Class D Audio GAN FET Amps

[AUDACC]

I would like to buy a quality American made (or assembled) ~200 watt/channel stereo amp. The American assembled amps using the Hypex and Purifi modules are in the range and have been testing well on Audio Science Review. However, there are no reviews of the Class D Audio GAN FET Amps. Has anyone here had experience with Amps from this manufacturer?

I bought a Topping D50 iii last month and played some downloaded HiRes recordings through the USB port on my Windows 11 computer. “Wow” DAC’s have improved! Especially at the $250 level. Thanks ASR Members.

 

[antcollinet]

Problem is Gan Amps have been failing to deliver any benefits - and tend to be sold at a premium price. None of the 4 GAN amps tested so far have achieved a "recommended", and the one tested from your preferred manufacturer scored the lowest Sinad of the 4.

Premium Audio Mini GaN 5 Review (Stereo Amplifier)

This is a review and detailed measurements of the Premium Audio Mini GaN 5 Stereo Class D power amplifier. It was kindly sent to me by a member and costs US $799 (recent price increase). The GaN 5 comes in a compact enclosure with plenty of ventilation at the cost of decent looks: Beside...

www.audiosciencereview.com

 

[AUDACC]

Problem is Gan Amps have been failing to deliver any benefits - and tend to be sold at a premium price. None of the 4 GAN amps tested so far have achieved a "recommended", and the one tested from your preferred manufacturer scored the lowest Sinad of the 4.

Premium Audio Mini GaN 5 Review (Stereo Amplifier)

This is a review and detailed measurements of the Premium Audio Mini GaN 5 Stereo Class D power amplifier. It was kindly sent to me by a member and costs US $799 (recent price increase). The GaN 5 comes in a compact enclosure with plenty of ventilation at the cost of decent looks: Beside...

www.audiosciencereview.com

Since the newest version of their offerings, the Premium GaN 6.5 Balanced 2 -Channel 200 watt/channel hasn’t been submitted for a review to show that the 2021 review issues have been addressed, I will pass on this amp for now. Thank you for responding.

 

[Hear Here]

I would like to buy a quality American made (or assembled) ~200 watt/channel stereo amp. The American assembled amps using the Hypex and Purifi modules are in the range and have been testing well on Audio Science Review. However, there are no reviews of the Class D Audio GAN FET Amps. Has anyone here had experience with Amps from this manufacturer?

I bought a Topping D50 iii last month and played some downloaded HiRes recordings through the USB port on my Windows 11 computer. “Wow” DAC’s have improved! Especially at the $250 level. Thanks ASR Members.

You don't say which brands you've tried or are considering. However I have a pair of Atma-Sphere Class D GanFET amps. They sound very good indeed. I also have a NAD M33 that uses the Purifi Eigentakt Class D amp, again very good indeed.

My main concern is how amplifiers sound when powering my Avantgarde Duo XD speakers and not with their measurements. I want to hear as near as possible the sound and excitement factor of the original performance. I previously had an exceptionally good measuring amp from an American brand that sounded not a patch on either of the amps I currently own. I can highly recommend both the Atma-Sphere and NAD amps. I hope this helps.

 

[MaxwellsEq]

I would like to buy a quality American made (or assembled) ~200 watt/channel stereo amp. The American assembled amps using the Hypex and Purifi modules are in the range and have been testing well on Audio Science Review. However, there are no reviews of the Class D Audio GAN FET Amps. Has anyone here had experience with Amps from this manufacturer?

I bought a Topping D50 iii last month and played some downloaded HiRes recordings through the USB port on my Windows 11 computer. “Wow” DAC’s have improved! Especially at the $250 level. Thanks ASR Members.

A properly designed GAN FET class D amplifier should sound identical to a properly designed amplifier of any class. GAN FETs in theory have the advantage of faster switching time, which in class D should allow higher switching frequencies, lower distortion etc. in practice, as yet, clear blue water has not opened up in terms of measured performance between GAN FET class D and other amplifiers. That doesn't mean to say it won't at some point.

 

[boXem]

At the exception of @orchardaudio all the GaNFET amps I know about have a load dependant frequency response.

 

[AUDACC]

At the exception of @orchardaudio all the GaNFET amps I know about have a load dependant frequency response.

Thanks for your and other replies.

