From: John Larkin
Subject: Re: Strange going-on with a floating MOSFET gate
Date: Sun, 05 Jan 2003 10:24:20 -0800
Organization: Posted via Supernews, http://www.supernews.com
X-Newsreader: Forte Agent 1.91/32.564
On 5 Jan 2003 06:39:07 -0800, Winfield Hill
>> Winfield Hill wrote:
>>> But what I didn't expect was what happened immediately when
>>> removing the clip lead to the gate. Without fail the current
>>> would instantly drop by about 2x, indicating a roughly -75mV
>>> change in gate voltage. ...
>> If you offered your hand (and croc clip) back to the
>> gate (without actually connecting to it) did you get
>> the corresponding +75mV change on the gate voltage?
> Generally the new readings were stable, movements of my hand
> nearby and other activity had little effect, that's why I went
> ahead and let the meter log 2 hours of data. The FET's tiny
> antenna wire and all that. But I haven't tried your idea.
> My latest theory, developed in the middle of the night, when
> thoughts are hazy, involves the instant just as the clip lead
> is being removed, and one's fingers are very close: what's the
> level of induced ac at that point? More than enough to drive
> the gate voltage into nonlinear regimes. Handwaving. :>)
> - Win
this is very nice, silly, useless mystery: the best kind. My goofy
1. The fet gate is unlikely to rectify AC or RF.
2. If induced AC were the cause of the 70 mv 'disconnect' step, the
magnitude and sign of the step would be random. I haven't a clue.
3. AS regards the positive gate drift...
a. If it's gate threshold drift, it's a semiconductor thing, like
a popcorn-noise sort of chemistry. I kinda doubt that.
b. If the Vg drift is real, charge is coming from somewhere,
probably somewhere external. The gate could be collecting
ions from the air, or emitting electrons photoelectrically.
It would be easy to tell at least if it's a) or b).
4. We are close here to being to resolve single electrons. Given a fet
with really low gate capacitance and leakage, you could digitize drain
current at a fairly high rate (several times the electron current
rate) and mathematically process the current samples to extract the
single-electron steps. I'm thinking highpass filter followed by some
sort of histogramming algorithm. The random electron departure times
makes this a bit messy.
Really, this has a lot of potential (pun!) as a basic classroom
learning tool. Battery-resistor-LED gives basic
potential-current-Ohm's-law stuff. Brief detour into nonlinear LED
curve and power dissipation. FET as an on-off switch. FET in its
linear region, drain curves, load lines. Charge on gate, Miller
capacitance, leakage currents, charge on an electron, air ionization,
alpha radiation, photoelectrics maybe. Fun with static electricity.
Charge pumping. A bit of semiconductor theory. I bet I could do a
full-semister course on this one circuit and its peripherials, largely
quantitative, and still be less boring than most of the EE courses I
had. No simulation, no Spice, simple math, and every student gets to
take their own rig home and play with it.