I have made about 30 amplifiers using IRFP240/9240 MOSFET's
I aligned one of the 30 amplifiers setting the output dc offset and the bias current.
After about 5 minutes the fuses blew along with the output transistors despite running at a quiet level.
The blown transistors were very hot.
The other 29 amplifiers have had no problems at all.
I replaced the output transistors and the amp has run fine now for a week !
Does anyone know if MOSFET's do just sometimes pack up?
Or perhaps the bias current ran away for some reason ?
I aligned one of the 30 amplifiers setting the output dc offset and the bias current.
After about 5 minutes the fuses blew along with the output transistors despite running at a quiet level.
The blown transistors were very hot.
The other 29 amplifiers have had no problems at all.
I replaced the output transistors and the amp has run fine now for a week !
Does anyone know if MOSFET's do just sometimes pack up?
Or perhaps the bias current ran away for some reason ?
Did you notice any oscillation? Mos devices are very fast and can become an oscillator at many MHz if certain parts of the ‘internal circuitry’ are not dampened properly or may have bad contact. RF oscillation can appear as DC on a multimeter and can cause frustrating destruction.
Maybe useful to put up a schematic (omit values if it conerns you). Maybe there is some marginal stability issue and you just happened to have all the "wrong" values with that one.
The last time I had a mosfet amp blow up was one I was repairing and was a C-Audio something or other, it turned out I'd been sent fake mosfets. I just replaced them with expensive ones and it worked great.
Yes but this one failed on soak test after about 5 minutes.
I always soak test the amps I sell for many hours to make sure they are OK.
The VAS capacitor is 220pf which is much more than otehr people use.
It is probably overdamped rather than underdamped.
Yes I replaced the blown pair of MOSFETs and it has been fine since.
I have kept this pcb for myself for playing at home. If there is any doubt about an amplifier I dont let them out to customers.
The MOSFET's are marked IRFP so I assume they are good quality from IR.
My apologies for being vague, I was regarding to the mosfet itself outside of the CL, they can be great oscillators because of the lead inductance and the device capacitances which can resonate at MHz frequencies. Vertical types particularly. At 50+MHz range the lead inductance and PCB trace/layout can become problematic. Oscillations at this speed will pass right over the global loop and affect the bias stability of the mosfet. Most good mosfet amps have some sort of damping within close proximity other than just a gate stopper. If you find this type of instability to be a problem, you might try adding a gate Zobel filter to each fet. Maybe with a zero around 40MHz, say 27pf in series with 150R from gate to gnd. I find it can be very helpful in stabilizing mosfet followers, vertical types, hexfets and such.
Last edited:
Before powering up for the first time I always put a scope proble on the output to look for excessive dc or ac.
I havent seen many designs with filters on the MOSFET gates, there is usually just a resistor.
Finally got to the bottom of this problem.
Turned out to be an unsolfdered joint on the bais preset.
The amp was sat on top of a speaker so its likely vibration move the leg in the hole and gave the amp full bias volts and killed it.
Turned out to be an unsolfdered joint on the bais preset.
The amp was sat on top of a speaker so its likely vibration move the leg in the hole and gave the amp full bias volts and killed it.
Im using safe limited current at 1st testing.
Most failure in production are wrong component placed and soldering failure (both short and open).
That was very often.
Most failure in production are wrong component placed and soldering failure (both short and open).
That was very often.
Often amplifiers can errupt without load or reason -- other than out of the batch of components one rogue component may have got through all the quality checks; just so happens to go critical inside your amplifier.
With your burn in time you are operating a quality policy for your customers so they're not disappointed. This is really nice to see in an age of instant profiteering and throw away consumerism. 😉
It's better the issues manifest prior to getting to the client.
With your burn in time you are operating a quality policy for your customers so they're not disappointed. This is really nice to see in an age of instant profiteering and throw away consumerism. 😉
It's better the issues manifest prior to getting to the client.
- Status
- Not open for further replies.