Hard Clipper Diodes - where to insert?

[Passaloutre]

I built a Tube Screamer last night (actually a "Son of Screamer" http://www.muzique.com/tech/scream.htm), and I'm very pleased with the sound.



However because I like to tinker, I'm toying with the idea of having a switch that goes from "feedback clipping" mode to "hard clipping" mode. In other words, the switch will remove the diodes from the feedback loop and send them from the signal to ground. I've seen something like this in John Hollis's "omnidrive":



What I'm confused about is the 1K resistor at the output of the first opamp. In the Tube Screamer the diodes take the signal straight from the opamp output, before the 1k. On most of the hard clippers I've looked at (DS1, Rat, DOD 250, even the Omni Drive above) however, the diodes take the signal from after the resistor:



I'm wondering how much of a difference it makes. If I can just move one end of the diodes, then I can use a spdt switch, but if I have to move the whole clipping arrangement, then I need a dpdt.

[Mark Hammer]

R4/C4 in the Tube Screamer form a lowpass filter that provides a sort of default rolloff, which the tone control can add to in different ways.

Since that circuit uses a virtual ground that biases the first op-amp stage to 4.5V via R1, you need to eventually remove that DC voltage before the signal heads out the door.  In the case of the TS-9, that function is performed by C6 near the volume pot.  In the case of the DOD250, however, there IS no second stage, so the DC-blocking function is provided by the 22uf cap on the op-amp output (22uf is actually pretty darn high for that circuit  ).  The 10k resistor works in tandem with the .0022uf cap to roll off some of the fizz, but it also drops the current a bit to hit the diodes a certain way (diodes conduct differently, depending on current).

Bear in mind the difference in gain between the TS and 250.  Max gain in a TS is around 118x, while max gain in a 250 is around 213x.  That can make a difference in how hard you hit clipping diodes.  Moreover, the TS rolls off bass so as to yield more equivalent clipping at any gain, across the entire fingerboard.  So lowering the bass, TS-style, and also reining in the gain TS-style, is not going to be great for hard clipping, right off the bat.

The other thing to keep in mind is that signal level will be "clamped" by D1/D2, such that you'd need to either add some gain somewhere to bring the signal up past the clipping threshold of another diode pair to ground, or else opt for diode with a lower forward voltage (clipping threshold).  That would suggest a pair of germanium or schottky types. (or even one of each).  However, lowering the clipping threshold would also severely limit your maximum output.  I'm assuming, here that both pairs of diodes would be in circuit.  If, on the other hand, you wish to select between feedback or hard-clipping pairs, then none of what I've said above applies; with the exception that "hard" clipping won't be particularly hard with a max gain of 118x

Where to insert?  The sensible place would be to insert a 2k2 resistor after C6, and ru your diodes to ground from the junction of that added resistor and R12.  Note that IC1B can be adjusted to provide some additional gain, but that would require changing some component values so as to maintain similar tone-control functioning.

[Passaloutre]

Thank you for the thoughtful analysis! Yes, I was planning to set it up for *either* hard-clipping *or* soft-clipping, but not both at once. In that case, do you think the first opamp wouldn't have enough gain for hard clipping? Lowering R2 would increase the gain.

If I put the diodes mid-circuit, they would have to be "grounded" to Vref right?

[anotherjim]

Changeover?
D1 & D2 connect AFTER the 1k R4 at C4 instead of at the amp output.
Change over switch to either connect the diodes as it is at the amp - input OR via 10uF (as per Hollis) to 0v.

[bloxstompboxes]

Thank you for the thoughtful analysis! Yes, I was planning to set it up for *either* hard-clipping *or* soft-clipping, but not both at once. In that case, do you think the first opamp wouldn't have enough gain for hard clipping? Lowering R2 would increase the gain.

If I put the diodes mid-circuit, they would have to be "grounded" to Vref right?

If you were to do both, then you would essentially have the same circuit as a large number of DOD pedals. This would include the Classic Fuzz, Super Distortion, a couple of others. Values of components are the only differences.