I guess any amp has load dependency if pushed hard enough toward clipping. Would the following 2 amps be audibly load dependent if operated in the low distortion power ranges before approaching clipping?

boXem A 4216/E2 Stereo Amplifier Review

This is a review and detailed measurements of the boXem A 4216/E2 stereo balanced amplifier based on Purifi 1ET6525SA class D module. It was sent to me by the company and costs €1,160 (USD $1,350). While not built out of billet of aluminum and such, the 4216/E2 still manages to look unique and...

www.audiosciencereview.com

Apollon NCx500ST Stereo Amplifier Review

This is a review and detailed measurements of the Apollon NCx500ST stereo class D amplifier based on Hypex NCOREx NCx500 modules. It was kindly drop shipped from the member and costs 1,090.00€ (US $1,168) with standard opamp. While the case is stamped aluminum, it is a heavily customized one...

www.audiosciencereview.com

 

[goryu]

Open loop amps are by design load dependent. Early class d amps were almost all load dependent due to the lack of post output filter feedback. This was solved and later designs by the likes of Hypex, Mueta, and later Purifi, Orchard, etc. do indeed have post filter feedback and show load independent frequency response. This is considered state of the art, modern class d design. Most of the GaN amps use old, out of date designs and are simply sold on the advantages of the GaN fet over the standard fets used by Hypex, Purifi, and others. The problem, in addition to the use of outdated class d design, is that none of them really exploit the advantages inherent in GaN fets. Outside of Orchard, I am not aware of a single GaN amp that has performance anywhere near the state of the art amps by Hypex and Purifi. It's mostly just a marketing scam at this point. Read Bruno Putzey's (leading class d designer) comments on the topic available by google search.

 

[boXem]

Thanks for your and other replies.

I guess any amp has load dependency if pushed hard enough toward clipping. Would the following 2 amps be audibly load dependent if operated in the low distortion power ranges before approaching clipping?

boXem A 4216/E2 Stereo Amplifier Review

This is a review and detailed measurements of the boXem A 4216/E2 stereo balanced amplifier based on Purifi 1ET6525SA class D module. It was sent to me by the company and costs €1,160 (USD $1,350). While not built out of billet of aluminum and such, the 4216/E2 still manages to look unique and...

www.audiosciencereview.com

Apollon NCx500ST Stereo Amplifier Review

This is a review and detailed measurements of the Apollon NCx500ST stereo class D amplifier based on Hypex NCOREx NCx500 modules. It was kindly drop shipped from the member and costs 1,090.00€ (US $1,168) with standard opamp. While the case is stamped aluminum, it is a heavily customized one...

www.audiosciencereview.com

This is a load dependant frequency response

This is a load independant frequency response

In each case amplifier is used well within it's limits.

 

[AUDACC]

This is a load dependant frequency response

This is a load independant frequency response

In each case amplifier is used well within it's limits.

Agreed. Thanks.

 

[DVDdoug]

MOSFET and transistor amplifiers are already very good (even good tube amps). They are often better than human hearing so it's hard to make an improvement.

Since there are many ways to make a good amplifier, it usually comes-down to economics. If/when a GaNFET can give you more for your money, it will start to make sense.

The faster switching of GaNFET can make the amplifier more energy efficient (because the partially-on time is shorter) but class-D MOSFET amps are already efficient... The closer you get to 100% efficiency, the less of a difference you can make.

Or the faster switching can be used with a higher class-D switching frequency which could potentially give you better high-frequency response, but again MOSFET amps are already pretty good.

 

[AUDACC]

Open loop amps are by design load dependent. Early class d amps were almost all load dependent due to the lack of post output filter feedback. This was solved and later designs by the likes of Hypex, Mueta, and later Purifi, Orchard, etc. do indeed have post filter feedback and show load independent frequency response. This is considered state of the art, modern class d design. Most of the GaN amps use old, out of date designs and are simply sold on the advantages of the GaN fet over the standard fets used by Hypex, Purifi, and others. The problem, in addition to the use of outdated class d design, is that none of them really exploit the advantages inherent in GaN fets. Outside of Orchard, I am not aware of a single GaN amp that has performance anywhere near the state of the art amps by Hypex and Purifi. It's mostly just a marketing scam at this point. Read Bruno Putzey's (leading class d designer) comments on the topic available by google search.

Do you see any movement to incorporate the faster switching speed advantages of GaN in newer generations of class D amps from Hypex and Purifi? I am guessing that faster switching would allow higher frequency internal oscillators that would perhaps reduce the size and cost of inductors and capacitors.

 

[AUDACC]

Open loop amps are by design load dependent. Early class d amps were almost all load dependent due to the lack of post output filter feedback. This was solved and later designs by the likes of Hypex, Mueta, and later Purifi, Orchard, etc. do indeed have post filter feedback and show load independent frequency response. This is considered state of the art, modern class d design. Most of the GaN amps use old, out of date designs and are simply sold on the advantages of the GaN fet over the standard fets used by Hypex, Purifi, and others. The problem, in addition to the use of outdated class d design, is that none of them really exploit the advantages inherent in GaN fets. Outside of Orchard, I am not aware of a single GaN amp that has performance anywhere near the state of the art amps by Hypex and Purifi. It's mostly just a marketing scam at this point. Read Bruno Putzey's (leading class d designer) comments on the topic available by google search.

