Distortion idea with schematic.

[rocket]

The first set of diodes should be dc-decoupled from the 2nd transistor, otherwise they won't act as clippers because one is conducting the dc and the other is just always off.

[ezanker]

Joe's is still posted at Effectronics:

http://www.geocities.com/j4_student/schematics.htm scroll down to homebrew

[Mark Hammer]

Cap in parallel with the first pair of clipping diodes.  I'm a big fan of double clippers, but nobody likes harmonics of harmonics of harmonics.  The key to satisfying multi-stage clippers is to shape the spectral content and gradually "spool out" the treble, such that 2nd and subsequent stages go to work on lower order harmonics, and not ALL harmonics.  Certainly what comes out at the end can have all the harmonics you want, but the balance will be more pleasing if you tame the stuff coming out of the first stage.  Consider sticking a 200-4700pf cap in parallel with those first diodes.

[fixr1984]

Cap in parallel with the first pair of clipping diodes.  I'm a big fan of double clippers, but nobody likes harmonics of harmonics of harmonics.  The key to satisfying multi-stage clippers is to shape the spectral content and gradually "spool out" the treble, such that 2nd and subsequent stages go to work on lower order harmonics, and not ALL harmonics.  Certainly what comes out at the end can have all the harmonics you want, but the balance will be more pleasing if you tame the stuff coming out of the first stage.  Consider sticking a 200-4700pf cap in parallel with those first diodes.

Would this be the decoupling that has been mentioned earlier? Would moving the .1uf cap that is before the first set
of diodes to after them make a difference?

[brett]

Hi
some filtering at the end might help (unless you want a fully "sharp" sound)
About 470pF to ground after the diodes should be in the ballpark for those 47k collector resistors. (Fc about 7kHz)
cheers

[edster]

I took off the pot on the input and added the gain pot as suggested earlier.
How does that look? Should I have a 10uf on each 680R? Baby sleeps alot so I havent been
able to get much done on the breadbord and have to keep asking ?'s to get everything down on paper first.

A single 5k pot and bypass cap doesn't seem like such a good idea in this case. The signal that develops at the two emitters is out of phase.

[fixr1984]

Heres an update to what I have so far. I finaly got a chance to breadboard this. Kinda weak till I
added the 22uf alongside the 680R. This schematic is similar to what i was toying around with. Changed a few values. I did draw up a
change that I havent had a chance to test out yet. I moved the cap going to the gain pot before the diodes to after them.
Will this help to DC block the diodes and the 10k pot?

[brett]

Hi
you need the 0.22uF cap before the diodes, so that one isn't forward biased.
I suggest that you run that gain pot to ground, and follow it with another 0.22uF cap.
0.22 uF and a 10kA pot will trim a little bass (Fc about 100Hz), but you may well like that sound, as it prevents "flabby" bass

[Steben]

To clarify brett a bit: you shouldn't have stuck the .22 cap on another place, but simply add an extra one. Mark's idea of an extra cap parallel to the first diodes to ground is indeed worth while. Maybe try some sockets and try different caps... on all places. 

[Mark Hammer]

Cap in parallel with the first pair of clipping diodes.  I'm a big fan of double clippers, but nobody likes harmonics of harmonics of harmonics.  The key to satisfying multi-stage clippers is to shape the spectral content and gradually "spool out" the treble, such that 2nd and subsequent stages go to work on lower order harmonics, and not ALL harmonics.  Certainly what comes out at the end can have all the harmonics you want, but the balance will be more pleasing if you tame the stuff coming out of the first stage.  Consider sticking a 200-4700pf cap in parallel with those first diodes.

Would this be the decoupling that has been mentioned earlier? Would moving the .1uf cap that is before the first set
of diodes to after them make a difference?

No.  You are talking about caps in the signal path itself.  I am talking about a cap to ground that will "dump" high end from the first stage.  Actually it doesn't eliminate high end.  Rather, it imposes a shallow (6db/octave) rolloff, such that the highest treble will be well below the clipping point when it arrives at the second stage and gets boosted again. 

Keep in mind that the harmonics in the original are well below the level of the fundamental.  Once the overall signal gets boosted by the first transistor, the lower order harmonics are raised to the level where they can be clipped by the diodes.  Not as much as the fundamental gets clipped, but clipped nonetheless.  Clipping adds more harmonic content.  When the sum total of the original (unclipped) harmonic content and the added (through clipping) harmonic content goes to the 2nd stage, it gets boosted and clipped again.  Keeping in mind that gain is multiplicative (x10 gain followed by x10 gain = x100 gain), that 2nd stage raises the level of those upper harmonics to the point where they can get clipped by the 2nd pair of diodes, thus adding even more upper harmonics.  This is normally what people might refer to as "fizz" and is the undesirable part of distortion. 

By sticking the suggested cap to ground between stages, you do a few things.  You attenuate any upper harmonic content originally in the signal such that even 2 stages of boost are insufficient to introduce clipping of it (but remember the fundamental and lower order harmonics are still ready and rarin' to get their heads lopped off....twice).  By sticking that cap in parallel with the diodes, you also attenuate some of the upper harmonic content that gets added in the clipping process.

For an analogy, consider a voice microphone that hangs from the ceiling over the edge of the stage.  Sure, it WILL pick up the singer's voice, but it will also pick up the noise from the crowd.  The crowd is also part of the room signal, but you'd prefer to hear mostly the singer and not the crowd when the mic signal gets amplified.  So, you point the mic at the singer.  The mic WILL still pick up some of the crowd, if the singer and crowd are close enough to the edge of the stage, but you'll hear mostly singer when it gets amplified.  In our case, we're trying to make it so that it is mostly fundamental and lower order harmonics that gets re-amplified and clipped.  The suggested cap is a bit like pointing the mic in the singer's direction instead of letting everything enter the mic.

And while I have your attention, that should have read 2200-4700pf, and not 200-4700pf.

[fixr1984]

Thanks for the analogy Mark, I think I understand what you mean.
I've updated it again.
Brett, Is this what you mean by grounding the pot and adding the cap after it?
Hows it looking. Cap values arent set in stone, I will tweek on the breadboard once I have the
time to be able to make noise

[Steben]

Thanks for the analogy Mark, I think I understand what you mean.
I've updated it again.
Brett, Is this what you mean by grounding the pot and adding the cap after it?
Hows it looking. Cap values arent set in stone, I will tweek on the breadboard once I have the
time to be able to make noise

SORRY! BUT!.... the second .22 cap shouldn't be in front of the second transistor. It blocks for instance DC, remember, so Q2 wouldn't bias...
IN OTHER WORDS: The second 2.2M resistor should be connected to the base of Q2 (after the cap), instead of the lug of the 10k pot (before the cap).

[fixr1984]

Gotcha

[mac]

I'd try also isolating the diodes from the pot (not necessary), and use different pot values.
Besides, I'd replace those 680R with 1K pots as gain1 and gain2.
... just for fun.

This draws 0.3ma from the battery... nice, no need for a dc out.

mac