???symmetric vs asymmetric clipping???

[stm]

I agree with Mark's comments regarding the higher headroom for one half-cycle has beneficial characteristics other than the assymetry itself.

In this respect, going back to something simpler, like comparing the "sound" of Ge, Si and Red LED clipping, you cannot make a fair comparison just substituting one type with because of the different voltage threshlods.

One way I want to try when I have time is symmetrical clipping with Ge, Si and Red LEDs in the following manner:

Let's assume nominal forward drop voltages are: 0.2V for Ge, 0.6V for Si, and 1.8V for Red LEDs.  (Let's assume these are representative voltage drops. I know in practice Red LEDs can be around 1.6V, as Si can be around 0.55V and Ge around 0.15V; just let's stick to some representative values for the sake of making an initial calculation.)

Then, to have similar clipping levels, and as such leave out the "extra headroom" effect one could try the following three diode combinations:

1) One Red LED one way + One Red LED the other way (1.8V clipping threshold approx.)

2) Three Silicon diodes one way + three Silicon diodes the other way (3x0.6V = 1.8V clipping threshold approx.)

3) Nine Germanium diodes one way + nine Germanium diodes the other way (9*0.2V = 1.8V clipping threshold approx.)

Of course most of us don't even have 18 identical Ge diodes to try this out!  But at least I think it would make an objective way of defining the "sound of diodes" debate for once and for all.

Regards,

STM

[WGTP]

Excellent point.  I have seen discussions of the "feel" being better with assymetrical clipping, which may be more noticible than the harmonic content difference.  

I guess you could have an average clipping threshold.  2 5v diodes would average 5v.  A 5v and 2x5v combo would average 7.5v.  2x5v on each side would be 10v.  An LED would be more like 1.75v.  I have generally found 2 SI diodes as you said, overly compressed and lifeless.  An SI and LED, on the other hand, makes me want to Howl.  

Also, I guess you end up with a situation where the SI will be clipping, but the LED doesn't.  (I think you already said that).  

That is why I think the Mosferatu is cool.  It only uses a clipper on one side, so the other side doesn't clip until the op amp does.

[MR COFFEE]

Check out the Peavey patent if you want to think about what it takes to create assymetical clipping.

It AIN'T a clipping difference of 2:1 ----it's 40:1

It AIN'T one stage ------it's 4 stages in series with 2:1 attenuators in between.

My breadboard experience and design listening agree.

Mark makes a good point also - the diode/clipping threshold in the TS design or the Boss design is relative to the virtual clean signal in parallel with the clipped signal because of the follower configuration. BTW, that isn't the only circuit topology to parallel a clean signal and distorted one. Other topologies allow different tone shaping for each path which adds a lot of possibilities tone-wise.  :idea:

[WGTP]

Does the PV circuit clip the same side of the signal each time, or does it alternate?  

[Fret Wire]

In this respect, going back to something simpler, like comparing the "sound" of Ge, Si and Red LED clipping, you cannot make a fair comparison just substituting one type with because of the different voltage threshlods.

One way I want to try when I have time is symmetrical clipping with Ge, Si and Red LEDs in the following manner:

Let's assume nominal forward drop voltages are: 0.2V for Ge, 0.6V for Si, and 1.8V for Red LEDs.  (Let's assume these are representative voltage drops. I know in practice Red LEDs can be around 1.6V, as Si can be around 0.55V and Ge around 0.15V; just let's stick to some representative values for the sake of making an initial calculation.)

Then, to have similar clipping levels, and as such leave out the "extra headroom" effect one could try the following three diode combinations:

1) One Red LED one way + One Red LED the other way (1.8V clipping threshold approx.)

2) Three Silicon diodes one way + three Silicon diodes the other way (3x0.6V = 1.8V clipping threshold approx.)

3) Nine Germanium diodes one way + nine Germanium diodes the other way (9*0.2V = 1.8V clipping threshold approx.)

I think you'll find that the three different types will sound different even though they'd have the same 1.8v threshold.

I noticed a while ago that 1N270 Ge's and 1N34a GE's sound different. The 1n34a's have more bass/dulled treble compared to the 1N270's. This was with a two diode symmetrical clipping ckt. The threshold voltage is nearly indentical, the composition material (Ge) is the same. I have a large quantity of each and the results were consistant. I thought I was hearing things, but I sent some 1N270's to Phillip Bryant when he was testing the Liquid Drive, and he heard the same difference.

With TS type ckts, I hear a difference between 1N4148/1N4148 and 1N4148/1N400x combo's. Again, not much of a voltage difference, but a composition difference.

Whether it's really asymmetrical clipping or not, the composition also seems to make a difference. So I think that three different diode types that have the same threshold voltage, will still sound different.

BTW, here's some voltage averages of 25 each type diodes I measured awhile back:

1N4148 - .604
1N4001 - .565
1N4005 - .558
1N4007 - .566
1N270 - .260
1N34A - .272
1N60 black case - .264
1N270 black case - .252

[dev]

I think the best way I can discribe the difference in a TS type circuit is as follows.

asymmetrical clipping:
increases the pedal's maximum volume (db's)
causes less distortion (less signal clipping)
causes less compression
at the same volume (db's) there is less background hiss
                             
                             
symmetrical clipping:  
decreases the pedals maximum volume (db's)
causes more distortion (more signal clipping)
causes more compression (especially with the gain maxed out)
at the same volume (db's) there is more background hiss

So there are the main trade offs.  

placing a pot in series with one diode of a symmetrical pair of diodes can alow you to dial in asymmetry.  

Hope that helps
Dev

[MartyMart]

Paul, good to see we agree.

Does anybody have any opinion on 1N4148 & 2x1N4148 v/s 1N4148  & 1N4148+1N34  for assymetrical clipping?

STM,
I LOVE the 2x1N4148 and 1x1N34a combo, have used it in my "SD-1 Blues" mod  very sweet sound  !!
1x 1N4001 and 1x 1N34a is also a nice sound
assymetrical floats my boat every time  

Marty.

[TheBigMan]

I tend to go for the 1N4148/1N400x combo in TS type pedals.  It sounds a bit sweeter and smoother to my ears.  However I lent a modded TS-5 to a friend and he doesn't like the combo.  With his setup he gets problems with octaves ringing out when he plays doublestops.

[MR COFFEE]

WGTP,

The Peavey circuit inverts the signal in each stage so clipping "alternates" which side is clipped.

The significance of the four stages in series, IMHO, is that a least one stage is going to be clipping assymetrically at nearly all signal levels.

And very assymetrically, just like a real tube circuit. The patent was based on devising a solid state circuit that has a good overdriven "tube" guitar sound. Peavey 5150 amps use four 12AX7 stages in series (with attenuators between stages that cut the signal by about half) to get their overdrive sound, and the circuit does basically the same thing without wasting all that filament power. The patent shows a JFET version and one using darlingtons with diodes to clamp the signal voltage just like grid conduction.

Damn I like surfing the patent office... ALMOST as much as hanging out here. The conversation is a lot better here at Aron's though <grin>