New GaN Transistor Class D Amplifiers now Shipping

Discussion in 'Audio Hardware' started by avanti1960, Aug 28, 2021.

  1. avanti1960

    avanti1960 Forum Resident Thread Starter

    Location:
    Chicago metro, USA
    Two new Class D Audio Amplifiers that use GaN (Gallium Nitride FET) transistors are now shipping to audio fans.
    The technology has been discussed on line for some time and has generated high anticipation and expectations.
    The GaN transistors allow for faster switching and low levels of negative feedback to reduce distortion to state of the art levels.
    A few user reviews on Audiogon forums are quite positive to say the least.
    The mainstream Peachtree audio version has 400 watts / channel and lists for $2999.
    The vendor direct Underwood HiFi LSA "Voyager" has 350 WPC and is $3000 US.
    These things are shipping without the ink being dry on their respective manufacturer's web sites.

    Anyone have a chance to hear these?

    Interesting times.

    [​IMG]
     
    hifisoup, Manimal and toasty like this.
  2. mkane

    mkane Strictly Analog

    Location:
    Auburn CA
    Honestly. We have 5 Class d amp here and none cost anywhere near three grand. Our most powerful spits our 275w @ 8ohm. And were using one as I type because it very hot here and these don't put out much heat.

    Thats the joy of DIY. Kinda dry compared to everything else here but, very detailed. The average consumer would be pleased. They're very silent in the quiet passages.
     
    Last edited: Aug 28, 2021
  3. toasty

    toasty Senior Member

    Location:
    Tiverton, RI USA
    Very interested. Haven't heard (don't get around much anymore) but reviewers are doing backflips about the upscale AGD amps.

    If that level of performance, without the bling, is available for less coin, well, curiouser and curiouser.

    Though I've been a Tube Guy for most of my audiofiend existence, I once owned NuForce Ref 9 SE mkII amps and still think highly of them.

    Even Ralph of Atma-Sphere expects to bring a GaN amp to market.
     
    unclefred likes this.
  4. Gibsonian

    Gibsonian Forum Resident

    Location:
    Iowa, USA
    Love to hear one.
     
    Monty12 likes this.
  5. Helom

    Helom Forum member

    Location:
    U.S.
    Dynamic range spec is a little disappointing but benefit of the power output could make up for that with many speakers. Definitely an intriguing product. Hopefully we’ll see some independent bench measurements posted somewhere soon.
     
  6. Mike70

    Mike70 Forum Resident

    You're talking about the Peachtree?
    I see at least 112dB ... CDs can have a theorical 96dB maximum, and you depends more on recording mastering and speakers / acoustics dynamic range. In real life, people doesn't have much more than 70 / 80 dB.

    It's my opinion, of course, but an amplifier with 110dB can be totally enough. If we compare only with the theorical measurements, the nova 500 "it's better", but that doesn't mean a better amplifier in real life for someone, because maybe we're comparing inaudible stuff.

    Think about vinyl ... i listened to pretty good systems compared with digital sources and let me say that the difference it's much more in the recording that in theorical numbers. Maybe in top high end systems you can have another experience, i don't know, but generally theory and practice are equals only in theory.
     
  7. Helom

    Helom Forum member

    Location:
    U.S.
    Yes, was referring to the Peachtree. All its specs are a fair order of magnitude lower than SOTA with exception of its power output.

    I don’t know where you got those numbers but the dynamic range limitation of human hearing is approximately 115db, some sources claim even higher.

    My anecdotal experience is that these seemingly minute differences in performance specs matter, even at SOTA levels. Over the past couple weeks I’ve had three preamps in house for audition: the Quad Artera, Bel Canto Pre5 and Benchmark LA4. Each is a fully-balanced design, and respectively, represent progressively higher performance specs. The Bel Canto was a clear and obvious improvement in resolution over the Artera, as was the LA4 over the Canto. It was even apparent with some of the most highly compressed recordings, and with speakers that are probably 15 to 20% less resolving than the best I’ve heard.

    It’s my experience and belief that distortion and noise are almost always accumulative. Listening enjoyment through my reference system has improved with each progressive step in minimizing these factors. This doesn’t mean I can’t or don’t enjoy gear of lower resolution though. I still own and enjoy tube and phono gear.

    Another issue worth noting is that manufacturers commonly fudge these numbers to appear favorable on the spec sheet. I wouldn’t take Peachtree’s claims at face value. Its power output is attractive though, so I’d still consider trying one if the specs are as good as claimed, however, I’d like to see independent measurement results before taking the plunge.
     
  8. mkane

    mkane Strictly Analog

    Location:
    Auburn CA
    Class D play well with some types of music. Bog Band, Classical, probably Electronica. When it comes to voices I'll use tubes.
     
    Melody50 and bluemooze like this.
  9. avanti1960

    avanti1960 Forum Resident Thread Starter

    Location:
    Chicago metro, USA
    Little blurb on the technical aspects of the GaN FET:

    "GaNFETs turned-on ~4 times faster and turned-off ~2 times faster reducing switching distortion (ringing) by several orders of magnitude. This, in-turn, allowed designers to use much less global negative feedback to achieve excellent measured results. The result is an amplifier with exceptional musicality and excellent measured performance."
     
    head_unit likes this.
  10. rcspkramp

    rcspkramp Forum Resident

    Location:
    Vancouver, WA USA
    Curious about how this sounds. I've auditioned several class D amps and they never seem to push my buttons.
     
