questions about diodes shunted to ground and capacitors....

[jishnudg]

Hey,all!

So I was thinking of some circuit snippets I've seen on the web that use hard clipping diodes after gain stages in funny ways -

1. signal shunt --> diode pair ---> capacitor ---> ground

Example here -

http://2.bp.blogspot.com/_y4AYtND8Hz8/R7v3RaU_omI/AAAAAAAAAHI/C7E23DPLUZo/s1600-h/LORT.jpg

2.  Signal shunt -> capacitor (non polarised small value) --> diode pair --> ground

(ie if the order of cap to ground and diodes were reversed)

3. capacitor in parallel with diode shunt to ground

Eg - here - - http://3.bp.blogspot.com/_y4AYtND8Hz8/R72W3KU_osI/AAAAAAAAAH4/pwKc0P3ST5c/s1600-h/cot50briggsmods.gif

In all these cases it seems like using the capacitor in conjunction with the diodes is giving some sort of a voltage dependent low pass filter...? How is this working, exactly? Also, in such a scenario, what would happen if we took a tone control like the BMP tone control, and instead of the capacitor and the resistor (in the the low pass and high pass filters respectively) going directly to ground, we stuck a diode pair in there...? Would it work differently from Jack Ormans "tone clipping" circuit snippets?

Link -

http://www.muzique.com/lab/tclip.htm

Also, do these principles also apply to diodes pairs and caps in feedback loops? I've seen a lot of circuits with diode pairs and caps in parallel in feedback loops, but only one where a cap is used in series (Jack Ormans muffer)

[GGBB]

I would think of it the other way around - frequency dependent clipping - but I suppose you could describe it either way (1 and 2 at least).

Numbers 1 and 2 are essentially the same - any frequencies that pass through the cap are clipped at the output. So instead of the cap shunting certain frequencies to ground, it only shunts the portion of the wave for those frequencies that is over the diode Vf. Therefore those frequencies get clipped - lower ones don't since they cannot pass through the cap to ground whether the cap is before or after the diodes.

Number 3 is nothing more than a low pass filter plus clipping. Ignore the fact that the cap and diode pair are in parallel and look at it as two separate things: the cap forms a low pass filter and the resulting low-passed signal is clipped. So just normal clipping, but the signal being clipped has been filtered.

At least I think that's what's going on. I can't explain how the actual frequencies are calculated for 1 and 2 - not sure if the diodes have an impact or not.

[GibsonGM]

My understanding is that those caps bypass high frequencies, both before the diodes conduct, and possibly AFTER they've conducted and clipped, creating all these often-nasty high harmonics that aren't even multiples of the fundamental.   Things like high 3rds, that sound harsh...the idea is to shunt that stuff so we don't have to listen to it!

When not conducting, the diodes are a very high resistance.....and almost zero resistance when on.  In either case, I think the cap is just picked for its reactance and the diodes don't really figure into that.  The 'stuff' we are interested in getting rid of isn't generally that prominent prior to clipping....  In the 2nd link, the 47n cap has a reactance of ~170 ohms at 20kHz, ~338 ohms at 10kHz, but ~3,380 at 1kHz.  If there were an UNISOLATED driving impedance (output impedance of driving stage), that WOULD factor into the filtering action.  So would a following INPUT impedance... 

Maybe somebody has a better take on it?  This is just how I've always thought of those bypass caps...