What the GaN?

[jeffhenning]

Hey, it was inevitable that something was going to supplant silicon as the material for transistors, audio or otherwise.

So now we have gallium nitride…GaN. All good and fine, but, other than for marketing, there is really no reason for any consumer of audio equipment to get the least bit excited about equipment made with it.

Going through all of the electronic equipment that gets tested here that winds up being in the upper echelon of the ratings, none (to my knowledge) use GaN transistors. The units aren’t ridiculously expensive either and they’re world class, SoTA, etc.

If using GaN technology was going to make them half the price with the same or better performance, sure, I’d see the excitement. Unfortunately, though, that is not the case.

It really is nothing but another buzzword used to sell equipment. I’ve been an audio nut for almost 50 years and the marketing still really hasn’t changed at all.

 

[Doodski]

Hey, it was inevitable that something was going to supplant silicon as the material for transistors, audio or otherwise.

So now we have gallium nitride…GaN. All good and fine, but, other than for marketing, there is really no reason for any consumer of audio equipment to get the least bit excited about equipment made with it.

Going through all of the electronic equipment that gets tested here that winds up being in the upper echelon of the ratings, none (to my knowledge) use GaN transistors. The units aren’t ridiculously expensive either and they’re world class, SoTA, etc.

If using GaN technology was going to make them half the price with the same or better performance, sure, I’d see the excitement. Unfortunately, though, that is not the case.

It really is nothing but another buzzword used to sell equipment. I’ve been an audio nut for almost 50 years and the marketing still really hasn’t changed at all.

I think first it was gallium arsenide doping and now nitride. The arsenide doping made the RF bandwidth operation much improved and I suspect the nitride doping is even better. How that translates into audio bandwidth I don't understand either but progress is being made.

 

[Killingbeans]

GaN mostly has the potential to make devices smaller and more energy efficient.

The idea of GaN FETs having a "sound" is completely nuts, IMO.

It will probably have its most effective use in audio products outside of the hobby. Lifestyle devices with lots of power in a small package and better reliability than what we see today, because of easier thermal management.

Would likely also be beneficial to the reliability of the internals of powered active speakers. Although, I imagine most audiophiles don't care about those anyway.

I'm not surprised that GaN is being shoehorned into the hobby as a "mythical creature". It's the only way to get sales in a place that's already neck deep in fairy tales

GaN is still in its infancy. It will take some time for it to become really cheap, and even then silicon is probably cheaper yet.

One thing I've noticed is reports about discrete GaN FETs being somewhat electrically fragile, in the sense that they don't have much leniency towards the circuit driving them. More and more companies seem to be addressing this by offering GaN-on-silicon solutions in the form of integrated chips with both a foolproof driver and a power GaN FET on the same die.

If I were to make a new design, I'd pick one of those chips and not look back.

I bet we'll see new myths formed in the future about discrete GaN FETs "sounding better" than the ones with integrated drivers

 

[Doodski]

GaN mostly has the potential to make devices smaller and more energy efficient.

The idea of GaN FETs having a "sound" is completely nuts, IMO.

It will probably have its most effective use in audio products outside of the hobby. Lifestyle devices with lots of power in a small package and better reliability than what we see today, because of easier thermal management.

Would likely also be beneficial to the reliability of the internals of powered active speakers. Although, I imagine most audiophiles don't care about those anyway.

I'm not surprised that GaN is being shoehorned into the hobby as a "mythical creature". It's the only way to get sales in a place that's already neck deep in fairy tales

GaN is still in its infancy. It will take some time for it to become really cheap, and even then silicon is probably cheaper yet.

One thing I've noticed is reports about discrete GaN FETs being somewhat electrically fragile, in the sense that they don't have much leniency towards the circuit driving them. More and more companies seem to be addressing this by offering GaN-on-silicon solutions in the form of integrated chips with both a foolproof driver and a power GaN FET on the same die.

If I were to make a new design, I'd pick one of those chips and not look back.

I bet we'll see new myths formed in future about discrete GaN FETs "sounding better" than the ones with integrated drivers

The rise time improvements are not going to make audio sound better that's for sure. Gate voltage operation and operational tolerance improvements makes sense.

 

[GaryY]

One of well known application is military radar system.

The benefits and challenges of using GaN technology in AESA radar systems - Military Embedded Systems

militaryembedded.com

 

[MaxwellsEq]

There is a valuable thread here: https://audiosciencereview.com/forum/index.php?threads/is-gan-worth-it-icepower-article.51693/

 

[radix]

I think it's mostly used in power supplies, not the audio chain. At least from the posts and literature I've seen. I've only seen a trickle of audio devices using it in the audio chain, and those make "removes dead time" or "adds new dimensions" claims that I tend to write off.

 

[jeffhenning]

WHAT?!?!?!?

ARE YOU ASSERTING THAT A SQUARE WAVE DOESN’T EXIST IN NATURE?!

Seriously, though, as a student of the science of acoustics and music, as well as a musician and an audio nut, no audio circuit can create a true square wave because it has infinite bandwidth.

The square wave is to audio what faster than light travel is to physics… unattainable. And that’s OK.

A mangled square wave sounds better to the human ear anyway. It’s a staple of modern synthesizers.

What I’ve always been wondering is whether the one sample, 0dB digital audio pulse signal used to test DACs filter behaves when hit with the transient that can’t exist in real life is actually something that is in any way audible were you to take the output and amplify it to speakers?

My guess is no.

 

[MaxwellsEq]

WHAT?!?!?!?

ARE YOU ASSERTING THAT A SQUARE WAVE DOESN’T EXIST IN NATURE?!

Seriously, though, as a student of the science of acoustics and music, as well as a musician and an audio nut, no audio circuit can create a true square wave because it has infinite bandwidth.

The square wave is to audio what faster than light travel is to physics… unattainable. And that’s OK.

A mangled square wave sounds better to the human ear anyway. It’s a staple of modern synthesizers.

What I’ve always been wondering is whether the one sample, 0dB digital audio pulse signal used to test DACs filter behaves when hit with the transient that can’t exist in real life is actually something that is in any way audible were you to take the output and amplify it to speakers?

My guess is no.

Who on this thread has said anything about square waves? The person to mention it is you! Are you picking a fight with yourself?

What I’ve always been wondering is whether the one sample, 0dB digital audio pulse signal used to test DACs filter behaves when hit with the transient that can’t exist in real life is actually something that is in any way audible were you to take the output and amplify it to speakers?

This is an extremely confusing sentence. As written it makes little sense and does not describe the tests which are used to measure technology that converts PCM to analogue.

 

[Gregss]

Hello, As far as I can remember, GAN devices have the advantage of being able to make class d amps better by with the extra speed you can make the dead time even less than previously. Less distortion to correct for. Otherwise is just slightly more efficient transistor basically.

 

[amirm]

Hello, As far as I can remember, GAN devices have the advantage of being able to make class d amps better by with the extra speed you can make the dead time even less than previously.

In exchange for this, they cost a lot more. Proponents of MOSFET say that with sufficient feedback, you get the low distortion anyway so why spend the extra money on GaN transistors.