Two Stages (op amp) Distortion pedal

[CharlieGen]

Hi, I'm trying to create customize "wampler's simple overdrive" to get a 2 stages distortion with soft clipping at the beginning and hard clipping at the end. I would like to know what you think about this schematic:

[ElectricDruid]

Welcome CharlieGen!

Both of those stages look like they're trying to be non-inverting op-amp gain stages, but both of them are missing the connection from the -ve op-amp input down to ground. Compare your first stage with the Tube Screamer gain stage (which is very similar in many ways):



The important part that yours is missing is that R4 down to ground. Since you have a dual supply, C3 is not required unless you want it for tone shaping.

The second stage has the same issue. The feedback loop has no ground connection, so it's effectively a unity-gain follower.

HTH

[ElectricDruid]

Incidentally, what result do you get from the simulation? Does it give you any gain or not?

[antonis]

Incidentally, what result do you get from the simulation? Does it give you any gain or not?

It depends on how much stonkered the simulator is..

[ElectricDruid]

Incidentally, what result do you get from the simulation? Does it give you any gain or not?

It depends on how much stonkered the simulator is..

Is that a technical term, Antonis?!?

[Vivek]

Since you have a dual supply, C3 is not required unless you want it for tone shaping.

That appears to be an 9V AC supply, with one lead erroneously labelled "-9V". And that -9V seems to be connected to ground.

[Vivek]

In LTSPICE

1m or 1M is 1 milliohm

1meg or 1MEG or 1 megs or 1MEGS is 1 mega ohm

[ElectricDruid]

That appears to be an 9V AC supply, with one lead erroneously labelled "-9V". And that -9V seems to be connected to ground.

In LTSPICE

1m or 1M is 1 milliohm

1meg or 1MEG or 1 megs or 1MEGS is 1 mega ohm

+1 agree, Vivek is right. I hadn't noticed the supply problem ("+9V" and "-9V" looks ok on the main schematic) but that's not a DC supply, and it's not a dual supply either. And 1 milliohm is *definitely* going to give you unity gain.

So...

1) Get rid of the -9V label and use the Ground symbol that LT spice provides. Connect the opamps between +9V and Ground.

2) Fix the values of R3 and R7 so you actually get 1Meg.

3) Add a resistor and capacitor in series from both op-amp negative inputs down to ground, as shown on the tube screamer schematic I posted.

4) Check that the ".step param" command isn't going to give you 999 separate plots.

Then it might work.

[PRR]

Stage 2 has no input bias.

1Meg in the 2nd stage NFB just adds hiss. (1 milli doesn't but is not a standard value.)

2nd stage at unity gain, what is the point?? The final diode network can just hang on the first stage, same thing.

There's probably issues with levels between the two diode sets. One or the other is liable to dominate. But let's get "passes signal" first.

[antonis]

Stage 2 has no input bias.

@ CharlieGen: If you want to save a bias resistor, delete C6..

[Mark Hammer]

I attempted something sort of like this a while back, although the clipping types were actually reversed.  The first stage was a warmer "hard" clipper, that fed a second "soft" clipper.  The signal amplification of the first stage meant that  the second stage didn't need to provide that much gain to hit the forward voltage of the LEDs.  The "Tone" control pans between the first and second stage outputs.  The first stage has some lowpass filtering, and the second stage applies a midscoop filter, such that rotating the Tone pot simultaneously shifts between two intensities of clipping but also two different filtering profiles.  This person here provides a Youtube demo, although his build has one or two of the controls wired backwards.  http://tagboardeffects.blogspot.com/2013/01/the-roseyray.html

It would be interesting to stick the clipping diodes in the feedback loop for the first stage, and the LEDs to ground for "hard" clipping in the 2nd stage.  I never attempted that, though now you've made me curious to try it out.  Note that, when the signal fed to an op-amp is not constrained to keep the amplified version within the voltage-swing limits, you run out of headroom and the op-amp clips.  If one then feeds the op-amp's output to a diode pair going to ground (such that the diodes play no role in the negative feedback of the op-amp itself), you end up with a sort of "double clipper".  The first clip results from slamming up against the headroom limitations of the op-amp, and the output is clipped a second time by the diodes.

[Mark Hammer]

Redrawn to switch the two clipping stages around.  A little more lowpass filtering on the first stage, such that it will be discernibly warmer than the output of the 2nd stage.  I'll try and build it later tonight.

[Vivek]

Dear Mark

On circuits like this

How do you decide how much gain should each stage have ?

[Mark Hammer]

That's an excellent question.  I have to confess it is mostly guessing.  Stage 1 can have as much gain as 213x.  This is almost twice as much as a TS-9 (which is 118x), but not more than an SD-1.  Keep in mind that, although gain is multiplicative, the diodes will hold the signal level to a maximum value of about +/-600mv.  Stage 2 has a fixed gain of 101x.  Although 101x multiplied by 213x is WELL in excess of what can produce intense clipping, Stage 1's output is limited, and Stage 2 has the burden of reaching the forward voltage of the LEDs, which is about 3x higher than that of the silicon diodes in Stage 1.

Of course, a person doesn't HAVE to use all the available gain.  Set Stage 1 to its minimum gain of about 11x, and you can still get good, if not intense, clipping from Stage 2, and use the Tone control to blend in a little of Stage 2 with mostly Stage 1.

I'm about 2/3 of the way through building the "Version 2" on perfboard,  If I think the finished product needs any adjustment of the gain stages, I'll let you know.

[antonis]

I think Vivek's interest about gain amount concerns clipped waveform shape..
(and here might come his beloved discussion about slew rate..)

[Mark Hammer]

So here is the almost-finished version of the Roseyray 2.  In response to Vivek's question, YES, I had to drop the gain of stage 2.  I still get a wee bit of instability on the 2nd stage at maximum gain, but the first stage behaves itself nicely.  I think I prefer this to the original version, where I was hard-clipping in stage 1 and soft clipping with LEDs in stage 2.   It's been a while since I had an original one functioning, but this seems to have a wider contrast between "personalities" as one rotates the Tone control, from blues at one end to death metal at the other.  In addition to improving stability, reducing the gain of stage 2 also tends to even out the volume balance between stages.  You wouldn't think that so much cumulative gain (max gain of 4780X !!!  ) would do so, but the midscoop filter eats up enough signal that the two "sides" tend to balance out.  I do need to figure out how to stabilize it, and may reduce the gain of stage 2 even further, perhaps replacing the 22k ground leg with 39k or maybe even 47k.

Sorry for the derail.  I hope the OP takes some inspiration, and draws some useful inferences.

[anotherjim]

I keep thinking it ought to be possible to replicate the Acapulco Gold with a dual opamp, probably MC1458. Just to save using up a pair of LM386. I've messed with it on breadboard and it's promising. I found a series "coring" diode pair between stages helpful not especially for the crossover distortion but for noise gating. x2 high gain stages are as hissy as a bag of snakes. It also makes metal palm mutes snappier.