FETs as diodes

[johngreene]

There are neato datasheets that you can find just by googling with the part number and "datasheet".

For the BS170, the gate threshold voltage is listed as 0.8 to 3.0V, with a typical value (whatever that is...) of 2.1V. The gain is the forward transconductance, which is listed as 320mSiemen, or 0.32 amps per volt.

For the 2N7000 the gate threshold is the same, but the forward transconductance (gain) is listed as 80 minimum, 320 mS typical.

For the IRF510, the gate threshold is 1.0 to 4.0, the gain is 1.2S minimum, or about four times the typical BS170 or 2N7000.

Then you have something like the Si1012R/X from Vishay that has a MAX gate threshold of .9V and a MIN of .45. Transconductance is a little high at 1S but in pedal circuits you are playing mostly in the knee area anyway because of the low amount of current. Something to remember too is that the gate threshold is most commonly measured at a DS current of .25 mA. You will start to see conduction at lower voltage with lower currents.

From this you can get a taste of just how much there is to chose from. What sounds the best? It depends.

--john

[stm]

My 2c.

John, your graphics are very informative.  This thread has been particularly interesting to me, since I had the gut feeling that when people talked about MOSFET diodes they were not talking about the same thing necessarily. Now that it has been properly clarified, I think it is in order to be specific when refering to the subject:

So, a MOSFET-AS-DIODE produces a softer knee with a higher threshold, as opposed to the MOSFET-BODY-DIODE, which is essentially another silicon diode, or not ?...

...I've always had the feeling that the IRF520 MOSFETS sounded better/different in the Shaka 3 because said MOSFETs have a very large Drain to Source intrinsic CAPACITOR in parallel with the integral body diode, acting as some sort of filtering.  In fact, this capacitance is in the order of 1000pF IIRC, thus, very near to what you may find in parallel with the clipping diodes  of a Ross distortion, liquid drive and others.  I think this explains why it sounds different/better than a standard pair of silicon diodes and was preferred by Aron.  Moreover, the addition of the Ge diodes in this case do add some additional softness to the sound.  Remember that some poeple has reported good results in the TS workalikes when using mixed silicon and germanium diodes in series.

-----------------------------
Now regarding different unexplored diode types:

a. Has anybody experimented with "lambda" diodes?  I've seen one schematic (Pearl Overdrive?) that used two back-to-back diodes in the feedback loop of an OpAmp. One os them said "lambda diode" and had a strange part number (like MA150 or something).

b. What about tunnel diodes (very nonlinear!)  I haven't been able to get some, but there would be something new in there.

[johngreene]

Thanks for the comments, I always understand things better where there's pictures involved.

I did a lot of work with MOSFETs as clippers back when I first designed this pedal:

http://www.greene-pedals.com/glasspak.htm

I think I built the first one back in 1999 or 1998. I must have tried a hundred different MOSFET parts. Each one sounded different. The one I liked the most was discontinued but I managed to find something like 5000 of them from a broker and bought them all.

The body diodes are not anything to ignore either. I've liked several versions I had put together only to realize later than I put the other diode in backwards and was actually using the body diode. They tend to have more spank than a 1n4148 or similar diode. I'm not too familiar with the transfer characteristic of the body diode but whenever you put 2 or more semiconductor/diode devices in series, the transfer characteristics will follow that of the lowest transconductance device in the chain.  So if you put a germanium diode is series with the MOSFET body diode and the germanium has the dominant curve, it probably doesn't sound a whole lot different than if you used a regular silicon diode with the same forward voltage drop.

--john

[WGTP]

Thanks R.G., but what would be really neato is if I understood the stuff on the data sheets. 

Thanks John and others for helping clarify this for the rest of us.  It ought to get some more play around here now.

About the time I think I've covered the cool stuff, something else comes along.   

[stm]

I've been studying the Fender designs and if you look at the pic I posted in a previous message they come up with something interesting but  they go overboard.  The have a pair of clipping diodes right before this more complex clipping stage (CCS).  What do they expect is going to happen?  They already chopped off a lot of "information" from the signal before it reaches the CCS...

...

Alf

Alfonso, if you look carefully at the schematic you posted, you will notice the first pair of diodes is located at the very (-) input of U2B. Under normal operating conditions, the (-) input is a virtual ground (just a few millivolts apart from the voltage set at the (+) input).  Only when U2B is heavily overdriven so that its output starts clipping, then the first pair of diodes will start conducting.  As such for low to medium gain settings it is indeed a soft clipping design.

Best regards.

[ahermida]

I should breadboard it and listen to it.  Thanks!

Alf

[amz-fx]

For the IRF510, the gate threshold is 1.0 to 4.0, the gain is 1.2S minimum, or about four times the typical BS170 or 2N7000.

While that (gain) is true of the IRF510 under the conditions of the datasheet, which are pulses of high current, I would expect the '510 to behave differently under small signal audio conditions with much less current. It is not made to be an audio amplifier and behaves differently from the BS170.  Some people prefer its softer (?) sound...  I think Aron did some tests with it.

regards, Jack

[NoFi]

Thank you for all the useful info in that thread.

I was just wondering, is the Ibanez MT10 mostortion (mosfet overdrive) made during the eighties using that sort of diode connected mosfets ?
The schematic is here but i don't really have the skills to understand it :
http://www.freeinfosociety.com/electronics/schematics/audio/pictures/ibanezmt10.gif

[ahermida]

Thank you for all the useful info in that thread.

I was just wondering, is the Ibanez MT10 mostortion (mosfet overdrive) made during the eighties using that sort of diode connected mosfets ?
The schematic is here but i don't really have the skills to understand it :
http://www.freeinfosociety.com/electronics/schematics/audio/pictures/ibanezmt10.gif

Its really a diode clipper but the dual op-amp chip itself has the mosfets.  I think those chips have been discontinued for a while now.

