[baybal2]

> Faster switching speed for the equivalent breakdown, at least from an RF amplifier perspective. SiC was an interim until GaN was developed, and is essentially obsoleted now.

Wait wait wait... GaN and SiC have very, very different use profiles, and characteristics. SiC is by no means obsolete, and nor are purpose made silicon switches.

SiC is here to stay because of one very unique trait among all other semiconductors - a very pronounced negative temperature coefficient, and without extreme non-linearity.

Second to that is more or less linear threshold voltage temperature coefficient. GaN has positive threshold voltage temperature coefficient.

Third, SiC can simply operate at higher temperatures, and have known longer lifespan. GaN's current limits are WAY lower.

Fourth, transfer characteristic... Si, GaN, and SiC are all very different. This is one of few measures on which plain silicon beats contenders.

[madengr]

I was thinking in terms of RF amps. Cree was making SiC FETs several years ago, but I think it’s all GaN on SiC now. There was a SiC fab in CA making RF parts, but I think they went belly-up. All the SiC FETs topped out a 2.5 GHz.

How does the negative temp coefficient help; stability for very high temps? I know the positive temp coefficient was an issue in RF BJT, requiring ballast resistors for stabilization, but those are not needed in any FETs. The positive temp coefficient should be useful for GaN as long as you have temp stabilization in the bias network; all of my amps did.

Good point about the high temp. I know of some oil drill electronics in SiC. I just though SiC was dead for RF.

[baybal2]

> How does the negative temp coefficient help; stability for very high temps?

You can put multiple switches in parallel, and have them self balance without resorting to active temperature compensation which is completely out of question for any consumer grade device.

For power electronics, GaN is nowhere near as big of a bang as SiC, with its current handling being the primary showstopper. SiC can switch 100A loads at one kilovolt and above with ease, and at very high frequencies. There are simply no equivalent GaN part for this comparison.

Second to that, GaN needs a tricky gate driver, and is normally an n-channel depletion mode device. SiC can still be driven driverless at lo

Second to that, SiC JFETs still have niche uses in audio amps exactly because of their "bad" shallow IV curve.