Seeking Ideas for RoadTesting Gallium Nitride Semiconductors

Hi Randall,

I like the GaN capability, but I think this is one of those products that will have to be a well focused test.

I am thinking about the work that Jan did a year or so ago where he looked at the basic device characteristics.

I could see doing a good power supply test if you provided a standard PS and a GaN one. We could conduct tests against both to see the basic advantages of one verses the other.

As far as a charger test goes I see two issues. One is filtering out the battery dependencies. The other is looking at the overall efficiency of the GaN over a traditional charger from a system stand point. I would be interested in seeing if there was any improvement over the current systems as far as power consumed verses power delivered to the battery. I would also be interested in testing those efficiencies over various battery types.

It is clear that this type of testing is a step deeper than element14 has traditionally done as you are almost at the component level.

DAB

Don,

Interesting comment. Thanks. I'm thinking that a comparative test will probably be part of any roadtest.

Randall

Hi Randall,

I'd been channeled into thinking GaN technology was only for solid state RF power amplifiers, so by asking your question it has made me research what other possibilities are emerging for this technology, including your informative links. An great subject to watch and I'd be keen follow any roadtests closely.

On the maker side of things, which most of us can relate to, a comparaive test to build two H-bridge drivers could be an interesting test - one using silicon based devices and the other GaN devices. To drive an equal load the design of the GaN would be much smaller and with less heatsink requirements. Alternatively a more simple test would be make a motor driver that lifts a weight (weight on a string, wrapped around the motor shaft.). Again - showing how small a design could be made.

Rod

I’ve written a blog series on how to evaluate GaN FETs here on element14: Checking Out GaN Half-Bridge Power Stage

It’s about GaN devices from another manufacturer, but the approach isn’t vendor specific. It could be used to review GaN devices from any supplier.

One thing we strike here is controlling DC motors.

Yes there are some excellent OTS driver boards and chips, but sometimes these aren't the best solution.

(i used some Mosfet driver chips to directly control a DC motor in my 300g bot)

Perhaps the RT could control a large DC load using an Arduino and motor (something in stock).

The testing could identify the driving requirements, voltage drop, heat (and means to handle it), as well as EMF measures required.

I'm sure the results would suit anyone looking to build a robotic platform that falls outside the specs for OTS driver boards.

Mark

In my last trip to the pcim power messe in nuremberg I encountered new Gallium Nitride Semiconductors GaN devices. They were very interesting but mostly targeted the high power transistors.

A technology sample card that from the pcim. I would also be interested in such devices.

Cheers

snidhi  wrote:

 

In my last trip to the pcim power messe in nuremberg I encountered new Gallium Nitride Semiconductors GaN devices. They were very interesting but mostly targeted the high power transistors.

 

...

 

Yes, because of their attributes, they are particularly good for very fast switch mode applications. They can switch high voltage and currents. Because of the fast switch time, they operate less in the linear mode during switching -> less heat and more efficiency.

Been wanting to do a project with GaN or SiC devices for a while, it's a big leap from traditional mosfets or IGBTs as you need new everything.

The new power devices require new driver chips and related parts.

The technology is new enough there isn't much shared knowledge to look at.

If a kit of parts for a project were supplied I would really be interested, however it needs to be complete.

A) Driver board

B) GaN switches and CaN diodes

C) Snubber network and filter devices

Infineon know these devices better than most of us, so would know what sort of snubber and filter design is needed to keep things from self destructing.

If I recall GaN driver boards can be used with traditional silicon devices too, not got time to check at this exact moment.

As for the idea of

> "Another idea was to roadtest a gate driver and scope out the switching rates."

I don't have access to the equipment to do that at the moment, those fast rise times are higher frequency than they appear.

Plus I'd want to see how the driver performs under load, and that complicates matters (electrical isolation).

I'd expect faster rise/fall times to lead to more overshoot, undershoot and ringing.

Be interesting to see how this is managed in their designs, still an easier issue to solve than removing heat from semiconductors.

Hi Jan,

Do you think the GaN device could drive a Tesla coil?

DAB