Alf

[luix]

Hi guys, sorry for the intromission but I like this argument to much...
I made some experiment whith JFET in the feedback loop of opamp (circuit like the tube screamer) and this is the result: CLICK

It seem like the distortion of a tube, the only problem is the gain of the circuit, is a bit low so for guitar use (i tested it with a CD and a sinusoidal track) so I made the 1.2k resistor in series with the drain of the jfet a 500k log pot...
The sound is not bad for my hear, if you try it let me know the result...

[WGTP]

The fender schematic looks like some of the dual op amp stuff around, but what is going on in the loop of the second op amp?  I guess that is where we would want to put the Mosfet clipper?  Sort of goes back to the thread about the "spacing" between clipping stages.   

[WGTP]

OK, I'm set up on the breadboard with the 510,s as clippers but the op amp contributes more distortion than the clippers.  Guess I need to try a rail to rail op amp, or get mosfets that clip at a lower threshold.  Any other good ones to try?   

[lovric]

Hi Ahermida,

On L2XH site was said that emulation of a tube amp goes in two(3) steps:

- the preamp stage soft clipping is emulated by two antiparallel Si diodes to ground and
- the power stage clipping is emulated with simmetrical opamp clipping
- speaker/cab emulator is a must

I've listened to fender amp on a gig and it does sound sort of 'retro'. But it is great when you know the subject, only the liking of the sound is a different problem altogether. It's great when I fantasize about old marshalls and voxes and tweed amps than all of a sudden I would like sustain and heavy tone of a rectifier or something simmilar. Companies offer us three channels and what not, but we probably can't have it in a singlebox. Is there a device with such a sensitivity span that coud translate my pick attack from country all the way to the santana sustain. gary moore sustain. Whatever. EH says it's a reverb...

[lovric]

oh, I meant Fenderr solid state amp advertised with dyna-touch'ability'. From a far it looked like blackface deluxe. A far => three tables away.

[WGTP]

In the continuing saga of clippers, I experimented with the IRF520 and BS170 in the feedback loop of an inverting op amp lightly driven by a non-inverting op amp this weekend and found:

1.  They both sound nice, seemingly more organic or earthy than LED's.

2.  The GE 1N34/BS170 combo has a lower clipping theshold than the GE 1N34/IRF520.  When adding and removing them from the FBL the BS170 is quieter and more distorted.  The IRF520 isn't much quieter or distorted than the op-amp alone.  I was using a rail-to-rail dual op-amp to maximize the distortion of the mosfets and minimize the distortion from the op amp. 

3.  The CMOS op-amp doesn't sound bad by itself.

4.  They both seem to have less treble content than LEDs, which I'm assuming has to do with the softer knee generating fewer high order harmonics.

5.  The body diodes measured under .8v for the IRF520 and over .8v for the BS170.

6.  As stm pointed out, the IRF520 has almost .001uf of capacitance. The BS170 has significantly less (almost X10 if I understand the info).

7. Using a switch to rout the clipping thru the body diodes or the mosfets would result in a nice sounding clipping distortion with 2 different thresholds.  Around .8v for the body diodes and 2+v for the mosfets.   

[RedHouse]

They are configured such that it will use the body diode of the MOSFET and not the MOSFET as a diode. The 1N34s would have to be connected the opposite way for that.

--john

Ok one more time, by "opposite" do you mean the 1N34 is flipped (in the Shaka picture above in Reply #18) so it's anode connects to the MOSFET ...or... do you mean the 1N34 is moved over to the G-D connection on the MOSFET?

[WGTP]

Look at my drawing.  You can hook it to either connection, as long as it is "facing" the right directionl.  IIRC in Arons drawing, the 1N34 is connected the same, just turned the opposite direction.

What I have found on the bread board is that the body diodes are usually lower threshold, so that connection results in a lower volume more distorted sound.  When the mosfet connection is used, the output is higher and the sound less distorted.  This is the one we are after.   When either connection is out of the circuit, it is the loudest and least distorted, just the op amp.

Remember, in a Dist+/Rat diode to ground set up, the op amp will start distorting at some point and you will have your smooth Mosfet distortion after the harder op amp distortion.  Could be an effect to be "balanced" for nice transition from hard to smooth distortion over the decay period.

If you just want the Mosfet distortion, you need to use them in the feedback loop like a TS or Muff Fuzz to minimize the op amp contribution.   

I have a nice Brown sound using dual feedback loop mosfet clipping and a BMP tone stack.  That is the way I have been dealing with it.   

[WGTP]

Edit: sorry wrong post location, still interesing thread.

How about using just one mosfet without the series diode as the clipping device. (and I mean just one mosfet replacing the whole "standard" two diode back to back thing) Would it give asymmetrical clipping where the tops of the waveform would be clipped very softly and the bottoms would be clipped much harder?

If it works like this the harder clipping would happen a earlier than the softer, but it could probably be biased towards the softer side.

I might have understood it all wrong though...

Edit: Based on the other MOSFET-clipper thread it seems that it should work as I thought.


yes, it would work just as you described.

Try it!

--john

[jrc4558]

Thanks for the comments, I always understand things better where there's pictures involved.

I did a lot of work with MOSFETs as clippers back when I first designed this pedal:

http://www.greene-pedals.com/glasspak.htm

Wonderful thread gentlemen! Very informative!
Mr Greene, are those mosfets as clippers? http://www.greene-pedals.com/Overdrive%20pedal_files/inside.jpg

[johngreene]

2 of them are.