Dave's powertrip #10: Next Round... GaN versus Si


In the late ‘80s, I was working on a method to use near infrared to determine the mineral content of reservoir samples. The goal was to provide an alternative to X-ray diffraction. I used a Fourier transform infrared spectrometer to determine the transmission spectra of potassium bromide pellets that contained very small amounts of core samples, a few milligrams. The data was collected by a PC and the corresponding spectra were mathematically compared to a set of calibration samples. Due to the amount of data and manipulation required, there was considerable benefit from a fast processor and math coprocessor. So I moved from an 80286- to 80386-based system with a math coprocessor. At the time, many considered the 80386 to just provide a faster DOS machine. Well, it certainly did. In this application, the data could be processed while the next sample was being prepared. The results were that we could do nearly 30 samples per hour. Yes, it was a faster DOS machine. 

We may be at the same crossroads for wide-band gap (WBG) power devices. For some time we have heard that next year will bring new high-voltage WBG FETs. The WBG devices have shown up as Gallium Nitride (GaN) and Silicon Carbide (SiC). These devices provide low RDSON with higher breakdown voltage. They also provide lower gate capacitance, compared to a silicon device with the same RDSON. There are other benefits like high-temperature performance and such. The first applications seem to focus on switching speed and power density. Switching speed or frequency reminds me of the faster DOS machine. Having a solution for greater speed or size reduction in the power passives is a good thing.

So where are we with these WBG devices?

Read remaining article at EDN.com

 

Semiconductor material overview*

*Philip Zuk, Vishay Siliconix – June 20, 2012, “High-Voltage Silicon MOSFETs, GaN, and SiC: All have a place.”

 

Power to the people! - Dave Freeman