Reason for the single Ge diode in the MOSFET clipping pair (eg. Fulltone OCD)

[fryingpan]

Hi everyone, I'm new to the forum, I dabble with electronics (but more academically than practically, I study EE in my free time).

I'd like to know what's the reasoning behind using a single diode in a MOSFET clipping pair. The diode is supposed to prevent the series MOSFET from functioning as a regular diode (the so-called "body diode", when V_DS<-0.7V or so). But the other MOSFET in parallel is free to do so. So, what gives? Using just one MOSFET is fine (and it produces asymmetrical clipping). If you want symmetrical MOSFET clipping I think you should put diodes on both branches (and use as low threshold diodes you can find, possibly? ie. Schottky or Ge diodes).

[antonis]

Wellcome..!!

I dabble with electronics (but more academically than practically, I study EE in my free time).

More wellcome...!!! 

Threshold of diode used in series with other clipping device is added on individual device's threshold..
(Only in case of forward drive, of course..)

I can't see any reason for placing a series diode to a pair of same clipping threshold devices other than for partial use, via a selector perhaps, for also asymmetrical clipping option..

[jellyjams]

Your schematic is probably wrong, since that diode should be in the same direction as the MOSFET body diode to produce asymmetrical clipping. On another pedal forum I can't link to here, they've traced an OCD clone (Joyo Ultimate Drive) where the diode was in the wrong direction which made the circuit sound wrong, with osciloscope readings and all.

(As a matter of fact the Ge diode was only added to the OCD circuit later on, but then scrapped in the new Enhanced Bypass version, going back to symmetrical clipping.)

[fryingpan]

Wellcome..!!

I dabble with electronics (but more academically than practically, I study EE in my free time).

More wellcome...!!! 

Threshold of diode used in series with other clipping device is added on individual device's threshold..
(Only in case of forward drive, of course..)

I can't see any reason for placing a series diode to a pair of same clipping threshold devices other than for partial use, via a selector perhaps, for also asymmetrical clipping option..

The point is that only one of the two MOSFETs is preceded by a diode, so basically that MOSFET is always off: it is excluded from the circuit by the diode when voltage is negative (while the other antiparallel MOSFET will slowly start to conduct as V_DS>0), while when voltage is positive before the Ge diode, considering that MOSFETs used as clippers (gate and drain connected together) have a high threshold voltage, the other MOSFET's "body diode" will start conducting much sooner. (Having low threshold diodes anyway help with keeping the MOSFET's forward threshold low enough to be practical and not require as high an amplification beforehand, but they only makes sense of you have diodes on both MOSFETs).

[antonis]

The point is that only one of the two MOSFETs is preceded by a diode,

You now refer to a MosFet in series with a diode in pair with a single MosFet [ (MosFet + Diode)//Mosfet ] and not to:

using a single diode in a MOSFET clipping pair.

[fryingpan]

Your schematic is probably wrong, since that diode should be in the same direction as the MOSFET body diode to produce asymmetrical clipping. On another pedal forum I can't link to here, they've traced an OCD clone (Joyo Ultimate Drive) where the diode was in the wrong direction which made the circuit sound wrong, with osciloscope readings and all.

(As a matter of fact the Ge diode was only added to the OCD circuit later on, but then scrapped in the new Enhanced Bypass version, going back to symmetrical clipping.)

I rechecked the schematics, and in fact the correct ones show what you say, and so the OCD basically uses MOSFETs as simple body diodes. The Ge diode on one of the two MOSFETs is there to allow for asymmetrical clipping (increasing threshold on one side the pair) but at practically no time the MOSFETs are exploited for their soft knee, slow activation characteristics (drain-to-source conduction, gate and drain shorted together).

[fryingpan]

The point is that only one of the two MOSFETs is preceded by a diode,

You now refer to a MosFet in series with a diode in pair with a single MosFet [ (MosFet + Diode)//Mosfet ] and not to:

using a single diode in a MOSFET clipping pair.

The Shaka MOSFET clipping arrangement calls for diodes on both branches, maybe I wasn't clear. Anyway yes, on the OCD (certain versions at least) you have


---Diode>--S G D [MOSFET]---
/                            \
\                            /
-------D G S [MOSFET]-------

[jellyjams]

drain-to-source conduction, gate and drain shorted together

Which is exactly what the body diode is. The "body diode" in a MOSFET is not a deliberately added feature, but rather it is inherent to the construction of the MOSFET itself. https://en.wikipedia.org/wiki/Power_MOSFET#Body_diode

[fryingpan]

drain-to-source conduction, gate and drain shorted together

Which is exactly what the body diode is. The "body diode" in a MOSFET is not a deliberately added feature, but rather it is inherent to the construction of the MOSFET itself. https://en.wikipedia.org/wiki/Power_MOSFET#Body_diode

Isn't it the other way round?

http://www.diystompboxes.com/smfforum/index.php?topic=90474.msg769017#msg769017

[jellyjams]

That's not quite correct. The body diode is forward biased from S to D, while the voltage in the post you quoted is from G to S, and thus from D to S since G and D are tied, making the body diode act like a zener; the zener voltage would be the gate threshold voltage. Hence, the body diode would behave like a normal diode when forward biased (S to D), but when reverse biased (D to S) by at least the zener voltage, it will conduct as well. However, in the Shaka schematic where G and D are tied, the Ge diodes prevent the body diodes from being reverse biased, so the zener voltage doesn't come into play and the MOSFETs only behave like normal diodes.

[jellyjams]

In addition, even if you leave out the diodes in series with both MOSFETs, one of them will conduct forward before the other one will zener, since the zener voltage is much higher than the forward voltage.

On the other hand, the zener property from tying D and G versus tying S and G might have an effect on the diode; in my tests the former configuration has a slightly higher forward voltage than the latter.

[fryingpan]

OK, (I'm sorry but we use different terminology in Italian, so I might not be able to come across as clearly as I'd like to be). You have basically confirmed what I was trying to say, I thought that the most peculiar behaviour of a MOSFET used as a diode was when the body diode is reverse biased (="zener" behaviour) as in fact the MOSFET works as a smoothly variable resistor up to Vth, when it practically switches on. I know that forward biased MOSFETs used as a diode have a somewhat different response to regular silicon diodes. By reversing both Ge diodes in the Shaka overdrive you'd get both MOSFETs clipping as a "zener" diode symmetrically, is that correct?

[jellyjams]

Yes I'd think so, and it's nice that we agree with each other all along I was thinking that your main question concerned the Ge diode and not the MOSFETs themselves.

[antonis]

it's nice that we agree with each other all along

No beneficial deduction ever came up with such a manner..