DS-1, Rat, etc... opamp clipping??

[ashcat_lt]

This has come up in a couple of different threads in the past, but it was a while ago, and I can't be bothered to search, and I think it deserves a thread of it's own.

So the question:  Does the opamp in these diode-to-ground clipping circuits actually end up clipping on its own?  That is, does the output actually swing close enough to the rails to be clamped down? 

It seems to be the common wisdom that, in fact, they do, and that this is an important part of the sound of these boxes, but I'm pretty sure that I've seen in at least one place where RG (I think it was) seemed to be saying that this was not the case - the diodes clamp the output to something far short of the rails, and the opamp never ends up having to try to get there.  It is quite possible that I misinterpreted what was being said there, but it almost makes sense in my head that this would be the case, though it would seem to stand on ear a lot of people's ideas of how these things work.

Can anybody shed any light here?  RG, got your ears on?

[slacker]

I think R.Gs comments were about the Tube Screamer type circuit, where the diodes are in the feedback loop of the opamp, not circuits with clippers to ground after the opamp.
In the TS type circuit apart from for signals significantly bigger than the forward voltage of the diodes the signal is clamped by the diodes so the opamp doesn't clip.
In Rat type circuits the diode clipping is completely separate from the opamp so if you give the opamp enough gain it will clip.

[R O Tiree]

Slacker's right - the voltage at the output pin of the LM308 (pin 6) will be slamming from rail to rail, given enough "Drive" and input signal. (Gain is well over 100 at full bore, so only a few 10s of mV signal will send it to the rails.) Then there's a 4.7µF de-coupling cap to strip out the DC (and a 1k resistor just before the cap), followed by the diodes clipping to GND which limit the signal going forwards into the tone circuit to within +/- 1 diode drop. If there was no de-coupling cap then, yes, the output at pin 6 would indeed be limited and therefore would never hit the rails.

[Mark Hammer]

I don't know that I was the only one, but I know I am among those who have noted that, when huge amounts of gain are being applied to op-amps, with a 9V supply, and a realistic voltage swing of maybe +/-3.5v, that chip is not going to give you a "clean" output.

But do note the distinction between diode-based clipping in the feedback path vs diodes to ground - the Tube Screamer vs Rat scenario.  Where diodes are in the feedback path, the voltage swing the op-amp is allowed to produce is constrained by the forward voltage of the diodes, such that it won't approach the limits of the chip.  In contrast, where the clipping diodes come after the chip's output and there is nothing constraining the chip's amplification apart from the supply voltage and max voltage swing, one would expect that the sound is a product of both the coloration provided by the chip running out of headroom AND the clipping diodes.  I won't speculate on how much each contributes.

It is worth noting that one respected pedal - the Black Cat OD-1 overdrive - relies solely on the headroom limitations of the dual op-amp chip it uses (OP275) to produce distortion.  Nary a diode to be seen, other than the power supply.

[ashcat_lt]

Oh, I know that they will clip.  I've got a monstrosity on the breadboard right now which purposely gives an LM324 barely enough power supply to function, and then slams the signal through with a bunch of gain.  There's a diode in power supply for polarity protection, but the nowhere in the signal path.  It is actually a pretty awesome sounding thing.  Once I get it tweaked the way I want it and actually build the thing I'll be sharing it...

It is possible that I'm mis-remembering some discussion of the TS-style clipping, but I could swear that it came up in a Rat discussion, and that the assertion was pretty specifically that the idea of opamp clipping in these circuits was a fallacy.  Course, I may have killed a brain cell or two since then...

[Beo]

Conventional wisdom is that opamp headroom distortion (slamming the rails) is nastyand to be avoided. I've never found it so bad sounding, but I don't play at gig levels, and can't really tell the difference between a lot of ODs and distortions.

What trips me up sometimes when I'm breadboarding various experiments is whether I'm hearing the clippers I'm testing, or if I've started hitting the opamp rails. I should probably pull out my oscope more often, or learn how to do sim analysis.

Also, I'm a sucker for thinking that louder sounds better, when A/Bing two variables. very subjective.

[ashcat_lt]

The sound of a 12AX7 distorting is actually not particularly pleasant either. 

Once you get to a certain point of squaring off a wave, it's pretty harsh no matter how you've gotten there.  We use filters, either electronic or physical, to shape it into something we can stand to hear.  I think opamps may have a tighter knee than a tube or a diode so that sudden peaks which clip might be a little harsher, but for most of the travel of the gain pot on a Rat you're well beyond that knee for most of the signal swing anyway, and I don't think there's much difference sonically.  Of course, this board is really all about subtle differences that the people who listen to our music would never notice...

[Lurco]

Oh, I know that they will clip.  I've got a monstrosity on the breadboard right now which purposely gives an LM324 barely enough power supply to function, and then slams the signal through with a bunch of gain.  There's a diode in power supply for polarity protection, but the nowhere in the signal path.  It is actually a pretty awesome sounding thing.  Once I get it tweaked the way I want it and actually build the thing I'll be sharing it...

It is possible that I'm mis-remembering some discussion of the TS-style clipping, but I could swear that it came up in a Rat discussion, and that the assertion was pretty specifically that the idea of opamp clipping in these circuits was a fallacy.  Course, I may have killed a brain cell or two since then...

To the rescue of your brain cells: http://www.diystompboxes.com/smfforum/index.php?topic=94720.msg818349#msg818349 

[Jopn]

Of course, this board is really all about subtle differences that the people who listen to our music would never notice...

Whoa, wait.  Your guys play music TOO! 

That's so cool.

[ashcat_lt]

Cool Lurco!  That's the one.  Reading it now I see that I have mis-remembered, or perhaps just misunderstood to begin with.

Thanks guys.

[Kesh]

The Rat uses a pretty low power op amp. I wonder if driving the clipping diodes through that 1k resistor, which are effectively earthing everything over 0.6V, actually overloads the op amp so it sags rather than clips.

[teemuk]

I wonder if driving the clipping diodes through that 1k resistor, which are effectively earthing everything over 0.6V, actually overloads the op amp so it sags rather than clips.

Probably not. Most opamps can drive 600 ohm loads without clipping due to current limiting - and that is with a 30V supply. At low power supply voltage levels they could likely tolerate even lower loads so it probably justy falls down to supply voltage being the limiting factor.

Sagging really depends on power supply's characteristics, like internal resistance and the filtering capacitance. Unless you intentionally limit current with something like 1K series resistors in the [opamp} supply line (coupled with intentionally underrated filtering capacitance) you probably won't be generating much sag with the current draw of the opamps.

Conventional wisdom is that opamp headroom distortion (slamming the rails) is nastyand to be avoided.

It's not conventional wisdom by slightest bit. It's merely a common false belief.

For instance, the early SansAmp designs relied solely on opamp clipping and despite that their distortion tone usually gets all kinds of rave reviews.

[Bill Mountain]

If you don't want the opamp to clip you can always but 2 Zeners anti-parallel totaling about 6 volts and it would give similar headroom with possibly smoother clipping.