The Uber eGaN Half-Bridge Module

[luka]

soldering should be easy if you do it dead bug style

[egberttheone]

I have read the thread but it is no clear to me what you want to achieve.
Do you want more efficiency?
Smaller dead time ?
Less jitter ?

I'm interested in a low jitter mosfet bridge, the best driver i could find has 1ns of jitter which is pretty much.

Nice work BTW!

[elevator]

Originally Posted by egberttheone

Do you want more efficiency?
Smaller dead time ?
Less jitter ?

All of these, basically.
In a nutshell, eGaNs promise to deliver music cleaner, faster and more efficiently.
But to achieve these goals, electrical control must be tight, and mechanics also. The module embeds some (not all!) critical pieces while minimizing EMC etc. issues.
It should also make it somewhat easier to design a high-quality amp, so that eGaNs become accessible even to DIYers: most of them have trouble with SMD let alone bare dies...

Originally Posted by egberttheone

Nice work BTW!

Thanks, work in progress!

[elevator]

Originally Posted by Baldin

How about the topside of the mos is it isolated, so that you can just press it towards a heatsink, to get proper cooling?

Kind regards Baldin

Hi Baldin,
sorry if this is late. Yes, the substrate (topside) is isolated. EPC has an appnote on assembling and soldering etc., pretty interesting and IMO a must read before kidding around with these critters.
It's here: http://epc-co.com/epc/documents/prod...aNassembly.pdf

[smms73]

Originally Posted by egberttheone

I'm interested in a low jitter mosfet bridge, the best driver i could find has 1ns of jitter which is pretty much.

Hello egberttheone.

I think the only drivers that have that much jitter are those that are based in optocouplers or digital isolators, normal drivers like irs2011 for example, don't specify jitter in data-sheet , how do you get that numbers for jitter, have you measure?

[egberttheone]

I wish i had the ability to measure jitter. I found some jitter analyzer equipment but most of these can only measure as low as 1ns.
I would need one that can measure in pico seconds.
These are VERY expensive.

I found these jitter numbers in the datasheet, the driver did make use of opto isolators.
If there are no opto isolators involved what can i aspect in terms of jitter?
Most datasheets dont specify jitter.

[ChocoHolic]

Originally Posted by EPCguy

Thx for the warm welcome. I'll try to answer as much as I can.
There are many options for improving the 'usability' of our devices.
Wider pitch, better miller ratio, higher VTH and VGSmax, larger / better thermal package etc. Some are easier to do than others and are being looked into...

Pretty glad to hear this, because when your devices became available through digi key - I had a look to them and decide not to use them.
Mostly because of low Vth and insufficient capability to handle losses.
Of course we do not want to have losses, but my assumptions were as follows:
+/- 75V rails, half bridge 8 Ohms load.
I trust that I can design a PCB that allows power stage designs with di/dt around 1000A/us.
Consequently my designs would need approx 10ns for commutation at a 10A load. At a 1MHz switching frequency this would result in 7.5W losses in the device. My personal guess was that I would not be able to transfer this heat from the die to the heat sink.

Finally I decided, to go for MosFets in my current design and stay with 400kHz. In a few years GaN is likely to become mature enough for my purpose - I will wait for this.

Nevertheless, I am loving this thread. It is pioneering the application of future power switches. Absolutely great!!!

[elevator]

Originally Posted by ChocoHolic

... Absolutely great!!!

Thanks Choco. As they say, road works ahead Will keep you posted
E

[bwaslo]

Has anyone gone further on using these eGan FETS as class-d amps? Wondering how it worked out...

[elevator]

still sweating over it...