Infineon CoolGaN Class D amplifier EVAL_AUDAMP24 vs Purifi?

[Matias]

To me GaN is a lot of hype and I am yet to see it's ultra low distortion turn to reality.

 

[ZestClub]

To me GaN is a lot of hype and I am yet to see it's ultra low distortion turn to reality.

What about the Technics SE-R1 which utilizes the faster 1.5mHz switching speed and higher filter available to GaN users, instead of Mosfet with 3x lower switching speed and filter?

 

[kipman725]

When you design a Class-D amplifier though you don't want any higher switching frequency than is required to meet the performance requirement as losses increase with increased switching frequency. As Purifi has a maximum switching frequency of 500kHz this would seem sufficient (using their modulator) for any GaN amplifier. What switching to GaN in this circumstance gives you then is lower switching loss and cleaner switching events (reduced EMI).

 

[Wes]

To me GaN is a lot of hype and I am yet to see it's ultra low distortion turn to reality.

why is that?

 

[Matias]

why is that?

There is a lot of hype that because of the faster switching rate the distortion is a lot better than conventional class D. Sounds good, but I am still waiting to see a measurement proving this beating a Purifi amp.

What about the Technics SE-R1 which utilizes the faster 1.5mHz switching speed and higher filter available to GaN users, instead of Mosfet with 3x lower switching speed and filter?

Sounds great on paper. Where are the measurements?

 

[ZestClub]

When you design a Class-D amplifier though you don't want any higher switching frequency than is required to meet the performance requirement as losses increase with increased switching frequency. As Purifi has a maximum switching frequency of 500kHz this would seem sufficient (using their modulator) for any GaN amplifier. What switching to GaN in this circumstance gives you then is lower switching loss and cleaner switching events (reduced EMI).

That's what the new Technics-SU-R1000 would appear to be doing - it should be out by the end of this year.

 

[ZestClub]

This is the new Technics SU-R1000.
I am not sure how some of these figures compares to the Purifi?

 

[boXem]

This is the new Technics SU-R1000.
I am not sure how some of these figures compares to the Purifi?
View attachment 95248

On the left it's difficult to judge since there is no vertical scale, but this is anyhow only a problem for the certification of the product, as a user you don't care.
On the right I fear that the curve is worth nothing because the FFT size is unknown. With a big FFT size, I can give you a curve from my amplifiers with a noise level at -140 dbV if I wish. The only thing we can understand is that the orange is a lot cleaner than the blue one

 

[ZestClub]

On the left it's difficult to judge since there is no vertical scale, but this is anyhow only a problem for the certification of the product, as a user you don't care.
On the right I fear that the curve is worth nothing because the FFT size is unknown. With a big FFT size, I can give you a curve from my amplifiers with a noise level at -140 dbV if I wish. The only thing we can understand is that the orange is a lot cleaner than the blue one

Technics quote: 300W + 300W (1kHz, T.H.D.0.5%,4Ω, 20kHz LPF)

The brochure below is quite informative regarding areas like - not feeding back the original signal but only the distortion component.
And correcting problems associated with the variation in inductance of the speakers and the output filtering.

The product brochure can be downloaded by clicking the amp image here:
https://www.remusic.it/EN/New-Technics-SU-R1000-Reference-Class-integrated-amp-98dbd700

 

[Wes]

There is a lot of hype that because of the faster switching rate the distortion is a lot better than conventional class D. Sounds good, but I am still waiting to see a measurement proving this beating a Purifi amp.


Sounds great on paper. Where are the measurements?

your hype comment only applies to the implementation - if at all

not to the technology itself

 

[JohnYang1997]

Technics quote: 300W + 300W (1kHz, T.H.D.0.5%,4Ω, 20kHz LPF)

The brochure below is quite informative regarding areas like - not feeding back the original signal but only the distortion component.
And correcting problems associated with the variation in inductance of the speakers and the output filtering.

The product brochure can be downloaded by clicking the amp image here:
https://www.remusic.it/EN/New-Technics-SU-R1000-Reference-Class-integrated-amp-98dbd700

No remotely near purifi, not even in the same universe probably.

 

[Matias]

your hype comment only applies to the implementation - if at all

not to the technology itself

I may be wrong, sure, but as of today I am still waiting to see GaN being the decisive factor that makes a class D amp measure a lot better than everything else... And I have not seen it yet.

 

[kipman725]

I may be wrong, sure, but as of today I am still waiting to see GaN being the decisive factor that makes a class D amp measure a lot better than everything else... And I have not seen it yet.

