I/V of less common clipping diodes?

[Fancy Lime]

Hi!

I am looking into options for higher voltage clipping diodes, around the 2-3V mark. I have a bunch of extremely bright blue 3mm LEDs, which have a nice clipping sound. I don't have a scope, though, and do not know what type of blue LEDs they are. Seems like blue ones come in very different nominal voltages and I/V profiles, depending on the semiconductor material. Has anyone taken a scope to a selection of different blue ones? AFAIK, GaN and SiC are common substrates, both of which have an intrinsic band gap around 3V, so they have to tune that down with doping to get blue. But I have no idea, what exactly that does to the I/V profile, especially at low currents. Mine light up visibly when driven from 9V via a 100k resistor and are retina-burningly bright with 10k. So current demand is rather small.

Power MOSFETs also have nominally high drop body diodes, according to the datasheets, which is why I though they might be interesting. But these are specified at high currents (e.g. 2.5V @ 28A for the IRF540). At "normal" clipping diode currents, these have normal Si diode drops.

I know that there are GaN and SiC Schottky diodes and power MOSFETs on the market for a few years now. Has anyone looked into these? Is any of this stuff readily available and affordable for our purposes? I mean, we don't *really* need a diode rated for kilovolts and kiloamperes but if there are small enough devices, it may be worth experimenting. These substrates have other virtues, too, that influence their sound as clipping diodes. And if I remember correctly, at least GaN MOSFETs have a somehow different junction to the substrate and therefore a different "type" of body diode. Does anyone know more about this? It's been over a decade that I last did serious research into SiC devices and much has happened since.

Cheers,
Andy

[amptramp]

I have (somewhere) a SiC blue LED made by Sanyo in 1981 on a 3/4" wafer and packaged like a modern diode in clear epoxy.  It was one of the first shipment of 20 to come into Canada but the price was prohibitive at the time.  SiC has a nice blue-jeans blue at 480 nanometres and I am glad to see it hasn't disappeared into history.

You could always look at selenium rectifiers - they have a reverse voltage of 45 volts per plate so the ones used for battery chargers would be single-plate devices.  They tend to gain series resistance as they age but you might find them interesting, especially single-plate versions that could easily meet the 2 - 3 volt requirement.

[merlinb]

[aron]

I think I have those ultra bright blue LED's. I need to try them.

[R.G.]

Then there's the whole world of Vbe multipliers. Transistor and two resistors gets you the amplification of the Vbe in the transistor.

Notice that all the diode I-V characteristics have an exponential region where conduction builds up, blending into a resistive region where voltage is a straight-line resistance added to the exponential.

Between these, there is a lot of turf to cover.

Oh, yeah. One can make a Vbe+diode multiplier where the transistor amplifies the Vbe AND the diode.

[iainpunk]

i think you can gain a lot of variability with an op amp with a 1K pot on its output, before the feedback loop and feed forward node. (see pic related[the blue pot is 1k, the other one 100k])
this limits the op amps output current and thus the harshness of the clipping.

also, fust for good measure and to "clean up the distortion" i like to combine led's in the feedback loop (soft clipping style) with other diodes shunting to ground, this really cleans up the noisey fizz in a simple clipper circuit, because it doesn't saturate the opamp



cheers,
iain

[antonis]

why not Ge or GaAs tunnel diodes..??
(their "negative" differential resistance should create interesting clipping..)

[aron]

Iain, do you have any sound samples of that circuit?

[Fancy Lime]

Well that is one of the set of curves floating around. Another would be this:

Similar for IR, red, green, yellow, and UV but very different for blue, about 1V difference at 20mA. Hence my question if anyone has compared different blue ones directly.

Then there's the whole world of Vbe multipliers. Transistor and two resistors gets you the amplification of the Vbe in the transistor.

Notice that all the diode I-V characteristics have an exponential region where conduction builds up, blending into a resistive region where voltage is a straight-line resistance added to the exponential.

Between these, there is a lot of turf to cover.

Oh, yeah. One can make a Vbe+diode multiplier where the transistor amplifies the Vbe AND the diode.

Excellent suggestion! I never considered an amplified diode as a clipper. With a pot instead of the resistors, one could make a nice warp control. I have not (yet) seen that on Jack's page but it must be there somewhere... Pretty sure he has every conceivable kind of warp control on there. Have not seen it on Geofex, either. Have you ever written about it?

i think you can gain a lot of variability with an op amp with a 1K pot on its output, before the feedback loop and feed forward node. (see pic related[the blue pot is 1k, the other one 100k])
this limits the op amps output current and thus the harshness of the clipping.

also, fust for good measure and to "clean up the distortion" i like to combine led's in the feedback loop (soft clipping style) with other diodes shunting to ground, this really cleans up the noisey fizz in a simple clipper circuit, because it doesn't saturate the opamp



cheers,
iain

Hey, stop reading my mind! That is pretty similar to what I working on. I'm going for a simpler implementation than in your image but the same basic idea. What that pot does (among other things), is limit the current through the diodes to ground and thereby change the clipping characteristic quite drastically. I would use a larger pot though, like 10k. That is on of the major reasons why the Rat with a 1k resistor before the 1N4148s sounds like a distortion, almost fuzz, whereas the DOD250 with a 10k resistor before the same diodes sounds much more like an overdrive. There are other important differences as well of course but this one is critical for the character at lower gains in my opinion. Of course Jack Orman has written about this:
http://www.muzique.com/lab/sat2.htm
But the topic seems to get little attention. I never even thought about it until recently.