On your recommendation I read Bruno Putzeys’ article in the October 2024 Home Theater Hi Fi. It answered all my current questions about Class D amplifiers. The use of GaN FETs and higher switching frequencies boils down to a cost tradeoff between the more expensive GaN FETs and the possibly less expensive passive components. Since real world amplifier performance of the best class D amplifiers such as Purifi and Hypex seems to be more than adequate, there is no reason to increase the cost to gain performance that can’t be heard.

He does point out that because of the high cost to develop a new class D high quality amplifier in house, it is more efficient for the Hi-Fi manufacturers to add value to their products by integrating the power amplifier modules into their products tailored to specific applications like receivers, streamers, or integrated speakers.

Maybe I should wait to replace my inexpensive Sony receiver and circa 1987 ADS 1590 towers and with powered speakers.

 

[MaxwellsEq]

I guess any amp has load dependency if pushed hard enough toward clipping

That's wrong.

Load dependency is not to do with level or hard you drive the amplifier.

Load dependency means that the frequency response is only flat when using resistive loads at a certain defined resistance e.g. 4 Ohms. If you instead have pure resistance 8 Ohm speakers or any sort of inductance and capacitance (i.e. real speakers) the frequency response is not flat

 

[AUDACC]

That's wrong.

Load dependency is not to do with level or hard you drive the amplifier.

Load dependency means that the frequency response is only flat when using resistive loads at a certain defined resistance e.g. 4 Ohms. If you instead have pure resistance 8 Ohm speakers or any sort of inductance and capacitance (i.e. real speakers) the frequency response is not flat

I was using the term “load dependent” referring to the situation where an amp’s maximum output voltage driving 8 ohms will most likely be lower if the load is changed to 4 ohms. The term “load dependent” was used too loosely and therefore in a misleading way.

Thank you for that clarification.

 

[boXem]

was used too loosely

Load dependant frequency response is a term widely used in the industry but... yeah...

 

[Sir Sanders Zingmore]

I was using the term “load dependent” referring to the situation where an amp’s maximum output voltage driving 8 ohms will most likely be lower if the load is changed to 4 ohms. The term “load dependent” was used too loosely and therefore in a misleading way.

Thank you for that clarification.

Apologies if I’m wrong but these responses seem to me to have more than a hint of being A.I. generated…

 

[Mnyb]

Thanks for your and other replies.

I guess any amp has load dependency if pushed hard enough toward clipping. Would the following 2 amps be audibly load dependent if operated in the low distortion power ranges before approaching clipping?

boXem A 4216/E2 Stereo Amplifier Review

This is a review and detailed measurements of the boXem A 4216/E2 stereo balanced amplifier based on Purifi 1ET6525SA class D module. It was sent to me by the company and costs €1,160 (USD $1,350). While not built out of billet of aluminum and such, the 4216/E2 still manages to look unique and...

www.audiosciencereview.com

Apollon NCx500ST Stereo Amplifier Review

This is a review and detailed measurements of the Apollon NCx500ST stereo class D amplifier based on Hypex NCOREx NCx500 modules. It was kindly drop shipped from the member and costs 1,090.00€ (US $1,168) with standard opamp. While the case is stamped aluminum, it is a heavily customized one...

www.audiosciencereview.com

It has noting to do with the level it applies at all levels, it actually means that the amp has to high output impedance in the treble region hence the frequency response would deviate slightly from flat depending on which speaker you have , meaning it would sound different on different speakers .

 

[MaxwellsEq]

I was using the term “load dependent” referring to the situation where an amp’s maximum output voltage driving 8 ohms will most likely be lower if the load is changed to 4 ohms

Again, "maximum" has nothing to with it.

If I have 500W load-dependent amplifier that only has a flat frequency response at restive 4 Ohms, I will hear a non-flat frequency response into real speakers (inductors capacitors etc) at 30W, 130W, 2W, 0.1W, 17.67W etc.

 

[Bending]

I built mono amps with the new Orchardaudio Starkrimson 25 amp modules and they sound better used above 150HZ (analog XO) than my Hypex Nilai (which I use for the lows) at that frequency range and I suspect it is because of their much higher switch frequency at 1.1MHz. Orchardaudio claim a better performance of these relatively low powered amp than the higher powered version but it would be good to see some independent measurements also relative to Nilai or Eigentakt amps.
One issue of many GaN amp producers (not Orchard and also likely Laiv) is that they try to simulated the sound of older technology especially tubes instead of trying to make the most neutral and precise amp