I recently had a MOSFET blow in one of my amps (unfortunately I had no extra so I'm waiting for more), but I know not why. It's strange because I've got four channels built at the same time and only one FET in one channel blew. It was an IRFP044N running at 30V as a source-follower, it had been in faithful use for about a year now and all of a sudden it just went. it didn't fry or blow up but the resistance between the gate and drain went down to a few ohms.
What kills Mosfets?
Too much voltage
Too much current
Too much voltage and current (that is, too much wattage)
Too high Gate to Source Voltage (20 V max)
Oscillation from no Gate resistor (try 150-270 ohms)
And sometimes they just croak.
Too much voltage
Too much current
Too much voltage and current (that is, too much wattage)
Too high Gate to Source Voltage (20 V max)
Oscillation from no Gate resistor (try 150-270 ohms)
And sometimes they just croak.
Source Followers
I have always had problems when using mosfets as source followers when the power supply voltage is higher than the gate to source breakdown voltage of the mosfets. It is just too easy to exeed that rating as the source tries to keep up with the gate. It is critical in source follower configuration for mosfets to be very careful that the gates are not overdriven by too much voltage.
I have always had problems when using mosfets as source followers when the power supply voltage is higher than the gate to source breakdown voltage of the mosfets. It is just too easy to exeed that rating as the source tries to keep up with the gate. It is critical in source follower configuration for mosfets to be very careful that the gates are not overdriven by too much voltage.
Well, thanks alot for the replies. I really wanted to know what the possible reasons are. I'll replace the MOSFET and in time if it happens again (I'm hopping as Nelson said one of the possible reasons was, that it just croaked) I'll know where to look, but I've got my eyes on the gate to source voltage... Thanks again.
JoeBob,
The possibility of oscillation cannot be overlooked.
You may have oscillation, which in a push pull circuit usually leads to cross conduction, which can easily kill mosfets through overcurrent.
Like Nelson, I specify 150-270R, usually use 220R. Gate leads must be short, and there needs to be some film caps on the drain to damp any HF oscillation.
At protection for the sensitive gate/channel interface, I suggest a 9.1V zener, cathode to gate on a n type. Often a diode is placed beneath the zener, cathode to source on a n type.
Cheers,
Hugh
The possibility of oscillation cannot be overlooked.
You may have oscillation, which in a push pull circuit usually leads to cross conduction, which can easily kill mosfets through overcurrent.
Like Nelson, I specify 150-270R, usually use 220R. Gate leads must be short, and there needs to be some film caps on the drain to damp any HF oscillation.
At protection for the sensitive gate/channel interface, I suggest a 9.1V zener, cathode to gate on a n type. Often a diode is placed beneath the zener, cathode to source on a n type.
Cheers,
Hugh
I was also thinking about the good point Nelson make about oscillation. I think oscillation could cause excessive voltage swings on the gates.
Well, I thought later about the oscillations and I've got a 100Ohm resistor on the gate of the MOSFET and all the leads are very short (the PCB is under a centimeter away from the heatsink, the MOSFET's connections to the PCB holds it there), but the stages before the FET might indeed be oscillating, so I looking at my tubes first and adding grid stopper resistors, maybe that'll have an effect.
Also...
While you are adding protection, have you got diodes between output and supply rails to counter inductive flyback of the speaker cables and speakers? A couple of 1N4007, reversed biased, is a standard precaution.
While you are adding protection, have you got diodes between output and supply rails to counter inductive flyback of the speaker cables and speakers? A couple of 1N4007, reversed biased, is a standard precaution.
- Status
- Not open for further replies.