Clipping diodes exposed

[commathe]

    i experimented at LENGTH (oh yes) with diode ladders and came to the conclusion that you dont have to rely on the smallish diode knee. you can make your own knee out off strings of diodes and resistances. VERY soft clipping knees if you want em. people seemed to worry about intermodulation distortion, but i gotta tell ya, intermod is inherent in soft clipping. the bigger the knee, the more apparent it is.
   this route is the only one i've found other than ts-type feedback loop clipping and really small signals 'barely' hitting diodes to ground type clipping that does this (ok tubes but thats cheating)  both of those are 'soft' type clipping but the ts style is limited by the diodes and -gain of one- above the threshold and the other is limited by the diode threshold and noise due to the signal needing pre attenuation...
      the diode ladder or better series strings of diodes/resistances to ground gives you alot more control over how steeply you clip the signal at each higher threshold... imagine just 5% dist for .6v 10% for .8v 20% for 1v 40% for 1.2v etc... a consecutive clipping slope. it can be done. you need a bit of extra headroom to avoid hitting the rails but nothing crazy... you can even tailor in separate odd/even order slopes. sounds like bs i know, and its not perfect, but you can do it.

I'd be really interested n a schematic. I've been trying to do something similar for ages but never got it to work

[R.G.]

    i experimented at LENGTH (oh yes) with diode ladders and came to the conclusion that you dont have to rely on the smallish diode knee. you can make your own knee out off strings of diodes and resistances. VERY soft clipping knees if you want em... you can even tailor in separate odd/even order slopes. sounds like bs i know, and its not perfect, but you can do it.

Yep. It's the old-school way to do arbitrary waveshapers. It's more accurate with transistor clamps than with diodes, as the transistor characteristic sharpens up the changeover points.

[Tall Steve]

The 4001s will have more junction capacitance, which might make a different if you have a 1Meg feedback resistor and no other capacitor there.

Okay, cool. This is something I've been wanting to understand better. Does the relationship between junction capacitance and frequency response stem from the slight continuation of current flow through the cathode as the diode changes states? For example, say a diode in a clipping circuit to ground has an internal capacitance of 15pF, and it is oriented so that it clips the positive half of a sine wave input. During the negative portion of the wave, is the leakage current that continues after the diode has gone into cutoff akin to a 15pF capacitor discharging to ground?

[tca]

    i experimented at LENGTH (oh yes) with diode ladders...

Look for diode function generators and analog computers. Tried this: http://www.diystompboxes.com/smfforum/index.php?topic=100710.0

Cheers.

[Digital Larry]

Well, it is *only* my opinion, which is worth what you paid for it.

Well, I'm willing to expose myself as a semi-sycophant and indicate my general agreement.  I think Ge vs. Si shows itself best where your passive guitar goes directly (or through a cap) to the transistor whose emitter is grounded.  If you start biasing things and using negative feedback to reduce the active device's non linearities, then why use Ge at all?

I think it was a fortuitous alignment between Ge's voltage gap and the typical output levels of passive guitar pickups.

But as the other guy said, there is the possibility of touch sensitivity beyond the knee. [Cue "Don't Touch Me There"]   If we could more accurately describe those things we could attempt to reproduce them. 

Interesting thread - my favorite type in fact.

[bool]

One Touch Too Much

[jishnudg]

Just saw this thread and it answered a ton of questions. Just one more - is there anything that would clip at a lower voltage than an Si or 1N400x? All other factors being equal.

[jishnudg]

Also, assuming that we're using a typical LPB setup, with a voltage gain of 8 to 9...if the voltage of the input signal is anywhere from 500mv to 1V, the output signal would vary from 4.5v to 9V. So if I stick the following diodes after the LPB gain stage, I'm looking at 2 factors - forward voltage (when it clips) and the current-voltage characteristic (how it clips)

1. MOSFET as Diode - high forward voltage, soft clipping
2. Red LED - moderately high forward voltage, middle ground
3. Si 4148 - low forward voltage, hard clipping
4. Ge - very low forward voltage, soft slipping

What if I want clipping at a lower voltage than a Ge, but no soft knee?
What if I want MOSFET as diode clipping characteristics, but at a much lower forward voltage?

Taking our peak output voltage as 4.5V to 9V, how can I calibrate my diode setup so I have Ges clipping when the signal hits 4.5V (say put 2 Ge diodes in series?), have LEDs clipping between 4.5V and 7V (the Ges would clip too but it wouldn't matter) and have smoother Si clipping when the signal peaks at 8-9V?

Also, if we replicate the above setup, could we have a succession of cascading stages, each boosting the signal by a small amount, and clipping it gently, would that approximate a tube amplifiers breakup to some degree?

