How does diodes clipping work *in between* gain stages?

[midwayfair]

I understand how (and where) diode clipping creates distortion, but I'm trying to wrap my head around how they work when they're in between gain stages.

Some background: I recently picked up a ton of diodes (like, more than 20 types) and, since I didn't have a lot of circuits in which to test them, breadboarded and built a few things that use diodes in different ways where they're really noticeable. So I put sockets in a klone and a Harmonic Percolator, and two things on my breadboard, a simple op-amp clean boost and a dual boost thing, which I ended up liking enough to do a layout (http://www.diystompboxes.com/smfforum/index.php?topic=98091.0). The last one is where I noticed something funny.

For simplicity, we're basically looking at two gain sections, or two discrete amplification circuits. Whatever's the easiest way to think about it. I'm sorry to use a circuit that isn't something everyone's familiar with, but I don't think there's anything particularly unusual about it.

If I put the diodes at the very end of the circuit, after the second boost, they clipped like crazy, just like you'd expect any distortion circuit to do. They did that even though the boost was clean -- there is no transistor distortion involved. The transistor is also getting plenty of voltage.

But then if I put them after the first stage, it was much more mild clipping, even when the gain was set to be the same as the second stage. It's mostly just a slight bassyness and maybe a little fuzzy warmth added. Not like a distortion circuit at all, back into OD territory. The volume output from both stages is pretty similar and the voltage on the collectors is similar enough. And clipping diodes in the middle of gain stages in other distortion boxes (muff, klon, etc.) have a much greater effect. Like I said, I was happy enough with the sound to bother doing a layout, but I had originally expected to need a series resistor to calm down the clipping. Instead I ended up not being able to get more if I want it.

So I started messing around with some stuff just to experiment.

1) Removing the second circuit - just lifting the input capacitor to the second stage and moving the "out" jack to the place where the diodes are - made the diodes clip more, just as you'd expect if you put them at the end of a boost circuit. Add the second stage back in and they go back to super mild clipping. It's like they're clipping less just because another gain stage follows. As far as I know, this isn't what happens when you run two boards or pedals off the same power supply in the same enclosure, even though to me it seems like the same thing.

2) I noticed that the diodes have more of an effect when the gain is turned up -- this causes transistor distortion, too, but the diodes clearly have more of an effect. It's not volume related. If that were true, a higher gain transistor would cause more clipping. It doesn't.

3) I reduced the voltage on the collector -- getting it down to 6v -- and this also seemed to make the diodes clip more. A lot more, actually. Again, though, it caused more transistor distortion.

So ... what's going on here? In the second example, allowing more voltage to pass to ground affected the diodes. In the third example, less voltage in that part of the circuit overall caused the diodes to create more distortion.

[ashcat_lt]

These are back-to-back diodes to ground?

Is there an AC coupling cap after the first gain stage.  You mentioned an "input cap" on the second stage, but is there an "output cap" after the first?

[midwayfair]

These are back-to-back diodes to ground?

Is there an AC coupling cap after the first gain stage.  You mentioned an "input cap" on the second stage, but is there an "output cap" after the first?

Yes, there's an output cap after the first stage. The diodes, and one side of each cap are all on the same bus.

Back to back diodes to ground.

[R.G.]

So ... what's going on here? In the second example, allowing more voltage to pass to ground affected the diodes. In the third example, less voltage in that part of the circuit overall caused the diodes to create more distortion.

The diodes always do the same thing. They each have their own relation of current through them to the voltage across them. That does not vary. Put X current through them, you get Y voltage across them. Each one is a bit different, even within a type number, but diodes are much more similar within a type number than, for instance, BJTs or (shudder!) JFETs.

What is different is exactly what you put your finger on - the circuits preceding the diodes and following them. One thing that many people forget is that the output of a circuit is not just a voltage, it's a voltage through what amounts to a resistor in most cases. And the resistor is sometimes a 'virtual resistor', the combined effect of how much current the circuit can supply to any load.

Likewise, the input of a circuit has an implied series resistor which limits current into it, and an implied shunt resistor to ground which interacts with whatever drives the inputs.

I think what you're hearing is the differences in how much internal 'resistance' the driving and following circuits have, not the diodes.

[midwayfair]

Thank, R.G. This makes sense. And since changing the voltage and gain both affect the series resistance of the entire circuit, that would explain why altering these two values would have a similar effect on the diodes' clipping.

[R.G.]

There is a related but not identical thing you should try. Set up one of your circuits where the diodes clip *hard*. Now, put a 10K pot wired as a 0-10K variable resistor in series with the diodes only. Between diodes and ground should do it.

Now dink with the pot and listen.

[midwayfair]

There is a related but not identical thing you should try. Set up one of your circuits where the diodes clip *hard*. Now, put a 10K pot wired as a 0-10K variable resistor in series with the diodes only. Between diodes and ground should do it.

Now dink with the pot and listen.

Yup, I've done the series resistor thing in a few circuits to soften the clipping & I like the effect. I also like the resistor in series with just one of the diodes to create more even order harmonics. I'm an asymmetric clipping junkie.