From: Terry Pinnell
Subject: Re: zero-power toggle circuit; was, how to master electronics
Date: Tue, 19 Nov 2002 15:00:34 +0000
References: <3DD39D90.email@example.com> <firstname.lastname@example.org> <email@example.com>
NNTP-Posting-Date: Tue, 19 Nov 2002 15:00:50 +0000 (UTC)
X-Newsreader: Forte Agent 1.92/32.572
Tony Williams wrote:
> Terry, is it a flakiness in Off-going,
> and is it worse for P-channel MOSFETS
> with a higher drain-gate capacitance?
> From another post just done, I'm thinking
> that if R2 is slow to come down, even for
> a mS or so, then that could be enough for
> R2 to charge C1 back to the Vgs(th) of Q1.
That seems to square with my results, Tony, such as the results I
assume you've seen at
As you see, of the 7 MOSFET examples I happened to try, only 2 of them
(IRF9140 and MTD20P03HDL) simulated as I would expect, i.e
- remained off when power supply first applied
- switched on at first press of button
- switched off at second release
The other 5 fell into 3 waveform types, ranging up to the 'worst' case
of the IRF9513, which as you see comes on with supply and is not
switched at all by any button signals. These tests were after I'd
first queried the circuit and Win and others had confirmed it was OK.
I've since checked again with similar conclusions, and you probably
saw my more detailed waveforms for the IRF9513 at
It also looks to me as if the failures are, as you suggest, associated
with higher powered MOSFETS, which have higher drain-gate capacitance.
I don't have full data for those 7 I simulated. The IRF9513 in
particular doesn't apppear in the AoE table,2nd edn., page 165. But
here's a selection that does:
Type Cgd Max current Cgd from pwrmos.lib?
---- ---- ----------- --------------------
VP0650N3 10 pF 100 mA
IRF9120 50 pF 1 A
IRF9520 50 pF 6 A 130 pF
IRF9640 150 pF 11 A
IRF9540 250 pF 19 A 470 pF
I think the IRF9513 is up-range with the IRF9540, whereas the BS110
(which Wafer said worked reliably), is a 170 mA type, and therefore
perhaps closer to the VP0650N3.
That last column came from my attempt to google my way to a datasheet
for the IRF9513. I did however find a library of MOSFET Spice models
which included this:
.model IRF9513 PMOS(Level=3 Gamma=0 Delta=0 Eta=0 Theta=0 Kappa=0.2
+ Tox=100n Uo=300 Phi=.6 Rs=.3715 Kp=10.54u W=.2 L=2u Vto=-3.923
+ Rd=.6523 Rds=266.7K Cbd=331.8p Pb=.8 Mj=.5 Fc=.5 Cgso=2.306n
+ Cgdo=551.8p Rg=.8973 Is=3.748E-18 N=3 Tt=2250n)
* Int'l Rectifier pid=IRFC9110 case=TO220
* 88-08-25 bam creation
If the parameter 'Cgdo' is meant to be same as Cgd (which the AoE
table also calls 'Crss' BTW), then that implies Cgd = 550 pF. But
comparing similar pwrmos.lib models for the IRF9520 and IRF9540, with
the AoE data makes this look a bit iffy; the pwrmos.lib figures are
around twice as high.
Anyway (despite fact that I see Win has sustained his earlier
position!), I'm crystallising my original view that the circuit is
very sensitive to MOSFET type.
This is in marked contrast to Win's own original zero toggle switch
based on a PNP BJT and an N-channel MOSFET. I posted my simulations of
that earlier. But I've now also breadboarded Win's corrected version
successfully with a variety of MOSFETS, bearing out my simulations.
This is why I'm increasingly confident that neither CM models nor my
simulation technique is at fault regarding the later 'Wafer/WH'
circuit. But it would be good to see another set of simulations from
someone more experienced, with a different package. I can't breadboard
it as I have no P-channel MOSFETs of any type!
Hobbyist, West Sussex, UK