LEDs as clippers

[asfastasdark]

What would be a good standard LED to be used as a clipping device (you know, two parallel diodes, one reversed, both going from input signal to ground)? I don't know if there is an audible difference in different colors/sizes, I would guess not, but still just want to make sure.

[Zen]

+1 on question.  Seems like I read a description of the subtle differences in colors due to slightly different input and output frequencies, eg red is harsher than green but has a lower threshold... or maybe I just dreamt that bit (it happens)

[bioroids]

I don't know about harshness, but different colors have definelty different thresholds (they may vary from maybe 1.2v the red ones to more than 2v the blue ones).
That would account for different sounds, because higher threshold means lower distortion at that stage, but also higher levels to hit the next one. Also a higher threshold may cause the opamp (if you are using one) to hit the rails before the threshold is reached.

Greetings!

Miguel

[kurtlives]

With diodes "you want" one going one way the other going the reverse usually.

With LEDs as clippers you usually just use two in the method stated above. If you put them in series the world voltage would be close to over a volt and they would do next to nothing clipping wise.

Some people claim they hear difference between 3mm and 5mm, red vs yellow. Personally I haven't heard that but I haven't really done a lot of experimenting with colors and sizes.

Water Clear High Brightness ve Diffuses do make a big difference imo. I always use 3mm Red Water Clear High Brightness  and love them. They light up when you play much easier than diffused.

Where you put them is up to you and the circuit,,,,op-amp feedback loop, right near the output etc.
Experiment

[DougH]

Red 3mm's are my favorite.

[Mark Hammer]

Red generally provide the optimum forward voltage, in terms of still providing suitable clipping and a nice big output level.  The trouble with  other colours is that you have to crank the gain so high to produce any appreciable clipping, that when things are not getting clipped, you still have to contend with the added hiss from all that boosting.  Better to boost a moderate amount, get the clipping you seek, and keep accumulated hiss to the minimum possible.  Trust me, a 1.5V signal is more than sufficient to pummel the input of your amp.

Note as well, that when the clipping threshold is exceedingly high, you will only get clipping on the initial peaks, and not a clipping that continues for any appreciable period of the note.  What we call fuzz is clipping that extends over a considerable portion of the lifespan of a held note, not just the first few milliseconds.  Because the overall envelope of a plucked note tends to produce a short peak followed by a quick decay, which is followed by a prolonged period where the fundamental and some lower order harmonics ring for a while at a much lower amplitude, you need to think in terms not of what the pickups will produce immediately when you smack the strings, but rather in terms of what they will be producing 300-500 milliseconds later, after all the peaks have disappeared.  THAT is where fuzz lives.  And if that signal is 15-20mv (or less, if we are talking unwound strings on the higher frets, a full second after plucking) and you need to goose it up to 3v p-p to get anything that sounds like distortion, you may be asking for trouble and disappointment.

If anything, many might find greater satisfaction in simply sticking a second diode in series (i.e., a 2+2 diode complement), so that you get the higher output, but don't need tons of help to get there.

It's not a question of what sounds good or bad.  Rather, it is a question of tradeoffs.  For many, the sweet spot in all those tradeoffs is something around a 2+2 diode complement to a pair of red LEDs. YMMV

[R.G.]

Distortion tone is almost independent of clipping sharpness (and hence LED color) the more the signal X gain would have exceeded the clipping threshold had it been allowed to. That is, clipping sharpness only matters for signals which exceed the clipping threshold only modestly.

As a wise old man once told me, all square waves sound alike.

[Zen]

Thanks a lot guys!  That was wildly informative and answered a lot of questions I didn't even know how to articulate! 

[earthtonesaudio]

See also:
http://www.muzique.com/lab/warp.htm