[anotherjim]

Just thinking out loud, but if it's a tube-like S-curve soft clip you want, why not use a CMOS CD4007UB complementary pair inverter set up as linear amp? You could control the supply voltage to the Vdd pin to vary the headroom from 3 to 9volts (use the pair that are internally connected to Vdd and Vss), also fit a limiting resistor to reduce the high supply current you normally get with a linear mode CMOS inverter. You also get 2 spare mosfet pairs to play with that can also be used as independantly (either P or N channel). Shame the gates are tied common to P and N's.



I'm building an "amp sim" idea I've had for ages using only the 4007 mosfets. It sounds very promising so far. But why is it not commonly used? Hmmmm...

[amz-fx]

Here is an experiment that I have planned for a long time, but never got around to trying it yet. What are the differences in sound of the following two clipper circuits?



X is strong enough that the signal is clipped, all diodes are matched and there is no dc bias on the input. Do they sound alike, or are there some small differences? What if the signal level is reduced so X barely clips the two diode version?

regards, Jack

[teemuk]

I'm building an "amp sim" idea I've had for ages using only the 4007 mosfets. It sounds very promising so far. But why is it not commonly used? Hmmmm...

Likely because there are so many other ways to achieve clipping too...

But not widely used...? I may have to disagree. Laney's still making preamps based on CMOS overdrive. Since late 1970's Lab Series, Sunn, Fender, KMD have all introduced amplifiers or complete series of amplifiers with CMOS inverter overdrive. Everyone should know Craig Anderton's pedal design but a much, much newer example of CMOS inverter utilisation is GizmoAudio's "Sawmill" pedal, which mimics a push-pull tube amp with two CMOS inverters operating in push-pull and having their output signals mixed with a differential amplifier. Due to gate-drain protection diode there's also a "built-in" emulation for grid conduction of tubes. And due to DC bias shifts introduced by extended overdrive the signal even begins to portray crossover distortion, just like in tube amps.

[anotherjim]

Teemuk, it's the 4007 I don't see being used much. I know of plenty with 4049UB or 4069UB. Most of those designs only use 2 or 3 out of the 6 available inverters, then are cursed with a high supply current, and if limiting it with a resistor, you get interaction between the stages (which could be good I suppose). The 4007 allows you to supply the pairs separately or use some as single transistors.
Anyway, I think you reinforced my point, that there are better ways than networks of diodes to produce tube style distortion.

[amptramp]

For INFINITELY VARIABLE CLIPPING DIODES Go here:

http://www.ti.com/ww/en/bobpease/assets/AN-31.pdf

on page 17, lower left corner, you will see a precision clamp circuit that prevents a signal from going above a limit that can be set by the user (Eref).  This diode has a series resistance equal to Rin so you can set diode voltage and resistance separately to any value you need.  Reverse the polarity of Eref and the diode to clip the other half.  Since this clips to ground (or virtual ground), it would work on a DOD250 but not a tube screamer.  You could have four controls to set each parameter separately.

The example uses an LM101A which is glacially slow and requires external compensation.  For faster op amps with internal compensation, C1 and D2 can disappear.

[PBE6]

+1 amptramp

Very cool, will have to give this a try! I imagine using a pot to set Vref would let you create the ultimate saturation control for a hard clipping circuit.

[Steben]

For INFINITELY VARIABLE CLIPPING DIODES Go here:

http://www.ti.com/ww/en/bobpease/assets/AN-31.pdf

on page 17, lower left corner, you will see a precision clamp circuit that prevents a signal from going above a limit that can be set by the user (Eref).  This diode has a series resistance equal to Rin so you can set diode voltage and resistance separately to any value you need.  Reverse the polarity of Eref and the diode to clip the other half.  Since this clips to ground (or virtual ground), it would work on a DOD250 but not a tube screamer.  You could have four controls to set each parameter separately.

The example uses an LM101A which is glacially slow and requires external compensation.  For faster op amps with internal compensation, C1 and D2 can disappear.

Old thread. But +1 on this. All classic diodes are a combination of logarithmic transient and a resistance. Fiddling with resistance results in clones of diode curves.
Yet! In a (non-inverting) feedback loop this changes. The transient is determined by impedance ratios far beyond the typical values of diodes. In comes design and ratio, rather than soft knee specs. A Tubescreamer/OD-1/.... is very sensitive to the diode choice by the knee voltage.  If such a circuit is set to maximum gain, the result is the original signal on top of the knee voltage. That's why - with same input levels - single germanium sounds dull and LEDs very tight. A stack of LEDs will sound hard clipped (if the supply rail is set high enough).
Most "marshall" sounds are achieved with harder clipping than many admit.