If the switching frequency is the same as a Si MOSFET implementation then all your getting in terms of fidelity is perhaps less dead time and 'more square' output at half bridge mid point. This will improve THD but not drastically.

 

[ZestClub]

No remotely near purifi, not even in the same universe probably.

It would be interesting to have a side by side comparison between the Purifi and the Technics - both Class-D but coming at it from different technological perspectives. I have heard the Technics via a YT video and it sounds amazing on my $10 computer speakers...

 

[JohnYang1997]

It would be interesting to have a side by side comparison between the Purifi and the Technics - both Class-D but coming at it from different technological perspectives. I have heard the Technics via a YT video and it sounds amazing on my $10 computer speakers...

For acceptable sound 1% distortion is mostly fine. The objective performance is hundreds of times better on purifi. It's like comparing a boat to a ship.

 

[boXem]

Technics quote: 300W + 300W (1kHz, T.H.D.0.5%,4Ω, 20kHz LPF)

The brochure below is quite informative regarding areas like - not feeding back the original signal but only the distortion component.
And correcting problems associated with the variation in inductance of the speakers and the output filtering.

The product brochure can be downloaded by clicking the amp image here:
https://www.remusic.it/EN/New-Technics-SU-R1000-Reference-Class-integrated-amp-98dbd700

Thanks for this.
I initially didn't look at the performance data and was positively interested in seeing another class D with feedback handled with an ADC.
The issue is that in opposition to Axign (https://www.axign.nl/) they stay stuck on the external oscillator solution for modulating the signal. This means that to keep the amplifier stable in all conditions, the loop gain cannot be above 3 to 4 dB, which is extremely low. So is the performance of a class D amplifier with such a loop gain. So on objective performance, this amplifier just doesn't hold a candle against Purifi, Hypex, some TI chips, and others...
But it's a beauty

 

[ZestClub]

Thanks for this.
I initially didn't look at the performance data and was positively interested in seeing another class D with feedback handled with an ADC.
The issue is that in opposition to Axign (https://www.axign.nl/) they stay stuck on the external oscillator solution for modulating the signal. This means that to keep the amplifier stable in all conditions, the loop gain cannot be above 3 to 4 dB, which is extremely low. So is the performance of a class D amplifier with such a loop gain. So on objective performance, this amplifier just doesn't hold a candle against Purifi, Hypex, some TI chips, and others...
But it's a beauty

Thanks for the feedback - so do you think that the way forward is the Axign way with ADC feedback + internal (self-oscillating) solution for modulating the signal +GaN ?

To prevent quality loss of ADC/DAC I guess its possible to use the digital component just for feedback only and keep the main signal passing through the amp as analog + modulator ?

 

[boXem]

Thanks for the feedback - so do you think that the way forward is the Axign way with ADC feedback + internal oscillator solution for modulating the signal +GaN ?

I am not pretentious enough to think I know the way forward .
I have been attempting to get the Axign datasheet to better understand how it works (and see if we could do something with their chips) but these people don't even bother to look at small companies like mine. ADC feedback is good for an all in one chip that receives PCM (or DSD). For the rest it is just adding more issues to be solved (mainly handling of delays). So for analog inputs, purely analog solutions like the ones from Hypex, Purifi, etc are working perfectly.
The problem of the oscillator is not to be external or internal, it is it's own existence. The issue of a loop is that if the loop gain is too high, the amplifier becomes unstable and oscillate. If the loop gain is low, the loop is purely decorative. These ocsillations adding to another oscillator just create a non working amplifier. Bruno Putzeys has had the idea of the self oscillating amplifier 15-20 years ago (UcD amplifiers). The oscillator is suppressed and a voluntary unstable amplifier is created. Its working very well and is used by almost all high performance class D amplifiers.
About GaN, lower switching noise makes the engineering work easier for certification. Ok. For the customer, I don't get it.

 

[kipman725]

Thanks for this.
I initially didn't look at the performance data and was positively interested in seeing another class D with feedback handled with an ADC.
The issue is that in opposition to Axign (https://www.axign.nl/) they stay stuck on the external oscillator solution for modulating the signal. This means that to keep the amplifier stable in all conditions, the loop gain cannot be above 3 to 4 dB, which is extremely low. So is the performance of a class D amplifier with such a loop gain. So on objective performance, this amplifier just doesn't hold a candle against Purifi, Hypex, some TI chips, and others...
But it's a beauty

I also asked for a datasheet from an email which should have weight and didn't get a response.

 

[boXem]

I also asked for a datasheet from an email which should have weight and didn't get a response.

Well, I had a response asking for potential volumes, and then I ceased to exist