Characteristics of "side effect diodes" of JFETs and MOSFETs?

Started by Fancy Lime, May 07, 2020, 02:38:27 PM

Previous topic - Next topic

Fancy Lime

Hi there,

I was wondering if someone has taken a scope to gate-channel diodes of various JFETs or body diodes of various MOSFETs. Since these are not really "designed for" properties of FETs and therefore not usually included in the datasheets, I wonder if there are interesting variations going on in the exponential region (knee). For example, people seem to feel body diodes of BS170s or 2N7000s sound different than regular old 1N4148s while others claim they "should" be identical. I always assumed Si diodes are Si diodes are Si diodes but I definitely hear differences between 1N4148, 1N4001, and forward biased Zeners in some clipping configurations but not in others. So what does a scope have to say about the "unintentional diodes"?

Cheers,
Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

iainpunk

yeah, the BF245 Jfet has almost the same diode curve as an 1N4148.

the body diode in a mosfet is really sharp cornered in most cases.

that's all i know

cheers,
Iain
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

Fancy Lime

I have read the claim that body diodes are sharp many times but I have never seen it backed up by graphs or an original source for that claim. Listening to BS170 body diodes as clippers, they *appear* to go into clipping more softly than 1K4148s. At least to me but I have learned not to trust my own ears.

Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

iainpunk

idk, ill try if i can find my diy transistor and diode tracer in my junkpile anywhere and i'll hook it up to my o'scope and ask my brother to take a picture for you (i don't own a camera of phone) i have no idea which FET's are in my FET pile tho. they are all shorted together to protect them from killing themselves by keeping their leads at the same potential, broken mosfets still have their body diodes tho.
1n4148's have other bad behavior that makes them sound "bad", a little wiggle in the knee of the curve if i remember correctly, same as BF245's i was told that tunneling started just before conduction in those semiconductors, giving you a little wiggle in the curve.

cheers,
Iain
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

Fancy Lime

Hmm, that is interesting! I did not know that the 1K4148 had a little tunnel region. That might explain some of the sound differences to other diodes, though. I always "felt" that the 1N4148 sounded a little "fizzier" or "more brittle" than rectifier diodes or forward biased Zeners on slowly decaying notes, without being able to put a finger on what causes that feeling. May all be imagination, but if there is a tunnel region, that would produce some rather unpleasant harmonics at the edge of clipping.

Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

iainpunk

for the snare drum im building, im using 4148's because of the fizz, haha.
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers

anotherjim

I just found out that MOSFET diodes are a thing you can buy. They are called SBR super barrier diodes. The body diode may well be ordinary Si, but with the gate tied to source, it's quite different.
Here's a datasheet. I don't really know that much on the subject, but the Vf curve looks boringly linear, especially at 25c. Mind you, I never entirely trust graphical information. When I was young, I was told to Imagineer a smooth line between points when drawing a graph. I have no idea how much smoothing is applied to datasheet curves.
https://www.diodes.com/assets/Datasheets/SBR0330CW.pdf
More diodes...
https://www.electricaltechnology.org/2018/12/types-of-diodes-their-applications.html

I thought 1N4148 are gold doped? In which case, different from plain Si.
Will the real plain old average silicon diode please stand up?


teemuk

I think there is no such thing as "plain" silicon diode.

There will be variance in V/I characteristics from diode type to another, but probably not enough to be any considerable effect in applications of plain voltage rectifying. They don't make the datasheets for us who use those diodes for audio clipping and harmonic generation where little details actually do matter.  ;)

PRR

> Will the real plain old average silicon diode please stand up?

Transistor B-C junction.
  • SUPPORTER

iainpunk

Quote
Will the real plain old average silicon diode please stand up?

the 1N400X series is quite "average" if you ignore capacitance
friendly reminder: all holes are positive and have negative weight, despite not being there.

cheers