ADHD use of 2n7000 doesn't make sense to me...

[Paul.Ruby]

Not sure if this is already known or not but internal to the 2n7000 mosfet is a zener diode from source to drain to protect against over-voltage spikes or ESD events. I see the ADHD (and OCD) are trying to use the mosfets in a diode connection as if the clipping voltage is set by the mosfet threshold. But two are wired in anti-parallel, so the mosfet part of the component is doing nothing. The forward-biased, anti-parallel zener diodes are doing all the clipping. This makes me wonder if the 2n7000 was really the chosen original mosfet for the OCD?? It's choice seems silly, even if they happen to sound good. It's just a couple anti-parallel diodes. Am I rehashing old news as a newbie to this design?

 

[benny_profane]

Not sure if this is already known or not but internal to the 2n7000 mosfet is a zener diode from source to drain to protect against over-voltage spikes or ESD events. I see the ADHD (and OCD) are trying to use the mosfets in a diode connection as if the clipping voltage is set by the mosfet threshold. But two are wired in anti-parallel, so the mosfet part of the component is doing nothing. The forward-biased, anti-parallel zener diodes are doing all the clipping. This makes me wonder if the 2n7000 was really the chosen original mosfet for the OCD?? It's choice seems silly, even if they happen to sound good. It's just a couple anti-parallel diodes. Am I rehashing old news as a newbie to this design?

It’s all body diode clipping and needs to be redesigned to actually use the MOSFET. You’re correct.

See here for helpful info:

AMZ - Guitar Effects & diode clipping

How to use mosfet transistors and zener diodes for clipping in guitar fx pedals.

www.muzique.com

 

[Chuck D. Bones]

I had the same thought. For whatever reason, pedal designers like using MOSFET body diodes for clipping. If the gate threshold is < the diode Vf, then the MOSFET never turns on. Some circuits, Like the VFE Alpha Dog (Dingo) use the MOSFET both ways for asymmetric clipping. One could put diodes in series with the MOSFET so that they only conduct in the forward direction, but I haven't found any pedals that do that.

 

[benny_profane]

I had the same thought. For whatever reason, pedal designers like using MOSFET body diodes for clipping. If the gate threshold is < the diode Vf, then the MOSFET never turns on. Some circuits, Like the VFE Alpha Dog (Dingo) use the MOSFET both ways for asymmetric clipping. One could put diodes in series with the MOSFET so that they only conduct in the forward direction, but I haven't found any pedals that do that.

Honestly, I mostly see the series diodes in DIY designs. I’m not sure if I know of any pedals that really implement true MOSFET clipping. VFE pedals sorta bridge the gap; that’s probably the best example of DIY/production that uses it well.

EDIT: Zendrive and others definitely include the series diode.

 

[Chuck D. Bones]

I just looked at the AMZ link. If all we want is a slow diode, then the 1N400x series diodes will suffice. BJFe does that in the Sweet Honey, Blueberry, Pink Purple, etc. We can always put a cap in parallel with the diodes, like the DS-1 to smooth out the harmonics. MOSFETs have high transconductance, which means that if they are used as clippers in the forward direction, the clipping will be very hard. I figure Marc Ahlfs knows what he's doing with pedals like the Top Fuel, so MOSFET body diodes must sound good.

 

[phi1]

Zendrive has the series diodes oriented such that the mosfet body diodes are still clipping, right?

I remember testing a while back and thinking that the mosfet body diode (zendrive or ocd) was a bit “smoother” sounding to my ear than 1n4001 or 1n4148, in both soft and hard clipping configurations.

 

[Paul.Ruby]

Thanks folks. This seems to suggest the choice was indeed deliberate, even if it looked silly to me at first.

 

[Paul.Ruby]

Also, since there is a 1nf cap in parallel with the mosfets, the input capacitance of the mosfets seems completely immaterial. 1nf completely dominates over any capacitance in the mosfet. Makes it hard to believe any choice of diode really matters, beyond turn-on voltage. Of course, that's my brain talking, not my ears... I'm building two specifically so I can compare variations side-by-side.

 

[phi1]

Yeah please report back, my tests (if I’m even remembering the results right) was not on an ocd, per se, so no 1n parallel cap.

If fact, maybe you can leave that undsoldered at first and include different cap values in your experiment.

 

[Paul.Ruby]

Good suggestion. Math tells me that 1nf is still only filtering up beyond 150khz, and the capacitance in the mosfet itself would be well into the Mhz range. So my brain says the 1nf will be of no effect and the sound will be the same using any old silicon diodes (plus the germanium). But... My ears have won battles with my brain in the past. Brain said there is no difference between polyester and polystyrene film caps in tube amps. Brain was wrong.

 

[Chuck D. Bones]

The filtering is hard to calculate because of the diode non-linearity. Makes more sense to think of it in the time domain. Or, like you said, let your ears decide.

 

[HamishR]

To be honest every time I have tried or built an overdrive which uses Mosfets like this I have found the sound to be harsh and tiring. They sound good at first but after playing for a while it gets fatiguing to listen to and there's a very subtle artifact (sorry, can't thing of a better word!) in the sound which is not part of the tone. Every OCD I have tried has this.

It could well be that I am just hearing things because whenever I mention this people look at me strangely. But I really don't like Mosfet clipping. To test this I built an OCD with 1n4001s instead and while it wasn't a world-beating overdrive it still sounded smoother and less harsh to me - and I have found the same thing with other ODs.

 

[Paul.Ruby]

This is exactly what I find fascinating about building pedals and amps. Engineering is often not quite enough. I always build two of everything, side-by-side to tweak. Sometimes I'm completely surprised and cannot explain why and it leads to things that seem silly.... Like a mosfet used for only it's bulk diode (which I have yet to hear first hand and didn't even know about until today).

 

[Chuck D. Bones]

Building two and A/B'ing them is definitely Engineering. I thin you mean that analysis by itself is not enough. It's a good start, but until you build one and test it, you won't know for certain how the circuit will perform. This is especially true with guitar pedals when builders design circuits that depend on undocumented properties of components, like how transistors and opamps behave as they transition in and out of saturation.

 

[Chuck D. Bones]

It could well be that I am just hearing things because whenever I mention this people look at me strangely. But I really don't like Mosfet clipping. To test this I built an OCD with 1n4001s instead and while it wasn't a world-beating overdrive it still sounded smoother and less harsh to me - and I have found the same thing with other ODs.

The EQD Palisades (Stockade) is a great platform to compare clipping methods because it contains most of the feedback-based diode clipping arrangements.

 

[Paul.Ruby]

The EQD Palisades (Stockade) is a great platform to compare clipping methods because it contains most of the feedback-based diode clipping arrangements.

Cool! I was just thinking about adding a rotary switch to an ADHD or a queen of bone. This place has everything.