which diodes sound like what?

[vseriesamps]

hey team

is there a site anywhere where someone breaks down the clipping character of different diodes and in different configurations? it seems like it can make a lot of difference . . .

thanks so much
K

[brett]

Hi.  I've got V vs I curves somewhere.  But it's fairly simple, anyway.
The conduction curves of most diodes are generally* similar, so what mainly* separates different diodes is their clipping voltage - lowest voltage = "softest" clip.
Ge = 0.3V
Si = 0.7V
LEDs = 2 to 4 V.

About generally* - the gold bonded 1N4148 is softer than a 1N400X.  Probably why it's so popular.

This all assumes that you'll use some sort of volume control to equalise the output level.

cheers

[Steben]

Hi.  I've got V vs I curves somewhere.  But it's fairly simple, anyway.
The conduction curves of most diodes are generally* similar, so what mainly* separates different diodes is their clipping voltage - lowest voltage = "softest" clip.
Ge = 0.3V
Si = 0.7V
LEDs = 2 to 4 V.

But LED's seem to be softer...
?!?
Schottkies are also worth a try
Sh = 0.4V

[MartyMart]

I agree with Brett, i've always found that LED's give a harder more raspy drive when pushed into distortion but if your circuit allows you to "eq" a bit and maybe put a high roll off cap ( 47pf-150pf )  around them, they can be quite "tubey"  sounding.
3mm yellow mojo !
Marty.

[Ge_Whiz]

...and Schottkys clip much harder than germanium... :?

Hence "try them and see".

[Steben]

...and Schottkys clip much harder than germanium... :?

Hence "try them and see".

but less than Silicon, no?!

[Ge_Whiz]

Well, I wouldn't wish to be pushed for an opinion on that...

[brrt]

It all depends on the configuration. A diode clipping around an opamp sounds very different than the same diode shunting to ground...

[Peter Snowberg]

Much of the sound depends on how the diodes are used and what other components are "helping" them out.

First there is the forward voltage of the diode in question.... then comes any series and or parallel resistance and/or parallel capacitance.... and then the circuit topology as feedback clippers and output clippers each have a slightly different effect.

It's hard to make direct comparisons when different voltages are involved, but in general the forward voltage (Fv) is generally stated at a given current flow. This voltage changes depending on the current and the diode chemistry and you'll see the relationship between voltage and current described as a "knee".

Here's how I see the knee being affected by chemistry...

(soft) Ge - Si signal diodes - Si rectifiers - some LEDs - Schottkys - other LEDs (hard)

[Mark Hammer]

I think Heisenberg would have something to say about this....Peter did, and he is absolutely correct.

There really IS no "sound" of diodes, since diodes only produce clipping in context, and the context shapes the sound as much as, if not more than, the diode does.  Diode-class changes (e.g., Si to LED, *not* 1N914 to 1N4148) WILL alter the sound of a device, and it is possible to make suggestions about differences in tone, but the diode itself has no sound.

Because of their clipping action, diodes increase harmonic content; what we call fuzz or distortion.  In order to clip, the signal must be of an amplitude at or above the threshold set by the diode.  As you know theguitar/bass signal varies a lot.  Stick it on a scope and you'll see that there are peaks interspersed with lots of "hills".  The peaks stand a greater chance of being at the level which will result in clipping.  To make sure as much of the signal as possible is around or above the clipping threshold, ALL distortion units will insert some gain into the circuit so that the signal spends as much time as desired at or above that threshold.  In many respects, the distinctive sound of some units is a product of how much gain is inserted.  You can make *any* distortion sound more severe simply by sticking some clean boost ahead of it.

Because there is so much gain needed to take the puny signal from limp wrists strumming thin gauge strings and turn it into a hefty enough signal to produce clipping,  most circuits will have some sort of bandwidth limiting to maintain stability in the circuit.  This will include both trimming of low end in the process of blocking DC, and trimming of high end to keep oscillations out of the picture.  It is the bandwidth limiting, in conjunction with the amount of gain applied and the diode choice that produces the kind of harmonic spectrum associated with different kinds of pedals.  Remember, harmonic distortion is multiples of frequencies already in the signal, if the low or top end is restricted by filtering of any kind, then that part of the spectrum will be essentially absent (or low enough relative to the clipping threshold), such that multiples will not be added.

Indeed, this is the very premise of the "smooth" sound of a Tube Screamer.  Since bass notes are naturally a higher amplitude, hence closer to the clipping threshold anyways, the bass is filtered on the way to the clipping diodes such that the diodes do not add MORE multiples of those lower notes (and their overtones in the unaffected signal) than they add to notes higher up.  Filtering of the high end at several points also alters the harmonic balance such that much higher order harmonics that add buzz and fizz are removed.  The result is a different sort of tone than if all the bass and treble is kept in all the way through.

