MOSFET Madness

[drewl]

I'm seeing more and more powered speakers with blown Fets in the power supply and power amp

Look up specs and they're 500-600 volt, 25-30 amps, out of TO220 devices?

You're talking crazy man!

Sorcery! Wizardry!

[Jon Snell]

Please rtead and understand the data sheets. Don't ignore the listings.
They can pass 25Amps only momentarily. The duty cycle at that level is less than 10%. That means they are looking at around 1/10th of the power = 2.5Amps on average with a S - D resistance of a few mili Ohms so not a problem..

[J M Fahey]

Well, the problem arises IF they are asked to deliver 5 or 10A ... on average.

IŽd expect 10% duty cycle on, say, a TV flyback signal but not on Audio duty, where I expect 50% when amp is putting out full power.

[Enzo]

Well he did say in the power supply, which is at least a cousin to the TV flyback. SMPS: That is spiky stuff, not like audio.

[SoulFetish]

Plus, traditionally, all fet power ratings were listed with the presumption that it is mounted to an "infinite" heat sink.

But here is the reason you see those power/performance ratings in increasingly smaller packages. Modern advancements in semiconductor technologies have resulted in some really impressive R_DS numbers. So, with lower on - drain to source resistances, you can pass higher currents with lower dissipation.

[J M Fahey]

True, power supply is "usually" less than 50% (each) duty cycle to allow for regulation .... which means that when they are into conduction they pass proportionately more current than average so in the long run it all evens out.

If amp needs, say, 6A on average, each MosFet will pass, say, 10A during 30% of the time instead of needed 6A during 50% under ideal conditions.

My point being that switchers live a harder life than what appears on first sight, and now and then .... sh*t happens.
Oh well.