How will GaN change the world of electronics design in the next decade?

RE: How will GaN change the world of electronics design in the next decade?

colporteur — Sun, 01 Sep 2019 11:08:38 GMT

Establish a panel of long tooths, those individuals around when glass transistor were the norm and silicon replacements were coming out of the wings, and ask them the same question regarding silicon?

As one of the long tooths, I sometimes sit in awe at the opportunities silicon introduced.

GaN offers a chance to improve the way some things have been done but also those things never been attempted because of silicon limitations. The formula for development just got more expansive by adding another variable.

Hurts my head to think about it.

Sean

RE: How will GaN change the world of electronics design in the next decade?

ralphjy — Fri, 30 Aug 2019 19:29:51 GMT

I agree with [mention:acaf6a9338de4eef8f6717d5561ed01d:e9ed411860ed4f2ba0265705b8793d05] that GaN will have specific applications but will not replace silicon. GaN has some similar issues to other III/V semiconductors that we used in the past when we were trying to achieve higher frequency performance- GaAs and InP. These materials are hard to scale in terms of wafer size and circuit density and because of higher cost.

RE: How will GaN change the world of electronics design in the next decade?

Jan Cumps — Fri, 30 Aug 2019 10:17:59 GMT

I've been validating GaN power FETS for a few years now.

Checking Out GaN Half-Bridge Power Stage: Texas Instruments LMG5200 - Part 1: Preview

I think they have a place together with silicon, won't replace it.

GaN has some very good attributes that help fast switching - perfect for digital switching and switch mode power applications - and for some RF solutions.

But GaN is several orders of magnitude more expensive to make than silicon (sand) - and that will be so forever.

It has disadvantages too compared to silicon, making the devices more sensitive for particular failure modes and not the best choice for use in the linear (non-switched) range.

RE: How will GaN change the world of electronics design in the next decade?

dougw — Fri, 30 Aug 2019 04:58:29 GMT

I am looking forward to using GaN devices in the near future because of their lower on resistance. It looks like you could come up with something closer to an ideal diode than what can be done today.

GaN devices have many great performance features which allow a device equivalent to a silicon power mosfet to be smaller, faster, higher voltage, higher temperature, etc. However the simpler GaN devices are depletion mode devices (normally on and they need a negative voltage to turn off) E-GaN (enhancement mode GaN) devices are possible, but not all semi houses have a good method to make them and patents are slowing adoption. It is also not yet feasible to shrink GaN devices as small as silicon transistors, so large scale integrated circuits are a ways off. It will take time to solve the scaling issues and then it will take time for the patents to expire before we see really cheap GaN chips on the market. In the mean time there will be lots of applications where the performance gains make it worth the extra cost to use Gan devices.

There is lots of money being spent on the race to get feasible chips first, so maybe breakthroughs will occur faster than most are predicting, and prices may erode faster than the patents, but I still expect prices will be at a premium for a while.

Of course this is all speculation and it is the internet, so I could be full of hot air...[View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-7e1284b6-a791-42ab-9ee0-25f98561b347/contentimage_5F00_1.png:16:16]

RE: How will GaN change the world of electronics design in the next decade?

shabaz — Fri, 30 Aug 2019 03:20:33 GMT

I'm not too clued-up on the physics of it all, but I'm especially interested in GaN LEDs. Some manufacturers concentrated on that, presumably because they'd invested in GaN technology for power applications too, and it takes years to refine their processes. Whereas other manufacturers might have started with a different technology (e.g. Silicon Carbide) and are expanding into that. Apparently there's many variants too, like GaN devices grown on SiC, or grown on sapphire, and again the manufacturers have to spend years /decades refining things, not knowing for sure if the research will bear fruit, or if an alternative technology will take over (especially if the processes to work with it end up cheaper or more reliable for instance).

So, it's very exciting, continually seeing the improvements in light quality and efficiency that they're achieving.