I have absolutely no doubt that if you examine the form of the clipped wave produced by feeding a tone generator through a back to back pair, you WILL see differences.  Guitars are not steady state devices/signals, though.  Diodes only clip what they get, and the player only hears what manages to make it out of the pedal.  Ultimately, the manner in which the low and high end is retained or filtered out at different points en route to the diodes has more impact on the tone than the diode choice does.

All of that being said, the same signal will produce less output volume and more uniform clipping across the note when diodes with lower thresholds are used.  If you want more output volume, you can either use higher threshold diodes like LEDs and push them harder (i.e., apply more pre-clip gain in the circuit) or you can use lower threshold diodes (e.g., germanium) and follow them with a gain recovery stage to compensate for the low threshold.  If you want more modest clipping, either use diodes with a higher threshold or more low-threshold diodes in series, or apply less gain.

[Phorhas]

Could be interesting if someone with an oscilloscope would test many type diodes in many types of configuration, under the same current/temp./voltage conditions, to see their clipping - take photos and post the result...

So - who's up to it?

[strungout]

Oy.

A little question: how does diode type affect the frequency content exactly? I thought different diodes (without any major filtering) clipped more in a certain frequency range... I mean Ge clips more lows, LED more mids, and Si more highs? That's the type of distortion I heard when I used each (thinking specifically about a hard clipping configuration).

Ciao.

[Peter Snowberg]

Bass notes have more voltage and current behind them so they will "adjust" the knee a bit more than high frequencies will, but the sound all has to do with the circuit that feeds them and how it responds to a given range of voltage.

Here's an experiment to try.....

Build up a feedback style clipper circuit and then make three strings of back-to-back diodes. The first diode pair is just a pair of yellow or green LEDs. Measure their forward voltage drop, and then make a string of Si signal diodes and one of Ge signal diodes that have the same forward voltage as the LEDs. You'll end up with a BUNCH of Ge diodes and that will make their “Ge” character effect more dramatic. With the same (roughly) forward voltages, now you can compare the response of the diodes in a consistent environment.

The clipping voltages of Ge diodes and of green LEDs is almost a 10:1 difference. That voltage change makes a big difference in the rest of the circuit if you simply change out one diode type for another.

[petemoore]

Interesting concept. If I had 20 Ge diodes, a bunch of Si's, and a some Ge's, and a scope I'd run some tests.
 Just swapping them in a given circuit, like I do, lead me to believe there's a tonal difference in the way the different material diodes clip...it's hard for me to accept it's purely a matter of voltage thresholds that cause the different tones...
 Someone that has many diodes a socket [in a clipper ckt.] could measure like thresholds by building Si, and LED strings, and Ge's that measure the same clippin voltage threshold,..I guess I haven't messed with diodes enough...I still have the strong tendancy to thing Ge's sound different...something about how fast they 'close' being different somehow...seems that the signal is so complex that ears might find nuances that scopes might not show [?]... ...

[Peter Snowberg]

For sure there is a variation in the character of the "knee". The harder the knee, the more odd-order harmonics generated.

I too would love to see some scope traces side by side with a spectrum analysis for all these different types. It still matters what's around the diodes, but getting the voltage environment to be the same will at least provide results that can be more closely compared.

[The Tone God]

I too would love to see some scope traces side by side with a spectrum analysis for all these different types.

Coming soon...

Andrew

[Peter Snowberg]

Coming soon...

Andrew

                 

[vseriesamps]

what about when you tie together the emitter and base of a germanium transistor and use it as a diode? does it behave much the same as a germanium diode?

hey i'm so excited - my first post that led to a real nice thread

thanks
K

[aron]

I know you wanted something more "scientific" but my simple mods page has had descriptions since forever:

http://www.diystompboxes.com/cnews/mods.html

Also, be aware that yes, diodes change the feel and the tone, but other parts of the circuit (tone control etc...) can have much more of an change on the overall tone.

[Mark Hammer]

For sure there is a variation in the character of the "knee". The harder the knee, the more odd-order harmonics generated.

I too would love to see some scope traces side by side with a spectrum analysis for all these different types. It still matters what's around the diodes, but getting the voltage environment to be the same will at least provide results that can be more closely compared.

Maybe so, but odd-order harmonics of what?  If the input signal were ruler flat from 20hz to 20khz, I'd buy into this.  But it's not.

Hell, I simply change the input cap on my Bosstone clone (thanks Aron!) and the quality of the distortion changes fundamentally in a very obvious way....for the low notes, but not for the higher ones.  I haven't changed the diodes but I HAVE changed what I feed them.