LED vs diodes in feedback loop

[Clint Eastwood]

Hello everyone,

Imagine a gain stage with clipping diodes in the feedback loop, what (if any) would be the the difference in sound between having three silicon diodes in series and a red led? I know, I should just try it out, but I am curious what y'all have to say about it.

[Mark Hammer]

Not an awful lot, actually.  Three diodes in series is roughly 1500-1600 forward voltage, which is approximately what a red LED has.

[MikeA]

To my ear, LEDs have a fizz as the note trails off.  I don't hear it with an equivalent Vf from multiple diodes.  I don't know the mechanism, but it seems to happen just as the signal level is falling through the LED's Vf level.

[Locrian99]

To my ear, LEDs have a fizz as the note trails off.  I don't hear it with an equivalent Vf from multiple diodes.  I don't know the mechanism, but it seems to happen just as the signal level is falling through the LED's Vf level.

I've noticed this with the one pedal I've built with led clipping (TS with switchable clipping one option being symmetrical red leds)

[Fancy Lime]

LEDs tend to have a sharper knee, meaning a more abrupt change from non-conducting to fully conducting, compared to the "usual" Si diodes like 1N4148, which have a softer knee and thus smoother transition into and out of clipping. The sharper knee produces higher order harmonics at low degrees of clipping, which we hear as fizzyiness on decaying notes. In most circuits that makes no audible difference but in some it can. An appropriately sized cap parallel to the diodes can be used to kill the fizz or adjust it to your liking. Choosing the currents through the diodes right by adjusting the impedance of the feedback loop is an even better method but more experimentation is necessary to get it just right.

[antonis]

LEDs tend to have a sharper knee, meaning a more abrupt change from non-conducting to fully conducting, compared to the "usual" Si diodes like 1N4148, which have a softer knee and thus smoother transition into and out of clipping.

Quite right but, for the vast majority of respective circuits, there is a cap across feedback loop which "softens" enough transition knee..

@Clint Eastwood: There are countless threads conserning "diode clipping" like the one below:
https://www.diystompboxes.com/smfforum/index.php?topic=125795.0

[merlinb]

There isn't much tonal difference for a given level of distortion, LEDs don't really have a much sharper knee, not significantly anyway (it tends to scale with forward voltage). But the higher forward drop of LEDs means you can't access 'as much' distortion for a given gain from the opamp. Basically LEDs sound the same as "diodes at a lower gain setting".

[Clint Eastwood]

Thank you all for your answers.  So there does not really seem to be consensus, some say there is no (audible) difference, others hear 'fizz'. I had read about that fizziness, but thought it was when using the led as clipper between signal and ground.
Alas, I shall have to experiment, and will report the results.

@Clint Eastwood: There are countless threads conserning "diode clipping" like the one below:
https://www.diystompboxes.com/smfforum/index.php?topic=125795.0

I will study it. But exactly because there are so many threads, it is hard to see the forest for the trees..

Best regards,

Clint

[Phend]

Youtube Clipping Diodes and get a bunch of demo's

[deadastronaut]

to my ears leds sound more open, and diodes sound a bit compressed.

but it may be just my old ears....

[Mark Hammer]

Thank you all for your answers.  So there does not really seem to be concensus, some say there is no (audible) difference, others hear 'fizz'. I had read about that fizziness, but thought it was when using the led as clipper between signal and ground.
Alas, I shall have to experiment, and will report the results.

@Clint Eastwood: There are countless threads conserning "diode clipping" like the one below:
https://www.diystompboxes.com/smfforum/index.php?topic=125795.0

I will study it. But exactly because there are so many threads, it is hard to see the forest for the trees..

Best regards,

Clint

And the reason for there being no real consensus is because they are never heard in complete isolation.  There is always some form of filtering before and after.  If not within the circuit itself, then certainly by the frequency response of whatever amplifier and speakers one happens to listen through, and even the cable connecting the last circuit in line to the amp.

So, there is likely to be SOME context in which a person swaps out diodes for LEDs and proclaims "Wow, what an improvement!" or perhaps "Jeez, that's awfully harsh!".  Could any of use here be able to accurately predict whether builder X's individual context will lead to the one declaration or the other?  I doubt it.  Try them, and see if you prefer it, but don't have any expectations about which way things should go.

[ElectricDruid]

Youtube Clipping Diodes and get a bunch of demo's

...and a load of nonsense commentary!

[Phend]

^ And I agree 100%. 
From what I have heard in my experimenting is that between Working Pairs of different diodes don't expect to hear a Night and Day difference. Do expect to hear the clipping effect. In the Harmonic Perculator build, I did settle on the two diodes I think sound the best (using my limited selection). So yes there is that "better" sound.  I am building a TS 808 now and will have a diode switch. Shall see or shall I say hear.
BTW the Harmonic Peruclator turned out to be great. Put quite a few hours in it.
(And expect volume differences)

[amptramp]

If you are using a breadboard or some other board and you are testing this circuitry in the light, be aware that an LED operates as a photodiode.  In my previous existence (before retirement), I was working on an optical touch panel that used LED's on two adjacent sides of the display bezel and photodiodes on the other.  We thought we could reduce the effect of sunlight by using green LED emitters and green LED photodiodes so infrared (which is a large component of sunlight that plays havoc with silicon photodiodes).  LED's work very well as photodiodes and are sensitive to the light wavelengths they emit but the leakage currents are high and vary with temperature.

If your design in inside a light-proof enclosure, it may work better than if it is out in the ambient light.  Some people have put the feedback LED's on the box so they get a visual indication of clipping, but this also means that the signal will be modulated by ambient light.  If you are playing under light sources that have a variation in output during the AC line voltage cycle, the 120 Hz hum will be impressed on the LED and on the clipping voltage.  Even if you are in daylight, varying levels will result in different clipping characteristics.  If you use an LED in a feedback network, paint it black or keep it in darkness.

This may account for the differences some of you are reporting with the use of feedback LED's.  You can always use an optocoupler LED (which is encased in black plastic) and use the output transistor to drive a LED for indication purposes.  That gives you the best of both worlds - indication of when clipping starts but with no effect on circuit action.