When placing diodes for asymetrical clipping...

[skiraly017]

...is there any way it should be done? Let's say we were using two 1N194's and a 1N270, the possibilities could be -

< 1N914  < 1N914
        1N270>

< 1N270  < 1N914
        1N914>

< 1N914  < 1N270
        1N914>

Is there a correct way to do would any of the combinations yield the same result? Thanks.

[Mark Hammer]

The forward voltages of each diode are approximately additive.  So if diode A has a forward voltage of 515mv, diode B has a forward voltage of 537mv, and diode C has a forward voltage of 222mv, then:

A+B -> / C <- gives  clipping at 1052mv/222mv

A+C -> / B <- gives clipping at 737mv/537mv

B+C -> / A <- gives clipping at 759mv/515mv

The first combo yields greatest asymmetry, then the third, then the 2nd.

As to which sounds better, that's a matter of your taste.

[Quackzed]

think of it as 2 separate thresholds of clipping.
so the 914 and 270 in series and the 270 and 914 in series (swapped) is basically the same thing and will have the same clipping threshold
where 2 914's in series will have a different threshold
and 1 914 or 1 270 will each have a lower threshold than any 'series pair' will..
id start with a 914 in one direction and a 914 + 270 in sereis in the other direction.
then try the 270 in one directon and the 2 914's in series in the other direction.
those are basically the 2 options you have for asymetrical clipping with those 3 diodes.

[skiraly017]

Got it, thanks!

[Mark Hammer]

The direction doesn't really matter.  If you can hear the difference between sticking 2 diodes in this direction vs that, you shouldn't be in the business of making dirty rock and roll.

DO note, however, that not all diodes have exactly the same forward voltage, and that it is worth measuring your diodes so that you can optimize whatever it is you are trying to do.  I find that Si types can vary between under 500mv to over 600mv, and Ge types can vary from under 200mv to over 300mv.  Increasing the difference between required forward voltage to conduct in one direction vs the other will determine the degree of asymmetry.  Note that the difference between paths is a separate matter from the absolute forward voltage.  I can produce a theoretically asymmetrical arrangement by employing 2 red LEDs in one direction and a single one in the other.  Will the signal ever clip in the 2-LED direction?  Not bloody likely.  Indeed, there may be no audible difference between having those two diodes...and not.

Sticking 2 Ge diodes (or a single Si) in one direction, and one Ge in the other, on the other hand, is similarly asymmetrical in the sense of needing roughly double the signal amplitude, but you won't need nearly the overall absolute signal level to produce clipping.  You should be able to easily extract clipping on both half cycles.

A simpler approach may simply be to use a standard 1+1 Si pair, and stick a variable resistance in series with one of the diodes to get variable asymmetry in clipping.  I've done it, and it works extremely well.