using a trimpot on clipping diodes in the op amp feedback path

[mordechai]

I've seen the neat trick where a trimpot is used along the path of diodes-to-ground to adjust the degree to which the diodes are able to clip the signal.  Is this possible in a Tubescreamer type feedback-path diode scenario?  That is, could the diodes be placed in the feedback loop of the IC, but with one endpoint connected to lug 1 of a trimpot, with lug 2/3 of the trimpot tied to the other rail of the feedback loop?  I'm wondering if this might allow one to dial in the degree of soft clipping created by these diodes within the feedback loop akin to how a trimpot would do the same for diodes-to-ground.

[Bill Mountain]

Yes it would work.  Just be wary of having too much resistance and the opamp clips instead of the diodes.

[Johan]

Do you mean in series with the diodes? If so, yes it works. It becomes sort of a clean blend rather than a soft clip thou...cool thing if you use your ts for boosting.
J

[Mark Hammer]

Here is the schematic for the Boss BF-2 flanger.  What you see in the 2nd op-amp stage is a pair of diodes in series with a 4k7 resistor.  All of that is in parallel with a 47k feedback resistor.  The R11/D3/D4 combination provides "soft clipping" so that when the Res control is cranked up, the sum of the feedback signal plus the input signal will not exceed the limits of the MN3007.   You'll see a similar arrangement in a variety of pedals, whose models escape me at the moment, for pretty much the same reasons.

The specific amount of resistance to use will depend on a host of factors, including how obvious you want the clipping to be, what the gain of the op-amp stage is in the absence of the diodes, what the feedback resistance, and any input or ground leg resistance is, and of course the forward voltage of the diode combination used.

FWIW, I sold a TS-808 clone to one of the region's better blues players (and winner of many different blues awards) that included a variable resistance in series with one of the diodes, to introduce varying amounts of asymmetry/symmetry in clipping.  Very effective.  I think I used a 10k pot.

[mordechai]

Thanks everyone for your comments, and Mark, that example you provided is exactly what I'm going for.  I put some LEDs in my Tubescreamer project, and they sound very good but are a WEE bit too crunchy in some circumstances.  Though they sound great for rock rhythm playing, I'd like to be able to tame them when I do something a bit more bluesy, so having them in series with a pot that I can adjust seemed like a good way to add some flexibility into the circuit.  I happen to have a 10KB lying around so I'll give that a try. 

The one thing that I guess seems like a wildcard here, though, is the smoothing cap.  I have a 47pF cap in parallel with the diodes right now.  Will adding in series resistance change how the 47pF will affect the clipping?  If I reduce the clipping but keep the cap the same, I suspect the softening effect will become much more pronounced, maybe more than I would like.  So is the key, then, to lower the cap value and perhaps also lower the gain resistor/cap network from 4.7K/47n to something like 3.3K/68n?