In looking at distortion pedal circuits, two types caught my attention: those that used shunt diodes after an op amp (Proco Rat), and those that used diodes in the feedback loop of the op amp (Ibanez Tubescreamer). Up until this point, I've had a bit of distain for getting distortion with silicon, prefering to plug directly into a tube amp and turn it up. But so many guitarists can't all be wrong, so I started playing with these circuits.
There's a lot of talk about circuits, op amp types, and the diode(s) used in these pedals. I breadboarded versions of the Rat and TS9, listened to the output of the different stages, and fooled around with diode combinations. Here are some notes:
The effect of the diodes is pretty minimal on the Rat. I think that most of the distortion is generated by the LM308 IC, and it sounds surprisingly good. Without the shunt diodes, you really don't get much of a clean signal, the op amp is distorting from the start. The shunt diodes make the onset of distortion start sooner on the Gain setting.
There is a mistake on the schematic I link to--the input cap polarity is wrong. But this version of the circuit really doesn't work well--the volume is WAY too low. If you delete the JFET buffers, and just use the IC section, it's fine. I don't know if this is a mistake I made and can't find, or whether there are additional errors on the schematic. I found other Rat schematics with slight differences, so I don't know what the *real* circuit is, but the LM308 section of the circuit is all that's really needed for comparisions.
The Rat tone filter is OK but nothing great, and is nice for rolling off the highs, but you loose some volume.
The input buffer on the TS9 really doesn't color the sound. Its just that, a neutral buffer.
Listening to the signal right after the first stage of a TL072 IC without the distortion diodes in place shows a fairly clean signal. If you set the gain up, or play hard, it will distort with an ugly, gritchy sound.
If you add the tone filter section (still without distortion diodes), the improvement is tremendous. Rolling off the high gets rid of all the gritchiness, and it sounds like a clean boost. (This may be something to play with later.) I think this might explain why a lot of guitarists tend to set the tone setting fairly low.
The distortion diodes matter on the TS9. The type of diode adds subtle coloring. I think the biggest area to play with is in combinations of diodes so the clipping is different on the positive and negative sides of the signal. My preference at the moment is for a 1N34a and 1N001 in series, paralled with a 1N4148 with the opposite polarity.
In the TS circuit, I don't really like the sound of shunt diodes (with the feedback loop diodes removed). It lowers the output volume substantially for one thing.
I tried making a hybrid of the Rat and Tubescreamer by seeing what sections of the circuits could be shared. The answer is none. The input buffer of the TS9 can drive the Rat, there's just a little volume loss, but it really doesn't need the buffer. The TS9 tone control is not fit for the Rat, it doesn't do enough and the Rat sounds buzzy. And the output buffer doesn't do anything positive either. I started out thinking these circuits were fairly generic, but now have a better appreciation for the design that went into the different sections to make them sound good.
I've been using these circuits for a while now, and prefer the Tubescreamer in front of a Fender amp. The Rat was too buzzy for me, but it gets way more distortion that the TS can. I added another cap in parallel to the one in the tone control to cut some more of the highs which helps a bit, but I only use it for a few situations. I occasionally use the TS in front of a Marshall Super Lead, but it doesn't add a whole lot, and don't use the Rat at all.
On the TS, I tend to set the gain around 9 o'clock, the tone rolled completely off, and the volume around 1 o'clock. This boosts the signal and it over drives the tube amp, and the gain adds just a bit a grit.