Sinking Heat From a Surface Mount Footprint to the Chassis

That most likely dominate.

A very low Rds-on fet could dissipate more power than a smaller device, if switching losses dominate.

Care to explain that? Makes no sense to me. I stock a low-voltage SOT23 that has lower Rds-on than a stocked TO-247 high-voltage part.

What I meant was that a low resistance fet may well have higher capacitances than a smaller geometry part with more Rds-on. The higher-capacitance part could dissipate more dynamic power itself, and in its gate driver, than it gains in static dissipation.

But you knew that, or sure should have.

It is PWM switched at 10kHz. It is a buck switcher.

Why so low? The inductors will be big, and it will be acoustically noisy.

I used them in a 500kHz x 4-phase boost. Not exactly that P/N but it worked well. ...except the one that caught fire (but the FET only did what it was told to do). :-(

Den tirsdag den 5. august 2014 20.32.27 UTC+2 skrev John Larkin:

and hot, coil DCR will be much more than the fet

-Lasse

You still haven't explained that LOL thing.

No it couldn't. If we're talking about i^2r loss in the FET (as we are), the lower Rds(on) FET has *lower* dissipation.

If you meant "a bigger FET can safely handle more heat" you're mostly right, but that's not what we're talking about.

Cheers, James Arthur

I know this subthread is not very serious, but seriously: why GaN? Its bandgap is as high as 3.4V, while e.g. InAs has just 0.36V. Wouldn't it be a perfect material for high-current diodes, far better than the current silicon Schottky rectifiers? PbSe is even better with its 0.27V.

Best regards, Piotr

Exactly, it's for high voltage device (500V+).

You are looking at the wrong direction.

I knew that; it's in The Cliche' Dictionary.

Explain the claim about fets.

There are SiC Schottky'es already available, which is much less exotic than GaN.

The forward voltage of a semiconductor diode is ~1*E_bandgap (PN) or 0.5*E_bandgap (Schottky). What's wrong with that? A PbSe rectifier would have 0.135V of forward voltage, which would be good at low voltage/high current applications. Otherwise one needs synchronous rectification, which is more complex and much more prone to failures.

Best regards, Piotr

Not high enough voltage or low enough resistance (Ron). GaN can potentially be better, but remain to be proven of course.