Op Amp biasing

[Steve Conner]

Executive summary: Measure the DC voltages of all the op-amp outputs (wrt ground) If any of them is far away from 4.5V, you did something wrong.

I'd expect a guitar pedal (other than a distortion pedal of course) to take the output of a Les Paul without distorting. Humbuckers give a couple of volts peak-peak when played hard. So, it's just doable with a TL07x type chip running off a single 9V supply with unity gain: the output can get to within about 1.5V of the rails, so the maximum unclipped output is about 6V p-p.

If I was building an EQ pedal, I'd include a soft clipping circuit to prevent ugly op-amp clipping when a lot of boost was applied. For example, in my Ninja Toaster EQ, the gain stage is a tube, and the gain structure is arranged so it always clips before the op-amps do. When you apply boost you get tube distortion.

[J M Fahey]

Way to go
And to our friend lowell:

Hmmm... Well how predictable is it whn u don't know what the input source is?? This begs the question how does any designer know if a singled coil, humbucker, active pickup, or a string of preceding pedals/boosters are the input source? How does one design with that in mind?

you do the same as tube designers do; stick a 1M resistor as input impedance and that's it.
As of signal handling: a typical 12AX7 is cathode biased to 1.5 to 2.2V so any peak voltage reaching its grid with same negative peak voltage is going to cutoff that tube, which is an *extreme*, very sharp form of distortion, no way to avoid that.
And *usually* , same positive peak will saturate the tube, although in that case clipping is softer and it alsop depends on the value chosen for the plate resistor.
But in a nutshell, a tube preamp can reasonably have 1M input impedance and handle no more than 3 or 4 V PP.
That performance can easily be achieved with Opamps.
And soft clipping, on one polarity or both can also be achieved easily.
Music Man is an early example of that, with their "amplified diode" technology:


EDIT: to complete the answer:
the Op Amp has a gain resistor which is: R13+R14+R15=514K.
It has 2 transistors in parallel with it (one for the positive swing, one for the negative one) which will turn on and clip when their base gets, 0.65V .
And we have a built-in diode from base to emitter.
Since there is a voltage divider from collector to emitter the BE diode clipping voltage will be "multiplied" by the divider attenuation.
So for the positive peak, Q2 will clip when it has 0.7V at its base. meaning it will have 0.65V*(514K/22K)=15.2V.
In fact we have to add an extra 0.65V to that, because of D2, which is also in series.
Same happens with the negative peak.
So we see that MM Op Amp clipper *just* reaches the rails, and in practice it has been slowing down a couple volts before that, the clipping waveform clips nicely rounded.
Later MM used unsymmetrical clipping, by replacing one of the 22K resistors by a 47K one, so the Op Amp clipped on negative 15V but on positive 7V.
Carvin, Crate and Hughes&Kettner use zener diodes for clipping, a 3.3V and a 6 or 7V one, but clipping is much sharper and harsher.
So we see the Op Amp *is* an universal building block.
I have been able to get unsymmetrical distortion, soft clipping, duty cycle variation, input signal "grid" clipping, compression, etc. , all desirable "tube" characteristics, without a single red hot piece of wire inside, just with "sand", go figure.

[Steve Conner]

I must admit I hadn't seen that circuit before.

[J M Fahey]

Well, they are not much used in 500KV+ circuits

[lowell]

JM that's a good point about preamp tube biasing and cutoff. I didn't take that into account because I generally don't worry too much about super CLEAN audio when it comes to tube audio circuits. Especially for guitar. A little clipping here or there may or may not be "heard." Where as with opamps I DO think it would be heard. Maybe I'm wrong. I'm pretty sure I've heard many a tube amp with more than 3 v p-p hitting its grid and have not been disgusted by the sound. :-)

[J M Fahey]

I'm pretty sure I've heard many a tube amp with more than 3 v p-p hitting its grid and have not been disgusted by the sound. :-)

Well, 3Vpp would *barely* be touching clipping, and probably only on some very narrow peaks, not the main waveform, so it would not be noticeable except on a scope.
And a d*mn fast trigger one, at that.

[lowell]

So gettin' back to this old thread here. I basically took a preamp from a great solid state amp, and built it into a pedal off a 9v supply. Having thought about this project over the last few months I think I may have figured out why it sounds so distorted and... well, bad. If the original circuit had an 18v supply, would I need to compensate the op-amp gain stages so as to be relatively HALF the gain of what they are? Or said another way, is having a 9v supply the reason this thing clips so easily? Seems obvious, and like a stupid question, and I believe this is the answer to my problem with this particular project, but please do chime in. Again, designing op-amp circuits is fairly new to me.

[loudthud]

The big headroom bottleneck is the first gain stage, or all gain stages before the first volume control. The signal can be clipped there before you have a chance to reduce the signal level with the pot. Realistically with a 9V power supply, the gain needs to be about 2, or some kind of variable gain structure. A gain of 2 will accommodate all be the highest output guitars. A variable gain stage will allow you to increase the gain for low output guitars.

Many stompboxes have an emitter (or source for a JFET) follower as the first circuit. It's function is just to present a high impedance to the guitar with a gain of slightly less than one and a lower impedance to the rest of the circuitry so a lower value pot can be used. An opamp with a gain of one can be used.

[lowell]

thanks LT. I'll be tweaking that 1st stages gain.