Audio comparison clips of various clipping diodes

[Killthepopular]

I breadboarded the electra circuit and tried out a load of different diodes. I recorded all the clips under lab conditions, more or less. Should give a reasonably good sense of how some of the more common diodes compare sonically. Hope this is helpful.

https://youtu.be/xHhWFdMY4rU
https://youtu.be/MkvSTcwu5Dc
https://youtu.be/jUU_OPEQTRs

[Mark Hammer]

I'm not really hearing very much difference if at all.  I think you need to increase the gain so that threshold differences start to emerge.  Admittedly, I'm listening through very puny speakers in my monitor, but the volume level seems more constant than I would expect from the different diode complements used.

[iainpunk]

I'm not really hearing very much difference if at all.  I think you need to increase the gain so that threshold differences start to emerge.  Admittedly, I'm listening through very puny speakers in my monitor, but the volume level seems more constant than I would expect from the different diode complements used.

youtube tends to add a ton of compression to the videos uploaded. thats probaby why

[Mark Hammer]

That's kind of what I figured.

[Killthepopular]

the volume level seems more constant than I would expect from the different diode complements used.

Do you mean that you would expect say germanium diodes to be quieter than LED because of the difference in headroom? I normalised all the tracks so that the volume level would be consistent. I also adjusted the gain and volume with each set of diodes so that i got a "Break up" tone and the same output level. The amp was set very clean. 100% of any dirt you hear on these clips is coming from the diodes (and possibly the transistor) not the amp.

I just wanted to hear the difference in tone and clipping characteristics. I wasn't really interested in the difference in gain/output that I would get simply from different amounts of headroom.

I personally hear a big difference between arrangements. Germanium is fuzzy and natural sounding but also sound a bit small. 1n4148 sound big and solid and the clipping is good but the tone of the clipping seems to add a lot of shrillness in a way that the other diodes don't seem to. 1n4001 have a less natural sounding clip but the tone seems to be fatter and the clipping somehow sounds less shrill and more focussed in the lower mids. LEDs somehow sound a bit thinner than silicon diodes, with a very unnatural sounding clip but without adding shrillness like 1n4148. Asymmetric setups are better at giving a subtle breakup distortion with more dynamics whereas the symmetrical combinations tend to have a more on/off distortion sound and fewer dynamics.

[Mark Hammer]

Maybe my hearing is shot, or maybe my patience for discussion about diode clipping is shot, but if you took the screen titles away, I wouldn't have had any idea what I was listening to, or whether anything had been changed.

Personally, I think the question one always has to ask and answer is: how much of the signal, and for how long, does it remain at or above the forward voltage of the diodes, at the given gain setting.  With a lower forward voltage, using the same gain, Schottky will continue to conduct for some time after germanium have stopped conducting, and germanium will conduct for some time after silicon have declared "Nope, not doin' it". 

Guitars are generally a kind of percussion instrument.  You hit the string and it vibrates, but dies out shortly after, changing not only volume but harmonic content as well; just like a drum.  Diodes have a fixed forward voltage.  If used to hard clip the signal what you hear is a function of what portion/s of the signal result in the diodes conducting.

In this scope image of a guitar signal (taken from here: https://www.keysight.com/main/editorial.jspx?ckey=866712&id=866712&lc=eng&cc=AW ), we can see that drop in amplitude.  If the clipping diode's forward voltage is equal to whatever the second horizontal vernier line above the midpoint is, then any clipping will be very very brief.  If the forward voltage is equal to whatever the first horizontal vernier line is, then the clipping will occur over a longer portion of the note's lifespan.  But since you can see that the string does not vibrate consistently above or below that forward voltage, the diodes won't conduct here and there, even though most of the signal exceeds the forward voltage.  Finally, the remainder of the note falls below the forward voltage, and cleans up.

If one adjusts the gain of the circuit feeding a pair of clipping diodes, and controls for the current  in the signal, one can effectively mimic any diode with nearly any other diode.  So, while I appreciate your efforts to provide comparisons that are "uncontaminated" by volume levels, doing so makes them all sound the same...at least to my ears.

If I make anything with a switch to select between diode types, silicon is always "louder" than germanium, and germanium is always louder than Schottky.  Or rather, a pair of one is louder than a pair of the other.  As you have no doubt learned, sticking several of a lower Vf in series gets you what a single one of a higher Vf gets you, assuming the Vfs add up to the same total.

One of the gripes people had about the old MXR Distortion+ was that, while they loved the tone, the damn thing wasn't very loud, unless you dimed gain and volume.  The DOD250 helped with that a bit by using essentially the same design, but silicon diodes instead of germanium, lifting the ceiling on the maximum output level, and roughly doubling the output level.