Diode/Notch Filter Experiment

[WGTP]

http://aronnelson.com/gallery/WGTP/DioNotch

Check this out and see what you think it will do.

I'm hoping to vary the asymetry of the distortion at different frequencies while adding a shallow Marshalesque notch to the response.  

[MartyMart]

That looks cool ! could you use 20k trimmers instead of the 15k's for some "notch" variation ?
Or a 47k pot A'la Mark hammer :
Wiper to the diodes/ground connection and lugs 1&3 either side of the 0.1
cap ... ?

Marty.

[puretube]

That looks cool ! could you use 20k trimmers instead of the 15k's for some "notch" variation ?
Or a 47k pot A'la Mark hammer :
Wiper to the diodes/ground connection and lugs 1&3 either side of the 0.1
cap ... ?

Marty.

ahem: MH&PT

:lol:

[Mark Hammer]

Being a card carrying, dyed in the wool, stark raving loony when it comes to double clipping, because of the way it allows one to shape the clipping quality, naturally this appeals to me immensely.  This little trick (and it really seems to be quite a day for clever little tricks, doesn't it?) accomplishes part of what double clipping does, with much less effort, and obviously fewer parts and, dare I say it, no current consumption.  Kudos.

Here's where I think you need to *maybe* re-engineer a bit, though.  The notch will undoubtedly reduce the overall amplitude of the signal that the SI diode is intended to clip the other half-cycle of.  In a perfect world, that shouldn't really matter, but the problem is that where there is a lot of choice in resistor and cap values, there is much less choice in diode clipping thresholds.

Wisely, you elected to stick an LED "first", presumably because you anticipated the drop in level over the notch, and the need for the second diode to have a lower clipping threshold.  The $64k question is whether the difference in threshold is enough.  On the other hand, it is an empirical question, and presumably there is no requirement to use the values shown, and any of a wide variety of values could obtain a similar or maybe even preferable notch, in such a manner as to still result in clipping from diode 2.

The trick would seem to be avoiding the sort of asymmetry that arises from having clipping at diode 1 and none at diode 2.

Finally, smoke on this for a bit.  Gain stage 1 pumps signal through the network you show/drew (much like a twisted sort of Dist +), and then goes to a second more conventional clipping stage à la TS-9.  If you wanted, you could also use the simple tone control (SWTC) I posted last night between the diodes, rather than a fixed notch, so that there is some variable lowpass filtering between the diodes.  That might be interesting too, although my guess is that it will sound MORE interesting if one uses it to condition a signal in preparation for a second active clipping stage.

But once again, neat idea, and sure to find its way into a lot places as a basic circuit fragment.  Good on you, lad!

[WGTP]

As I was reading about Mark's tone control, it occured to me that there were posiblilities for it with this deal.  Great minds...  

I got really twisted and started thinking about a different EQ for each diode, in addition to various thresholds.  Using an LED for one and an SI with a cap before ground, for example.  Bass boost with less distortion?  Or a cap between the diodes to reduce bass and make it less symmetrical or more depending on the diodes used.  Hmmm  

I seem to like my treble assymetrical and my bass symetrical.   :twisted:

[Mark Hammer]

I got really twisted and started thinking about a different EQ for each diode, in addition to various thresholds.  Using an LED for one and an SI with a cap before ground, for example.  Bass boost with less distortion?  Or a cap between the diodes to reduce bass and make it less symmetrical or more depending on the diodes used.  Hmmm  

I seem to like my treble assymetrical and my bass symetrical.   :twisted:

Great minds indeed!!

What you have here is an opportunity to tinker with the degree of clipping generated (for halfcycles) forthe entire passband, and different parts of the passband.  You haven't mentioned it, but note that yu can also stick a diode in parallel with the cap to ground to clip the lowpass filtered signal available at the junction of the two resistors.  Hell, for that matter, stick an LED up front, a similarly oriented SI in parallel with the cap to ground (double clip the low end) and a reverse polarity GE at the output of the notch.

Lots to play with there.

[WGTP]

I guess, you could use an LED pair before the filter, and an SI pair after and get dual symmetrical clipping.  By setting up the op amp really bright and using the notch to reduce the highs and upper mids, you would be clipping the highs more than the lows???  

If the signal is reduced too much, you can use a dual op amp at the output for signal recovery.  

[WGTP]

Since when did July become the wet season in the mid-west?

While waiting for the rain to let up, I tried increasing the 1K resistor to 10K and 100K and it seems to be rolling off the treble.   Hmmmmm   :twisted:

[WGTP]

http://aronnelson.com/gallery/WGTP/DioNotch2

In the spirit of stupidly wonderful...

[Mark Hammer]