NeoVibe motorboating

[tubeguru]

Hi,
I'm having a motorboating problem with  a neovibe that I built from a GGG PCB. The PCB is  a year or so old.
It's a stock build  with the using all recommended parts with the exception of a trim pot in place of R35 and R 36.
I used the project file PDF from GGG as an assembly guide when I built it and checked it against  the wiring diagram from here which is a much clearer print. The components are all correct on  on the board and the voltages on the transistors (2n3904) seem nominal. It passes audio through the oscillation and and the frequency of the oscillation varies with the LFO. I am powering it with an Electroharmonix 24VDC adapter. The B+  on the board is 14.9V. I checked and re-checked the components and the soldering and wiring all looks correct. Any thoughts? 

[R.G.]

That's kind of a new one.

All I can think of is that maybe a wiring or board mistake is leaking some LFO into the signal path.

Does it go exactly in sync with the light bulb pulsing, or is it a different frequency from the light bulb, but affected by the speed pot?

[tubeguru]

As nearly as I can tell it's in the signal path. As you turn the intensity up it's modulated by the LFO and sounds as though it's being phased. The whole board acts like a theramin and the frequency of the oscillation is changed by the proximity of my hand. I'm thinking maybe one of the transistors is bad and oscillating or a resistor is way out of tolerance and shorted. If that were the case though then wouldn't the voltages be off? I'm going to stick it on the scope tomorrow and try to track it down. If this were a tube amp I would look for a cold solder joint or lead dress issue and shunt it if all else failed. The solder joints all look good, I went over them with a magnifying glass and retouched most of them. The wiring follows the diagram. Maybe a loose terminal on a pot? Oh well, I was hoping perhaps it was a known..... I'll let you know what I find....

[R.G.]

That helps. Check to see whether the three-transistor preamp/splitter on the front end is oscillating. That setup can oscillate if the 330pF compensation cap is open, possibly for other reasons. A broken/shorted resistor or cap could do it as well.

If you can manipulate a scope, that will tell the tale quickly. Check the collector and emitter of Q3 first.

[tubeguru]

OK, I checked it out on the scope. Q3 seems to be the source of the oscillation. I replaced the transistor and measured the values of the other components. All seems to be in order. There was also no oscillation when Q3 was not in the board so Q3 is playing a part... maybe a bum wall wart?

[R.G.]

What's the power supply voltage doing when this is going on? Is it steady, or does it dip and wiggle too?

[tubeguru]

The power supply voltage is a rock steady 14.9V. I tried a couple of other 18v wall warts that are known to be good and the same story with the oscillation. Possible that signal from the LFO could be leaking in through the power supply? or that there might be a problem with the lamp generating RF? Could happen I suppose if there were some arcing in the filament. I know, I'm grasping at straws here.....

[R.G.]

Weird.

I'm suspicious that the three-transistor preamp is oscillating. The 330pF to ground from the collector of Q2/base of Q3 is supposed to keep it stable, but I've seen them oscillate if this isn't just right. You might want to pull it out, test it, clean up the traces, etc.

I actually like changing that out for a 30pF from collector of Q2 to base of Q2. This not only seems to work, the simulation says it's more stable.

Motorboating is classically either a gain-phase oscillation caused by feedback over multiple capacitor coupled stages or a relaxation oscillator setup.

If some of the components in the path leading back from the output of the phase line to the emitter of Q3 were wrong value or shorted, it might result in gain-phase oscillation. A missing component in a bias network might make a relaxation oscillator setup.

A lot of times I've had to result resort (!) to killing an oscillator to debug it. By that I mean that when you know all the parts which make it stop oscillating, you know where to look.

You might pull out the 1uF caps from Q3 collector and emitter to the rest of the circuit. Then use your scope to see if it still oscillates. If it does, the problem is localized to the preamp. If it doesn't still oscillate, the rest of the circuit has to help it oscillate somehow. We can go further from there.

Edit: RESORT, not result. spell checkers aren't a whole lot of help when you automagically type real, but incorrect words. 

[tubeguru]

The oscillation is present in Q1,Q2 and Q3. Its a fairly neat sine wave.tried the 33pf cap to no avail.

[R.G.]

The oscillation is present in Q1,Q2 and Q3. Its a fairly neat sine wave.tried the 33pf cap to no avail.

OK, so it's isolated inside the preamp. I would check the orientation, value and soldering on all the capacitors in that section; likewise, the orientation and pinout of all the transistors, and the value of the resistors.

Sine wave indicates gain-phase oscillation because relaxation oscillators are pretty much non-sine. I would suspect particularly the capacitors bypassing the emitter of Q2 and the split collector load of Q1.