• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

HEMPT GaN based Class D - What's the holdup?

PHD

Active Member
Joined
Mar 15, 2023
Messages
252
Likes
331
Hi

This little devil is a half-bridge power inverter, which we've designed and built at our power electronics lab for a large-scale Modular Multilevel Converter project. It is based on dual G-HEMT 650V GaN and two GaN power diodes that can operate at 1.5MHz and handle up to 65A of current without the need of a heatsink!


Why aren't Class D amps based on such a technology? This could bring class D performance and price to a new level

1727027718465.png
 
What is with what appears to be a very small mounting strap over this component? I have seen such mounting methods with high accuracy crystal oscillators.
1727027718465.png
 
What is with what appears to be a very small mounting strap over this component? I have seen such mounting methods with high accuracy crystal oscillators.
View attachment 394103
I'm not sure they're mounting straps. It seems like a grounding tap. I'll have to ask our lab engineer. The modules are custom-assembled by a Chinese vendor. We just send them the PCB-CAD files.


1727029227635.png
 
Last edited:
That appears to be a ground test point.
That makes sense. So it must be elevated over the thing that is soldered on the PCB there I imagine. If there is a thing soldered there at all. It's a tricky viewing angle.
 
Do GaN HEMTs offer significant advantages over GaN FETs now in use for power (and audio) circuits? I do not know, but am curious. My experience with HEMTs was for mW/mmW applications and have very little experience with GaN HEMTs; GaAs and InP HEMTs were the leaders back then. I know some varieties, while great at RF frequencies, exhibited high flicker or 1/f noise near DC that made them unsuitable for audio circuits (not as bad perhaps as traditional GaAs FETs). Less a problem in a class D amplifier where the noise bursts would (should) be small due to the short switching interval.
 
FAQ #8:
 
FAQ #8:
Nice article, I had forgotten you did that.

I have no data for audio amplifiers, but power supplies and battery chargers do exhibit higher efficiency and less heat dissipation in the devices using GaN over Si. I really doubt that it matters for an audio amplifier, given a small reduction in power loss is pretty much in the mud, but for applications where every little bit helps in efficiency and reduce thermal load they are useful.

As for measured performance, theoretically faster switching should help, but again capturing any useful advantage in a class D audio amplifier seems unlikely.
 
Infineon has an evaluation board. An ASR member tried it. https://www.audiosciencereview.com/...ss-d-amplifier-eval_audamp24-vs-purifi.17467/

Maybe someone will offer a finished boutique GaN amplifier?

I associate GaN, SiC, GaAs with high cost and frequency or high power we don't usually encounter in audio. You also see Indium and exotics in infrared image sensors. I just heard about a commercial 80GHz point to point radio. We need ever increasing power conversion AC>DC and DC>AC in the electric grid for solar inverters, bi-directional battery inverters, and ultra high voltage DC power lines.
 
Back
Top Bottom