Diode Options

[soupbone]

I was messing with some different diode options using alligator clips.I was wondering if anyone had a couple of diode diagrams.My two ideas were to have 2 different diode combos on a swith using an on-on.The other idea was the same idea,but using a on-off-on swith.That way you have a "no diodes" option.Any links or pic's on how to do this would gladly appreciated.

[amz-fx]

3 diodes options with the on-off-on switch:



regards, Jack

[soupbone]

3 diodes options with the on-off-on switch:



regards, Jack

Thanks!

[thermionix]

With the switch in an ON position, are the LEDs essentially "shunted" due to the lower forward voltage of the other diode pair selected?

Sorry if this is a dumb question, but I'm still new to this kind of stuff.

[Phoenix]

With the switch in an ON position, are the LEDs essentially "shunted" due to the lower forward voltage of the other diode pair selected?

Yes, that's exactly right. When placed in parallel, whichever diode has the lowest forward voltage will conduct before the other, until the forward voltage of the other diode is reached, which won't happen here, the LED forward voltage is much higher than the 1N4148 or 1N34, but can lead to uneven power sharing in power rectifier circuits if you try to use multiple packages in parallel to increase current handling, even if the diodes are the same type, as each junction will have a different tolerance on the forward voltage, so balancing resistors must be used if trying this method. Most of that rarely is of concern with effects pedals though, but definitely something to worry about if building amps.
Also note, if diodes are placed in series, their forward voltage adds together, which is often used in asymmetrical clipping circuits, with two diodes in series in one direction, and a single diode the other.

[thermionix]

Cool, thanks for the reply.  I suppose I'm learning.  I have never considered paralleling diodes in an amp rectifier, but I've never built anything where 1N4007s weren't more than enough.  And mostly I use tube rectifiers.  There was a 50's tweed Fender that used two (5U4s?) in parallel, can't remember which model off the top of my head.  I have asymmetrical clipping in my Tube Screamer clone and I love it.  In the future I'd like to build another with enough space for a switch to select clipping diode setups, hence my interest in Jack's diagram.  Much simpler than a 3-way rotary.

[antonis]

In the future I'd like to build another with enough space for a switch to select clipping diode setups

Actually, there aren't many setups (with audible difference..) beyond the "standards"..
(Ge, Si, LED, MosFet)


But your ear definately isn't mine..

[GibsonGM]

In the future I'd like to build another with enough space for a switch to select clipping diode setups

Actually, there aren't many setups (with audible difference..) beyond the "standards"..
(Ge, Si, LED, MosFet)


But your ear definately isn't mine..

Yes....I think that's something worth trying out, make a high-gain opamp amplifier snippet and start plugging in different diode/combinations...then you'll have a mental file of what sounds like what...3, 4 choices.   

Some in series, or LEDs, which will make your output LOUDER and seem to have more PUSH, too!  < I use that a LOT

But after a couple beers (or hours), you'll probably be done and not want a 'permanent circuit'...when you make up a dist. in the future, you'll just try a few combos again, see what sounds best

[ashcat_lt]

I have asymmetrical clipping in my Tube Screamer clone and I love it.  In the future I'd like to build another with enough space for a switch to select clipping diode setups, hence my interest in Jack's diagram.  Much simpler than a 3-way rotary.

Note that this diagram is for diode-to-ground "hard" clipping like in a Dist+ or Rat.  It can work about the same in the feedback "soft" clipping of a TS, but would need to be adapted.

Now for a bit of an unpopular take on it.  Unless you're really looking for that change in volume as you switch diodes, you can accomplish much the same thing by just changing the gain of the circuit feeding the clipper.  Taking the broad approximation that the forward drop of a Ge is half that of a Si which is half an LED, you could just build it with Si, drop the gain to half for an LED sound and bump it up to double for Ge.  People will say these devices clip differently, but the VI curves are pretty close to dead on exact, just scaled up or down.  The differences you hear are dominated by the relative level of the signal compared to the Vf of the diode, and probably some by the change in output level.  Adjust the gain, normalize the output, you'll never know the difference.

Course that assumes that nothing else changes when the gain changes.  In most gain stages things will change.  The filters move, you might clip the gain stage differently, and theres that whole GBW/slew rate thing.  You might be better off attenuating after the gain stage, but resistive dividers will affect the current through the diodes which will change the way they clip, so you might need to buffer the attenuator output...and it just won't work in a TS...

Since it was touched on above, I'll also put forward that it is actually a bit more complicated to get asymmetrical clipping out of a diode-to-ground arrangement like this than most people realize.  Since the diodes go to actual ground, we have to assume that it's AC coupled to the circuits around it.  That means it's got a big old cap in series at both input and output.  Those caps won't allow this circuit to maintain a DC offset, but that's kind of exactly what asym clipping is.  The caps charge or discharge and drag the signal with it until it's centered between the diodes and the output is symmetrical.  If you watch it on a scope or sim it out, you can actually watch it go from the asym form you expect to perfectly symmetrical over the course of a short time as it goes from just turned on and the caps charge up.

In order to get reliable asymm clipping here, we should probably remove the cap coming from the gain stage and reference the diodes to Vref instead.  Then we can use asymmetrical diode pairs and it will actually work.  Or we could "mix in" a little bit of DC voltage after the gain stage on the way to the diodes and adjust the gain appropriately and get about the same thing again.  Again this doesn't work on TS, but it's not necessary because it works fine with asymm diodes anyway.

[TejfolvonDanone]

In the future I'd like to build another with enough space for a switch to select clipping diode setups

Actually, there aren't many setups (with audible difference..) beyond the "standards"..
(Ge, Si, LED, MosFet)

But your ear definately isn't mine..

