Why is my Ring Mod squealing like a pig?

[jpiddy118]

So I put together an Anderton balanced ring mod from a member's pcb: http://www.diystompboxes.com/smfforum/index.php?topic=92057.0, based on the original schematic with a modification  to run off one 9v battery. Worked and sounded great initially, but something went wrong between the bench and the box... it now oscillates constantly, as if it was being fed a constant input signal. Everything else works fine... clean guitar signal still passes, controls all work, Freq control changes signal accordingly... but i can't turn the d@mn tone off. Swapped 565 chips, no difference. Checked solder joints, ground scheme, etc., can't find the problem.

[glops]

I built one of these from one of his PCB's.  Did you etch your own board or use one of his? Or Perf? Got pics or answers?  I was JUST playing mine. I love this effect and would be glad to help... 

[jpiddy118]

I used one of Scott's pcbs, been in touch with him too and he's great... this is pretty baffling, since it did work & sound great initially. I'd like to know what could cause this circuit to go into self-oscillation like this, otherwise i'm not sure where to start. Signal gets to the chip, power gets to the chip, buffered clean signal gets to output, loud wild self-oscillation also gets to output.

Btw here's the original schematic:

[jpiddy118]

I don't have a scope, but all my voltages look good, here's what i
got, all with a fresh battery:

TL7660 pin 5  -8.73
LM565 pin 1  -6.01v
LM565 pin 10  +5.69v

audio probing:
TL072 pin 7 same as input
LM565 pin 2 same as input (thinking it should be clipped... can't
figure out why it isn't)

here are a couple pics too:

[Mark Hammer]

The 7660 operates off a HF clock (though not THAT high).  Gotta wonder if the 7660 and 565 are heterodyning.

Does the whining abate when the null control is set properly?

Just for the heck of it, consider sticking a back-to-back pair of Schottky or Germanium diodes between the wiper of the Ring level pot and the 4k7 resistor.  They will act as a very crude gate such that no ring-mod signal will be available to be mixed in with clean unless it exceeds the forward voltage of the diodes.  Ideally you want that voltage to be high enough that whining during the quiet bits is rejected, but sofly-played signal still sneaks through.  That's why I suggest diodes with the lowest forward voltage.

Normally, I'd avoid something like that, because the diodes will introduce some crossover distortion, and the gating will introduce a certain unnaturalness.  But since you have the clean signal to provide the continuity (i.e., always clean signal to fill the dead air), and since the ring-mod signal is not exactly undistorted, what the diodes do will be largely unnoticeable....with the exception of what happens to the whine.

[glops]

I need to take a look at mine today because I have a couple of extra of those boards and need to look at it. I'll take a better look at your pics then.  I couldn't see anything weird at my initial peek at your pictures but I had one simple thought.  Could your null trim pot be off? Maybe it got nudged when you were wiring? Just a thought...

If it worked fine when you had it hooked up to the bread board then maybe it has something to do with the off board wiring?

[jpiddy118]

When i first wired this guy up on the bench, the null trimpot did its job... this is a different tone altogether though. The trimpot does nothing to this one. This is much louder, and controllable by the Freq & Ring level controls, as if the 565 was being fed a constant, consistent input signal and this was the modulated output.

Mark, i'll try your fix when i have a chance, thanks. I'd still love to know what went wrong with this though, since it was functioning fine initially. What exactly is heterodyning? I'll google it in the meantime. This is my first experience with this sort of power supply, and my next project is similar.

[Mark Hammer]

Heterodyning is the essentially what ring modulation IS:  the combining of two AC signals in a manner that produces their sum and difference frequencies.

And the behaviour you describe suggests that the oscillator in the 7660 and the 565 are doing just that.

Just out of interest's sake, what exactly IS the frequency the 7660 operates at?  The datasheet indicates that the 7660 is likely somewhere in the neighbourhood of 10khz.  If your modulating frequency for the audio signal is, say 1khz, that will likely result in a fairly audible differences signal, and, given the properties of guitar amp speakers, a somewhat less audible sum signal.

I guess the easy test is to ask whether the whining noise goes down in frequency as you turn the modulating frequency pot higher.  If it does, then you're hearing the difference signal.

[jpiddy118]

The 7660 does operate at 10kHz, and it's inaudible with the 555 timer. Just for fun, i swapped out the 7660 with another i had on hand... no difference. I also removed the 555 entirely, which only added the expected 10kHz whine to the mix. Swapped the TL072 with another one and also an NE5532... no difference at all.

As for the Freq pot, the pitch of the oscillation follows it accordingly, just as it did when it was working correctly. One thing to note... the Frequency pots' lowest freq is dead center... turning it to the left or right from there raises the frequency of the oscillator. This was also the case when the effect was working properly... there was a difference in the sound of the modulation whether it was to the left or right though, not sure how to describe it technically. Just thought it was a quirky feature. Not sure if this is correct, or if i miswired something.

[Mark Hammer]

Regarding the symmetrical change in whine frequency, I'm not entirely sure how it works but there is the difference of A minus B and B minus A.  Perhaps the change in whine frequency occurs when you reach a point where the one becomes more audible than the other.

[glops]

I looked at my board.  Looks like your pot wires, 3 and 2, are switched from board to the pots.  This wouldn't matter if there are traces on the PCB connecting 3 and 2 if they are meant to be wired as variable resistors. I couldn't tell if there were traces on my board.  Also, I have my cap (underneath the board) soldered at different points than yours noted in the pic below. Not sure if this matters as there may be traces leading to the same points.  My cap is also soldered with the + going to the IC.

[jpiddy118]

Hmm, i followed wiring according to Scott's files... all my pots were working correctly at one point and still technically are, i.e. levels & frequency, proper direction, etc. As for the 10uF cap on the back, there are traces from my solder points to yours... followed a pic that Scott sent.

I've double-checked any questionable solder joints, wiring, grounding, components, you name it... what's driving me nuts is that the d@mn thing worked perfectly fine two days ago, and all I did since was pop it into an enclosure.

Btw, double-checked the pot wiring... the order on the pcb is different than the actual pot lugs, if that's what you're referring to... the pcb is labeled 2-3-1, with a wiper symbol for 2. Mine are all good.

[Mark Hammer]

Sometimes, when boxing something up, leads which had been kept at safe distance from each other when the board was sitting on the desktop , get brought dangerously close in the more confined quarters of the box.

Take it out of the box and spread out the wires and see if your problem goes away again.

[jpiddy118]

Tried that earlier, went back to the exact scenario that had it working the other day, same problem.

A few observations re: 565 chip... 

Pin 2 should be getting a clipped version of the input (buffered clean guitar). It's not actually clipped at all though, it's exactly the same as the input. Checked the soldering of those diodes, everything looks fine. 

Also, when i audio probe pin 1 or pin 10 (+/- voltages), i faintly hear the tone that i believe is being modulated & amplified further down in the chip (loudest at pins 4&5, slightly lower at 6&7). I'm not sure what the norm here is... i know the voltage getting to the chip is correct, but should that tone be there? The 7660/555 power supply is new to me.