W
westnet
Guest
Hi,
I am attempting to use an 'unprotected' (no zener or TVS) N-channel MOSFET
in avalanche mode, (it is avalance rated) to switch an inductive load.
My application is a Peak and Hold, current mode PWM solenoid driver which I
have designed. It is a 'fancy' design in that I have a switchable current
recirculation path (schottky diode from load, in series with an upside down
P channel FET connected to +ve rail) for efficient Peak and Hold operation,
but at turn off this path is 'opened' (PFET is turned off) and I use a
'zenering' down of current to achieve fast turn off. I use avalance
breakdown of the FET to achieve a fast turn off. (as voltage rises to about
80V in avalanche mode), though the FET is Speced at 60V max Vds. I was
planning on using a 60v zener, but decided to try avalanche mode first.
I have never used / relied on avalanche breakdown before so am new to it.
Checking out current / voltage waveforms for my circuit I find I the NFET is
only absorbing about 4% of the maximum avalanche energy at solenoid turn
off. But I am killing N-channel fets! (The low side switch that absorbs
the avalanche energy) . Sometimes they work for hours, sometimes 5 seconds!
When not blowing up, the N FETs are barely luke warm (from 2A average
current or so). Gate waveforms have a fast risetime. There are very few,
very short gate pulses (during peak to hold current transition) that only
charge the gate to the point where the Cgd effect ('appears' if you know
what I mean.).
Any ideas appreciated,
Thanks, Nick
I am attempting to use an 'unprotected' (no zener or TVS) N-channel MOSFET
in avalanche mode, (it is avalance rated) to switch an inductive load.
My application is a Peak and Hold, current mode PWM solenoid driver which I
have designed. It is a 'fancy' design in that I have a switchable current
recirculation path (schottky diode from load, in series with an upside down
P channel FET connected to +ve rail) for efficient Peak and Hold operation,
but at turn off this path is 'opened' (PFET is turned off) and I use a
'zenering' down of current to achieve fast turn off. I use avalance
breakdown of the FET to achieve a fast turn off. (as voltage rises to about
80V in avalanche mode), though the FET is Speced at 60V max Vds. I was
planning on using a 60v zener, but decided to try avalanche mode first.
I have never used / relied on avalanche breakdown before so am new to it.
Checking out current / voltage waveforms for my circuit I find I the NFET is
only absorbing about 4% of the maximum avalanche energy at solenoid turn
off. But I am killing N-channel fets! (The low side switch that absorbs
the avalanche energy) . Sometimes they work for hours, sometimes 5 seconds!
When not blowing up, the N FETs are barely luke warm (from 2A average
current or so). Gate waveforms have a fast risetime. There are very few,
very short gate pulses (during peak to hold current transition) that only
charge the gate to the point where the Cgd effect ('appears' if you know
what I mean.).
Any ideas appreciated,
Thanks, Nick