Diode in Loop/Diode to Ground Pan

[WGTP]

http://aronnelson.com/gallery/WGTP/OPtion

I'm wonderng if this will work.  The idea is to use the 5K "Texture" pot to pan between the 2 standard diode distortion methods.  I have drawn this up before using a switch, but panning would be even cooler.  I have it drawn 2 different ways, but I think it is the same either way.  A larger pot may be needed.   

[Stompin Tom]

So, in the same circuit, what's the difference in sound between diodes in feedback loop and diodes to ground? I haven't really experimented with this, so I'm curious.

[dano12]

That looks very interesting. You could probably take it even further by having both diode pairs in the circuit, and add an opamp mixer at the output. That way you can control not only the "texture" of each clipper pair, but the mix of both pairs.

[Mark Hammer]

The basic principle to keep in mind with feedback loops is that all feedback is negative, and whatever conserves (as opposed to dumps) feedback will constrict or reduce gain.  Whatever dumps feedback to ground or limits how much feedback is applied will increase gain.

Without the diodes and texture pot, what you have is essentially a Rat gain stage, but with only one ground leg of 100R/1uf.  That gives a low-end rolloff at around 1.6khz, and a maximum gain of roughly x1000 with a high-end rolloff around 3.4khz.  With the gain control dimed, the 5k texture pot and 10uf cap will provide a gain of x21 for content below 1.6khz, since that bigger ground leg "conserves" more feedback signal.

The additional feedback path will apply more negative feedback, as determined by the forward voltage of the diodes, the ground leg of the texture pot and its interaction with the 10uf cap (e.g.,  if there is only 500R between the wiper and ground, the the rolloff would be at around 30hz).

What I'm wondering is whether the diode path (given the 1k resistor in series with it from the output) would provide any appreciable increase in clipping over much of the texture pot's range.  If you have the gain pot cranked, then the output signal will be well above the forward voltage point and settings of the texture pot won't matter much as any sort of "alternatue" to the other ground leg.  If the gain is turned down, then why should variations in the texture pot matter much when the signal feeding back through the diodes can take the 100R resistor as a path to ground, rather then the texture pot.

I'm not saying it won't work, OR that I fully understand and anticipate the behaviour of the circuit.  My gut feeling, though, is that you might need different component values for that texture control to do very much, or much more than the gain control does.

make sense?

[WGTP]

I understand what you mean.  As the resistance to ground is reduced, it is Rat like around 1K.  After that, the bass continues to get boosted until you end up with a 3db/oct. slope, sort of, instead of 6db/oct.  With the resistance small enough, that "leg" takes over and the freq response returns to normal with an overall boost.  Might need to set a minimum resistance for that around 500.  Reducing the 10uf cap to 1uf significantly reduces the bass boost effect, but I know you like single knobs that do lots of stuff.

There may not be values that make this work just right with out getting complicated.  Surely someone else has tried this before. 

The diodes in the loop are referred to by some as soft clipping and the diodes to ground as hard clipping.  The Tube Screamer is soft and the Dist+/Rat are hard.   

[Stompin Tom]

mmmmm.... distortion vs. overdrive... soft=more subtle? interesting. Should of already figured that one out for myself, but there you go... Thanks!

[aron]

What I'm wondering is whether the diode path (given the 1k resistor in series with it from the output) would provide any appreciable increase in clipping over much of the texture pot's range.

Agreed. Jack posted about resistance in series with clipping diodes to "equalize" op amps long ago and I did tests with it and it works, but it just made the pedal sound more "generic" for lack of a better word.

Give it a try and let us know.

[Fret Wire]

It's a cool idea. Soft clippers usually don't have enough ass when pushed, and hard clippers don't seem to lively when the gain is turned down low. A panning pot would allow good OD and heavy Dist. settings.

I've been bouncing ideas around to do this with just the gain pot and nothing else. As you turn up the gain (halfway), the diodes to ground start to take over, and the second half of the gain pot rotation becomes heavy distortion. And the bass is rolling off with the gain. I guess you could do it with a dual gang pot maybe. I keep thinking about doing it with CMOS switching, but can't get it from my brain to planet earth.

[christian]

This has been done before to a DIY design. This really works. I dont know about the "blend" functionality, but it should work, sorta.
You have some extra components there, just insert the diodes from opamp output to wiper of the "blend" which has 1st lug directly connected to the -input of the opamp and 3rd lug to ground via 10uF cap. The pot itself will give some resistance so you can ignore those resistors you put there.

ch.

[WGTP]

Thanks for the replies.  Another way to do this, would be to use a dual gain 10K??? pot and 2 sets of diodes with the pots wired in reverse.  As the resistance increased for one set of diodes, it would decrease for the other set.  The diodes sets could be adjusted to balance out the 2 extreme settings volume/distortion wise.  This would also remove the "boost" aspect of it, which may or maynot be desireable.  Using a 1uF cap in both places, reduces the Rat dual leg aspect of it to a minimum, which may or may not...

Check out the AMZ diode Warp article for further inspiration/explaination.

The desired result of having 2 distinct yet blendable distortions, which this should do, but the values need to be worked out.