For asymmetrical clipping you can use a low voltage Zener in the feedback of an opamp. The BZX55 series has 1.4V and even a 0.8V type (around 0.09c each here). It can give you a really harsh sound because at the Zener voltage the breakdown is much steeper than the forward curve so this might not suit your needs. I think it's worth a try.
If you try to abuse the Zener effect you have to use only one diode. If you use 2 in paralel one will always conduct before you reach the Zener voltage of the other.
You can use two in series and it gives you the hard clipping of the Zener but for both sides of the curve.

Also you can try Schottky diodes and they clip relatively soon around 200-300mV. I haven't heard if they have audible difference in their clipping (despite the lower forward voltage) but i haven't heard about many Schottky mods.

There is a last option called tunnel diode but i haven't heard of anyone even thinking about putting it in an audio circuit. I'm not really sure if it's even worth a try.

I'm thinking about creating an ultimate overdrive like this: http://diy-fever.com/effects/the-ultimate-overdrive/ but with other diode options: Si, Ge, MOSFET, Schottky, LED, tunnel (for experimenting), and a 0.8V Zener for one side and another 1.4V of the other.

I think i covered all of the diode types.

[TejfolvonDanone]

> Now for a bit of an unpopular take on it.  Unless you're really looking for that change in volume as you switch diodes, you can accomplish much the same thing by just changing the gain of the circuit feeding the clipper.  Taking the broad approximation that the forward drop of a Ge is half that of a Si which is half an LED, you could just build it with Si, drop the gain to half for an LED sound and bump it up to double for Ge.  People will say these devices clip differently, but the VI curves are pretty close to dead on exact, just scaled up or down.  The differences you hear are dominated by the relative level of the signal compared to the Vf of the diode, and probably some by the change in output level.  Adjust the gain, normalize the output, you'll never know the difference.
Grab a scope and prove it! Just kidding. But the little scientist in me begs for proof. The question is if we could create a measurement where everything stays the same except for the diodes. As Yoda would say "I'll meditate on that".

Am i going really off topic?

[thermionix]

Actually, there aren't many setups (with audible difference..) beyond the "standards"..
(Ge, Si, LED, MosFet)


But your ear definately isn't mine..

The possibilities are virtually endless, what with all the different diodes and combinations you can use.  Some sound similar for sure.  Some are radically different.  I've been truely amazed just by my limited experiments.  I was thinking of having 3 switchable options, you list 4 as "standards."  If you took one of each type, how many combinations can be made by mixing them together?  Then...how many different types of Si diodes are there?  Thousands?  They don't all sound the same.

[PRR]

> Grab a scope and prove it!

Why do original research?

As you know, it is easy to confuse yourself.

And there's really nothing new in audio.

Ponder this graph:

http://www.oldradioworld.de/gollum/fig04.jpg

Ge and Si have the same shape of a curve but Si voltage is about twice as high.

As Ash suggests, you could run the other type and adjust gain before/after to get the same final result.

LEDs are not on this graph but they follow the same curve at 2X to 3X the voltages of Silicon.

FWIW, the water flow and level in my ditch follows very similar laws for similar reasons.

The shape of the "ditch" does matter some. Vacuum Diode curves are a little different.

Nearly all devices "have" three different curve-laws, but the "useful" range tends to be mostly in one of the three basic curves. The ultimate curve for a high current density vacuum triode is the same as a BJT (except temperature), but we can't make grid wires so fine and tight to reach that regime. FETs follow a square-root law, except at insanely low currents they follow the exponential of a BJT.
------------------------

Ash's point about asymmetry versus AC coupling is well taken. AC coupling will ALWAYS remove the asymmetry average; and we are always AC-coupled somewhere, even if only at the speaker-air interface. However there will be a lag before the asymmetry is reduced, and the final wave may distort different. And what does that *sound* like? Much of this is not worth heavy analysis. BB it, beat it, and listen.

[Beo]

With Diode switching, my problem has always been my perception of distortion/tone vs volume. I thought of calibrating a diode switch that would select between LED, SI and Ge, with accompanying gain changes to keep volume level consistent. Never did this, but I suspect my ear wouldn't hear as much of a difference as I do with the volume change without gain compensation. Has anyone tested diode clipping options with gain compensation?

[teemuk]

Ponder this graph:

http://www.oldradioworld.de/gollum/fig04.jpg

Ge and Si have the same shape of a curve but Si voltage is about twice as high.

Actually, what I like about the graph most is how it displays the inconsistency of diode types. For example, 1N34A and AA112 are both germanium diodes but while characteristic curve of the 1N34A is almost similar to those of silicon diodes (except lower Vf) characteristic curves of the AA112 are distinctly different with more gradual "sloping".

As is, with identical input signal AA112 would clip "softer" than a 1N34A diode, generating less higher order harmonics and "flat topping" at higher peak voltage. But how audible is that? Well... based on wittnesses accounts most seem to regard all germanium diodes "the same" even though they evidently are not. So I suppose audible effects of differences (excluding Vf) are minimal at least.

---

Not to mention, one can add that kind of gradual sloping characteristic TO ANY DIODE simply by adding a low-ohmic series resistor to it. Much easier than "hunting for the perfect diode" with ideal characteristic curve. Vacuum tube diodes have very high series resistances so their characteristic curves generally have plenty of that "sloping" in comparison to semiconductor diodes.

[antonis]

Actually, there aren't many setups (with audible difference..) beyond the "standards"..
(Ge, Si, LED, MosFet)
But your ear definately isn't mine..

The possibilities are virtually endless, what with all the different diodes and combinations you can use. 

I definatelly agree...!!
(refering to the case of mathematical possibillities combination..)

But it doesn't ensure that your audio feeling "threashold" is as sensitive as the combination's lowest one..