The diodes in the feedback loop before hard clippers to ground is precisely what I want diodes with a 2-3V clipping threshold for. For exactly the reasons you describe. Do you have a crystal ball? I usually use blue LEDs for that and I like the sound. What I'm planning now will have 2v7 Zeners as to-ground clippers because they have an exceptionally flat I/V curve in the relevant current region for smooth overdrive. With a 10k resistor they should start clipping at just under 2V according to the Nexperia BXZ84 datasheet, so about 2.2V with 1K5817s to block forward conduction. I might end up using the same 2V7 Zener diode arrangement in the feedback loop but its nice to have some other options as well for experimenting. I will probably include a switch to go to very hard clipping as well.

why not Ge or GaAs tunnel diodes..??
(their "negative" differential resistance should create interesting clipping..)

Hmm, quite possibly interesting. Or a Gunn diode. Do you have any device numbers of such diodes which may be readily available in the usual shops?



Thanks for all the input, guys,
Andy

[antonis]

>Do you have any device numbers of such diodes which may be readily available in the usual shops?<

1N3717/19 (THD) or E28X (SMD) can be easily found..
(although, USSR military Ge/GaAs tunnel diodes exhibit more "interesting" I/V curves..)

[amptramp]

If you want a diode that is very abrupt in turn-on and turn-off, check out the precision clamp circuit in the National Semiconductor Linear Applications manual on the lower left of page 17 of this:


http://natasha.eng.usf.edu/gilbert/courses/instrumentsystems/roadmaps/0_1page_opamps/13_0b_op_amp_config%20notes.pdf


You can add a resistor in series so you have a diode to ground with a variable clipping voltage and a variable resistance.  You can configure it for any voltage and any resistance but it is always configured as a clipper that acts like a diode to ground.

[R.G.]

Excellent suggestion! I never considered an amplified diode as a clipper. With a pot instead of the resistors, one could make a nice warp control. [...] Have not seen it on Geofex, either. Have you ever written about it?

I've mentioned the technique here and in other places, but no article on it. Here are some things about it:
Transistors with base tied to collector have long been used as more-ideal diodes. This connection produces an effective diode that is closer to what the semiconductor physics say a diode should act like than a single junction. You can diode-connect MOSFETs too, and produce a MOSFET "diode" with a much softer forward knee than most LEDs this way by tying the gate to the drain.

The amplified diode uses the fact that a diode-connected transistor can be fooled into turning on at higher voltages by a resistive divider. The base-emitter junction starts conducting when its voltage gets up to 0.4V and really gets into it by 0.6V. That lets current seep into the base and starts conduction of the collector. The resistor string from collector to base to emitter delays this by preventing the collector voltage from getting to the base til there's more voltage on the collector, and making the cut-in voltage be when the base-emitter resistor has that 0.4 to 0.6V across it. A high gain NPN where the base current is quite small compared to the collector current will then show an amplified version of Vbe at the collector.

I haven't done much with this because it's primarily useful if you want to clip up at several volts. Most pedal use is for clipping down in the sub-2V range. I have designed power amp pre-clippers with diodes and/or MOSFETs to give a very soft fold over into clipping before the power amp can clip. This turns the power amp into a soft clipping type where it never gets the sudden sharp clipping that seems to be what is annoying about most solid state amps.

This is not a new technique. The Thomas Organ Vox amps in the early 1960s used it.

[iainpunk]

jusk for chitz and giggles, how about reverse avalanche mode transistors to clip and create weird oscillations in one of the half cycles of the input signal...and just use a simple clipper for the other side of the input waveform both to limit the amplitude of the oscillations and to normally clip the input wave...

dear Andy,

i have known that we think really really alike since you posted the mini sawmill pedal. i had just finished the design for my "pink" pedal, that also relies on crossover distortion to get its sound, also designed for 'raw' metal sound because i dislike polished modern metal tones. i think we just have to live with the fact we really have similar minds.

cheers,
Iain

[Fancy Lime]

@R.G.
I was considering the amplified diode not as a clipping diode but as an adjustable reference for diode-to-ground clippers. So that some forward or backward DC bias can be applied to change clipping thresholds. However, everything I came up with was fairly complicated and the same effect could be achieved much easier with clipping in the feedback loop of an inverting opamp. Thanks for the info anyway, I have a feeling I'll come back to this in relation to envelopes...

@Iain
Maybe we should start a collaborative line of Heavy Distortion pedals. We can call it Jagged Edge Effects or Chipped Tooth Audio or something like that (did not check if those were taken). I suggest limiting ourselves to one sound per pedal and one control per pedal. 1590A if we can. All pedals shall be black. Or neon pink with ponies. Cannibalistic Zombie Ponies from Hell! Wait a minute, this started as a joke but now it sounds like fun... I have done some research recently into how to get the hardest possible clipping from a simple circuit... I suck ad designing boxes, though.

Andy