Looking for a fet

[Rob]

Has anybody got a clue on how I might get or even perform a search for a high voltage power fet (>250v), preferably n-channel, that
doesn't have an internal zener? It must also have a low on-resistance.
I'm attempting to change the damping of a large back emf on the fly and the internal zener turns on defeating the exercise in any
configuration I can imagine. The fet will be protected by other means from over voltage so this isn't an issue. Early line output
transistors needed an external diode, does anyone know of a fet that was produced without one built in? This has had me stumped for
years and I would like to sort it out. Patents also show a fet being used in this configuration so they must exist somewhere.
Rob.

 

[Phil Allison]

"Rob" <

Has anybody got a clue on how I might get or even perform a search for a

high voltage power fet (>250v), preferably n-channel, that doesn't have an
internal zener? It must also have a low on-resistance.


** The gate zener is essential to protect the gate junction.

I'm attempting to change the damping of a large back emf on the fly and

the internal zener turns on defeating the exercise in anyconfiguration I can
imagine. The fet will be protected by other means from over voltage so this
isn't an issue. Early line output transistors needed an external diode, does
anyone know of a fet that was produced without one built in?


** The Drain - Source diode in a mosfet is an inherent part of the
structure, it cannot be taken out.

You can always put a diode is series with a mosfet to prevent reverse
conduction.



.............. Phil

 

[Bob M]

Rob,

Phill is correct. The drain-source diode is not an "added feature" but
is the result of the structure of the FET itself. You can think a FET
as a diode that has an extra connection (the gate) that allows you to
adjust the PIV rating of the diode from very high (the "off" state) to
very low (the "on" state).

Faced with a similar challenge as you are, my solution was exactly
what Phill has suggested. Use a high voltage ultra fast shottky diode
in series with the drain lead of the N channel FET. The effect of the
drain-source diode will only last as long as it takes the shottky
diode to turn off. Depending on your choice of diode, 10 nsec would be
possible. International Rectifier was my choice.

Cheers from Canada!

Bob M.

 

[Rob]

Thanks for the response. Instead of switching on the fly, I have fitted a
diode so that it can conduct at will as a test without the fet in place, and
unfortunately this hasn't worked as the diode presents a resistance and
doesn't conduct all the way to zero.
I haven't tried a high power shottky diode but now will if I can get a
suitable one easily. Like I said, the circuit must have a very low
on-resistance or it won't work. If I do get this working then I'd like to run
this by someone to see if they can explain the effect as it is just plain weird
but highly beneficial to a certain application. I'd say it is definately novel!
The only way I've been able get this working was accidental and by
using a mechanical means which was a very short section of a burnt track in
a small carbon pot. It appears the burnt track was ignored and couldn't pass
current at coil switch on but conducted extremely well when hit by the back
emf and ionisation probably did the rest. Obviously this idea isn't practical
and still can't be switched in and out on the fly.
Thanks again for your input,
Rob.

Rob,

Phill is correct. The drain-source diode is not an "added feature" but
is the result of the structure of the FET itself. You can think a FET
as a diode that has an extra connection (the gate) that allows you to
adjust the PIV rating of the diode from very high (the "off" state) to
very low (the "on" state).

Faced with a similar challenge as you are, my solution was exactly
what Phill has suggested. Use a high voltage ultra fast shottky diode
in series with the drain lead of the N channel FET. The effect of the
drain-source diode will only last as long as it takes the shottky
diode to turn off. Depending on your choice of diode, 10 nsec would be
possible. International Rectifier was my choice.

Cheers from Canada!

Bob M.