On 29/08/07, Bob Axtell <engineer@cotse.net> wrote:
> Crist=F3v=E3o Dalla Costa wrote:
> > Hello
> >
> > My mosfets are misteriously short circuiting. I had the circuit working=
 up
> > to yesterday using Microchip=B4s TC4427 mosfet gate driver then when it
> > overheated (due to me failing to account for increased switching losses
> > under pwm and insufficient heat sink) it took my last TC4427 chip with =
it.
> > However manual switching was OK many repeated times so I know up to the=
n the
> > circuit worked.
> >
> > Now while I wait for more TC4427s to arrive in a couple of days I wante=
d to
> > continue working so I copied the discrete driver circuit from here:
> > http://users.tkk.fi/~jwagner/tesla/tv-driver/flyback-mosfets-large.gif<=
http://users.tkk.fi/%7Ejwagner/tesla/tv-driver/flyback-mosfets-large.gif>
> >
> > However using that circuit caused the mosfet to immediately short-circu=
it
> > when power is applied (with 0 gate voltage). Using a 1u cap between gat=
e and
> > source cured this first problem and when I turn on the transistor it wo=
rks
> > ok (gate voltage is mantained) until the point I turn it off when I can=
 hear
> > a small bang and the mosfet short-circuits.
> >
> > The load is inductive, using 15 A 120VDC, I have the necessary anti-par=
allel
> > diode and also experimented with 100n and 50u polyester capacitors in
> > parallel to attempt to control the kick back voltage, to no avail. The
> > transistor is a IRPF460 20A, 500V.
> >
> >
> But I can assure you, kickback IS the problem. You will need a very FAST
> spike-killing device; this will be
> a varistor or TVS. I recommend the PK1E6 series, with the voltage rated
> at the voltage that the MOSFET
> will handle without breaking down. People here will argue with me about
> this, but to protect a MOSFET a
> diode will NOT work as well as a TVS. To placate people, I put a diode
> across the line as WELL as a TVS.
> With that problem solved, then you can examine other things. MOSFETS are
> very sensitive to pulse noise, and
> inductive pulses can be as high as 500-600V.
>
> The reason is that most diodes have a lengthy switching time. TVS and
> varistors will switch in a few picoseconds.
> The voltage across a diode will continue to rise until it decides to
> switch; TVS devices switch so fast  the  pulse
> has no chance to rise.
>
> --Bob A
> --Good luck.
>
>
> > However the fact it switched on and off repeatedly with Microchips mosf=
et
> > gate driver leads me to believe the problem is not in kick back voltage=
 but
> > rather in the gate control circuit I built. Would anyone have any idea =
why
> > the switch off failure occurs?
> >
> > Thanks.
> >
>

You may be getting RF oscillation on the MOSFET. Try placing a ferrite
bead directly om the gate connection. Possibly add a 10 ohm series
resistor also.

Make sure that the flyback spike does not exceed the rating of the
MOSFET. Connecting a series diode/zener combination between the drain
& gate of the device may help. If the drain voltage is sufficient to
make the zener conduct, the MOSFET turns back on & clamps the spike.

RP

-- =

http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist