I had an idea!!! Biasing diodes...

[Hal]

I was going to first try it myself, but I thought i'd share.

Last night when i couldn't fall asleep, I was wondering..."well, you can bias transistors, so why not diodes?"

The idea is to set bias voltage on a diode.  For example, when using a normal silicon diode (1n4148), put somewhere around .6 volts across it.  If logic works, this would allow the diode to clip with less voltage, and give a harsher clipping sound.  Adjust a pot to lower the voltage across it, and you can clip less easily.


Would this work ?  Anyone tried it?

[Fret Wire]

Are thinking along the lines of, say you have two 1n4148's and you'd like to bias one, to act like a ge 1n34a? Along those lines?

[Hal]

pretty much :-D

I thought of a way you could get both with voltage across then, but i forgot  :lol:

it was like...2 am.

[R.G.]

Good idea!

That's what the "limiter" circuit on the old Thomas Organ Vox amplifiers did in the 60's. This is what I put up on GEO as "Mystery Circuit" number 1. The trick is actually to feed a trickle of current through the diode so that it's already partially turned on, and the trickle circuit doesn't eat any signal on its own, but the same idea applies.

You can see the same trick in the Foxx Tone Machine and Green Ringer where diodes are fed a little current to turn them on slightly, although in these pedals the objective is to minimize the turn on voltage, not to get a variable clipper.

The simple way to do this is with a complementary current source and a big capacitor tying the current sourced ends together, but the Thomas circuit actually did it with a single current feed and the diodes in series for DC but parallel for AC.

[Mark Hammer]

Not to rain on anyone's parade, but I do believe that this is incorporated as part of Jack Orman's "Warp controls" document over at AMZ (http://www.muzique.com/lab/warp.htm).....or maybe you're thinking of something different.

[Transmogrifox]

I have thought of biasing diodes before, but more the other way, where it looks like the effect of turning off the BJT.  You bias the diode, which is in series with the signal path.  When the signal voltage reverse biases the diode, it turns off, thus clipping the signal, however, as the signal begins to further the diode, the signal path becomes more linear.  You could incorporate two stages of this with an amplifier in between to clip both signal polarities.  It may look like another cheap vacuum tube amp emulator trick.

[R.G.]

Not to rain on anyone's parade, but I do believe that this is incorporated as part of Jack Orman's "Warp controls" document over at AMZ

I don't think so.

If I read the note right, he's talking about DC biases to change the effective threshold of a diode, while the warp controls stuff is a catalog of ways to use a blend control to change the minimum forward resistance of a miscellaney of diodes when turned on, no biasing involved.

[zachary vex]

i'm working on a new octave design right now, and it was necessary to bias the rectifying diodes with DC to smooth out the decay tail of the signal, which is really nice and pretty now.  without the biasing, it was a very gated cutoff with some sputtering, so obviously dc biasing on diodes can have great use... just isolate them with blocking caps and set up a dc flow from positive to ground through the diodes using moderate-valued resistors (but not so low that they load down your audio circuit).

[Peter Snowberg]

I thought this up a while back but never did anything with it.

Code Select Expand

                   Vr1      Vr2
                    |        |
            +---||--+--->|---+
         In |                |      Out
         ---+      Vr3       +---||---
            |       |        |
            +---||--+---|<---+

             Vr1         Vr2         Vr3
              |           |           |
V+ ---\/\/\/\--+--\/\/\/\--+--\/\/\/\--+--\/\/\/\--- GND


Imageing a filter cap and series resistor between each Vr supply and the Vr points around the diodes.

The next step would be to replace the middle two Vr supply resistors with pots to adjust the bias.

Take care,
-Peter

[Hal]

I didn't think it would be origional :-D....

But i wonder how it would sound with a simple circuit, like Joe Davisson's "simple overdrive."  I'll try it, when i can...i've been so crazy busy lately.


And yea, putting a cap between the 2 diodes should be the best way to put different bias on each one.

[Fret Wire]

Sounds like something you could use a trim pot to adjust. You could adjust two 1n4148's to both sound like 1n34a's, or just one, or both to sound like LED's, or just one LED and 1n4148, and so on. External user adjustable knobs for each diode. It could get out-of-control adjustable, just what we like!

[brett]

Ge has some some fundamental reasons other than clipping threshold why it is softer for clipping than Si.  I *think* it's about valency and doping and energy levels and such.  In any case, Ge has a slower rate of change in conduction per unit voltage.  Gold-doped Si devices (eg a 1N4148) are also slower/slopier/smoother than regular Si (eg 1n4004).

But if you took the threshold WAY down (say, to 100mV), the ratio of sharpness to clipping level might be as good or better than Ge diodes.  (eg if Ge diodes clip, say over a 30mV range at 300mV, and Si diodes clip over a 20mV range, then a clipping level of 200mV would give an  equivalent effect).  Note that these are hypothetical figures only, but the logic applies and the relativities are appropriate.

good luck with this.

[Ge_Whiz]

Biasing diodes nearer cutoff is a trick that has been employed for decades by crystal radio enthusiasts to increase the sensitivity of their radios - using everything from LEDs to galena + catswhisker. Do a web search for "Ben Tongue" (yes, the inventor of the Blonder-Tongue inductorless studio EQ) and visit his website.

I know - I am one of those nuts.

[zachary vex]

I thought this up a while back but never did anything with it.

Code Select Expand

                   Vr1      Vr2
                    |        |
            +---||--+--->|---+
         In |                |      Out
         ---+      Vr3       +---||---
            |       |        |
            +---||--+---|<---+

             Vr1         Vr2         Vr3
              |           |           |
V+ ---\/\/\/\--+--\/\/\/\--+--\/\/\/\--+--\/\/\/\--- GND


Imageing a filter cap and series resistor between each Vr supply and the Vr points around the diodes.

The next step would be to replace the middle two Vr supply resistors with pots to adjust the bias.

Take care,
-Peter

seems like the signal might shrink a bit if you use that approach.

[Mark Hammer]

Not to rain on anyone's parade, but I do believe that this is incorporated as part of Jack Orman's "Warp controls" document over at AMZ

I don't think so.

If I read the note right, he's talking about DC biases to change the effective threshold of a diode, while the warp controls stuff is a catalog of ways to use a blend control to change the minimum forward resistance of a miscellaney of diodes when turned on, no biasing involved.

Buried in the warp article (which DOES tend to focus largely on passive means to achieve assorted asymmetries, etc., so you are not entirely incorrect), Jack includes several sample circuits for using DC to bias diodes in feedback loops and other locations.  These may well be qualitatively different from what started this thread.  Even if they are, I suspect that less technical readers may find a nice bridge from that article to this thread via the steps from passive to DC-biased arrangements.

[R.G.]

Ah - I didn't read it in depth, just looked at the pictures 8-).

[Gus]

It looks like R.M. used resistors to leak some current around the octaving diodes in  the octave petal of his.

[Hal]

can someone link me to this warp article?  I cant seem to find it.

edit: I think i found it....right here:
http://www.muzique.com/lab/warp.htm

[puretube]

http://v3.espacenet.com/origdoc?DB=EPODOC&IDX=US2703382&F=0&QPN=US2703382

[Doug_H]

I don't mean to poop in the soup but IIRC the tyco and octave up sick box both run some dc through the diodes.

Doug