Tech21 double drive - clipping curiosity?

[Pojo]

So during my lunch-time internet meanderings I stumbled across this schematic:



I highlighted the areas that I'm particularly interested in. But I'll start by mentioning that I'm not familiar with this pedal and haven't so much as heard a demo yet. Anyway, it appears that the 2nd half of the TL072 is biased in such a way that the signal sits at something other the Vsupply/2 (I really need to work on my electronics math) causing asymmetrical hard opamp clipping as the gain is increased via the Drive A pot. Then over at the 1st half of the TLC2262 we have a series pair of 1n914s in the feedback loop which I think would soft-clip only half the wave form. But I also noticed that the signal is coming into the inverted input with VB1 going to the non-inv input. What would be the purpose of setting up that stage in this way?

To me this seems like a clever and interesting circuit, I'm just hoping someone can help me fill in the blanks a little

[slacker]

You've got it pretty much right, the first stage is biased around about 7.4 Volt so it will hard clip the top of signal, the lower part of the signal has more headroom so will stay cleaner for longer. The next stage is set up as an  inverting amplifier the diodes will clip what was the bottom of the original signal, so that gets a bit softer clipping than the other side. This clipping is different than you get in a Tube Screamer type circuit with diodes in the feedback loop of an opamp, that's probably why they did it like this. Note that the next stage is another inverting amp so the signal gets flipped back the right way up again.

I think this is probably an attempt to get something vaguely like tube amp clipping, with enough gain the result will be similar to figures 5 and 7 here http://www.dwfearn.com/tubes_vs_transistors.html

[Pojo]

Awesome, thanks Ian! My breadboard awaits, though I wish I had an oscilloscope!

[Pojo]

Very cool link too.....I enjoyed the read!

[chi_boy]

Sooooooooo......  What's going on with those JFETs and the diodes to nowhere?  Is that a switching circuit?

[slacker]

Yeah they're part of the switching circuit, the rest of it isn't shown, but it's probably similar to what you'd find in a Boss pedal.

[slacker]

Just checked out some demos of it sounds good, I hadn't realised it had two gain knobs, for some reason I assumed gain A and gain B were a dual pot.

[Govmnt_Lacky]

Could this be done with a TL074 and a TLC2262? Looks like it should! 

[slacker]

Don't see why not.

[Govmnt_Lacky]

I really HATE not knowing how to do PCB artwork! 

[Govmnt_Lacky]

If this were to be set up in a "true bypass" configuration, I would assume that the 2 JFETs could be eliminated with R18 being directly connected to the non-inverting input of the TLC2262 and the connection marked "A" could be eliminated as well.

Am I correct?

Also, is the inverted input on the TL072 near the circuit's input directly connected to the op amp's output? I HATE when there are no connection points! 

[slacker]

Am I correct?

Yes.
If you want to build it true bypass stylee then with a few component tweaks you could get it down to two dual opamps. The input and output buffers could be ditched.

Also, is the inverted input on the TL072 near the circuit's input directly connected to the op amp's output? I HATE when there are no connection points! 

Yeah, I agree there should be a dot at the crossroads to show that they connect and not just cross.

[chi_boy]

So now I'm curious.  What parts could go to eliminate the switching?

Top is the original, Bottom is a first swing at it.

I'm confused about caps in the signal path too.  They are all polarized and in different directions.  What kind of caps are those supposed to be?

[Govmnt_Lacky]

WOW!!! Those are some BIG schemos 

[chi_boy]

WOW!!! Those are some BIG schemos 

Wow is right!  Preview is you friend, right?

[chi_boy]

OK, should be better now.

[chi_boy]

Changed the in and out.  The one above didn't look right.

[slacker]

That looks good to me. I would change C2 to 1n and R3 to 1Meg, that will give the same bass cut as the original C13 and R7 in the schematic Pojo posted, about 150Hz but will give the same high input impedance as the original buffer. I'd also put R12 before R13 like in the original.

You can use electrolytic caps for C4 and C8, C4 should be negative to ground, C8 should be negative towards R13, like you have them drawn. For C5 and C6 you should use non polarised caps, they're only 0.47u and 2u2 which are easy enough to get as film caps, C4 and C8 could be replaced with film as well if you want.

[Pojo]

Nice...looks like there's some interest in cloning this bad boy

I just heard a demo too, I though it sounded pretty good too tho maybe a bit too high gain for my taste. So I think I will borrow the general idea and twist it into something else eventually. I just like the idea of soft clipping half the waveform and hard clipping the other side. Just seems like a good compromise because, for me, the tube screamers, timmy, and the like which are all soft clipped just sound unnatural (as if it's still my clean tone just with some fuzzy hair mixed in). And the pure hard clipped circuits kill too much dynamics (or perceived dynamics maybe). So far the cascaded jfets come closest to what I'm looking for but I still find that too much dynamics are lost. I know there's also other circuits that aim to mix soft and hard clipping (boss os-2, plimsoul). But they seem to take the approach of: split signal, clip separately, blend together.

Would it be possible to achieve the asymmetrical hard/soft clipping using discrete components? Or any other ideas to accomplish that general idea?

I was thinking about using the cascaded jfets which gives a more pleasant hard clipping to my ears, and biasing them so that they clip only one side. I imagine I would have to alternate the drain voltages on each fet because each stage inverts the signal. For example, a biasing scheme using 3 stages with a 9v supply that would have the drains at something like 7v, 2v, 7v respectively. Not sure how I would go about the soft clipped side...maybe just copy the opamp scheme used here. Thoughts?

[chi_boy]

That looks good to me. I would change C2 to 1n and R3 to 1Meg, that will give the same bass cut as the original C13 and R7 in the schematic Pojo posted, about 150Hz but will give the same high input impedance as the original buffer. I'd also put R12 before R13 like in the original.

You can use electrolytic caps for C4 and C8, C4 should be negative to ground, C8 should be negative towards R13, like you have them drawn. For C5 and C6 you should use non polarised caps, they're only 0.47u and 2u2 which are easy enough to get as film caps, C4 and C8 could be replaced with film as well if you want.

OK, so this should be it.  All values added and changes made.  Left C4 as electrolytic.  C5, C6, and C8 are replaced with film.  I didn't have the big grey BC Comps in my library, so I used the Panasonic 1uF and verified that the BC's will fit in the same space.