shrink.

[g1]

I'm going to order an optocoupler and will be back if still having problems. Might replace the old caps as a precaution (both cathode bypass caps have been replaced at least once already) but it seems to work fine with them in place, so I'll probably just let it be.

Can you try just the opto first and report back? I'd be glad to eat my words ("I don't think your meter can make a bad or weak neon work") if I can learn something in the process!

[bobshbob1]

Can you try just the opto first and report back?

Sure, I will. I just ordered one so it will be a bit. Since it's not my amp I'm playing it safe and just getting a whole unit rather than just replacing the bulb.