    SonicCzar, timind and marcb like this.
  11. Harris11235

    Harris11235 Forum Resident

    Location:
    Minneapolis, MN
    I’m pretty sure I saw the Peachtree amp on Darko’s rack, so we’ll probably be getting his take on it soon.
     
    avanti1960 likes this.
  12. head_unit

    head_unit Senior Member

    Location:
    Los Angeles CA USA
    Yeah...but it's really the residual noise that matters. There used to be, and surely still are, like 400W/channel amps specifying 120 dB dynamic range, however compared to 1W that is -94 dBW. Not better than a 20W amp of 107 dB dynamic range (if I'm doing the maths right).
    When I used to do trainings I said it was like stacking panes of glass, each one could be fairly clear but stack several and the view becomes less so. Plus of course each stage amplifies the noise of the stages before (well maybe not the preamp depending on design. I fondly recall my little NAD of yore with its volume control range switch.
    - All this is why our for our next move from the Denon AVR-X3600H we want to find external power amps with like at least 10 dB less residual noise.* How to check that I don't know because I've never seen anybody measure that. Denon AVR-X3600H AV Receiver Review measures SINAD but no residual hiss noise measurement that I see.

    *and ideally 3-channel, and at minimum 300W/channel at 4 ohms. Any nominees?
     
    jusbe likes this.
  13. head_unit

    head_unit Senior Member

    Location:
    Los Angeles CA USA
    Tell us more, tell us more!
     
  14. Mike70

    Mike70 Forum Resident

    I don't understand why you think i'm talking about human ear ... i'm talking about "real life" dynamic range ... the dynamic range that you will hear from your recording and speakers / acoustics / noise floor. I repeat ... i don't talk about theorical numbers on whatever (ear / amplifier / sources / ...). Theorical dynamic range is one thing, real dynamic range in your room have other numbers.

    better SQ it's not about any isolated parameter.
     
  15. MaxBuck

    MaxBuck Forum Resident

    Location:
    La Quinta, CA
    orchardaudio, head_unit and EnochWWG like this.
  16. Helom

    Helom Forum member

    Location:
    U.S.
    Well residual hiss will be slightly different for all systems depending on connection quality, cable routing and sources of interference such as a transformer anywhere within a couple feet of low-level circuitry. As for self-noise, there exist some components that have none, for all practical purposes.
     
    head_unit likes this.
  17. Ntotrar

    Ntotrar Forum Resident

    Location:
    Tri-Cities TN
    Yet to see how they will fare but I went from tubes to Hypex Ncore and I haven't looked back. Technologies keep emerging...
     
    adaug, jusbe, HiFi Guy and 4 others like this.
  18. Helom

    Helom Forum member

    Location:
    U.S.
    IDK about your room but mine is extremely quiet, quiet enough to hear a mouse fart, so these seemingly minute differences are quite audible.
     
    Ntotrar likes this.
  19. Ingenieur

    Ingenieur Just a dog looking for a home...

    Location:
    Back in PA
    Dynamic range is defined differently for Class AB and D, not comparable.

    AB
    10 log((S + N ) / N) dB
    (S + N)/N = S/N + 1
    S/N is in the x00,000,000 range so the +1 can be ignored
    So dynamic range is ~ S/N

    Class D
    Largest discrete power package/smallest
    This is determined by switching time since output is PWM (constant magnitude and varying duration, the product is power magnitude when integrated by an inductive load/filters). Basically the largest signal magnitude/smallest... (NOT noise).
    Noise is moot with a switching device.

    These should not be confused with audible dynamic range.
    If background level is 40 dBZ (pretty quiet) and music peaks are 110 dBZ (loud), dynamic range is 70 dB, about the best you can get in a home.
     
    Last edited: Aug 29, 2021
    SteveKr, head_unit and Mike70 like this.
  20. Ingenieur

    Ingenieur Just a dog looking for a home...

    Location:
    Back in PA
    Here's a Class D example, numbers make it clearer sometimes.

    amp switching rate 10000/sec
    Assume pulse is 1 V and takes 1 sec of switches or 10000 to get 1 V.
    0 = no signal
    1/10000 = smallest signal or 0.0001 V
    10000/10000 = largest signal

    20 log ((10000/10000) / (1/10000) ) = 80 dB dynamic range
     
    guestuser and head_unit like this.
  21. Mike70

    Mike70 Forum Resident

    That's what i try to show ... maybe with math is more clear :edthumbs:

    Then, you have the dynamic range existent in the recording (you can't improve that) ... and the dynamic range that your speakers can show (you can't improve that) ... and all the fantastic theorized number bytes the dust.
    It's crystal clear in my opinion, even without the math.
     
    Hanks3 likes this.
  22. Helom

    Helom Forum member

    Location:
    U.S.
    Hear a Benchmark stack in a quiet room and you’ll be eating crow.
     
    head_unit likes this.
  23. Ingenieur

    Ingenieur Just a dog looking for a home...

    Location:
    Back in PA
    What will be heard?
    The limitation is the room and the source

    For a given background level and max SPL it will be the same? You will not hear a difference between 100 or 110 SN or dynamic range.

    My amp, DAC input, no signal, 100% volume
    0.1 mV at the speakers.
    That can't be heard
    Nor can you listen at 100% volume 140 W...for long
     
    head_unit likes this.
  24. Ntotrar

    Ntotrar Forum Resident

    Location:
    Tri-Cities TN
    I have a quiet room, you won't hear a peep out of my Marantz 30. Top brands wouldn't be putting "class D" in their premium products if it didn't meet their standards.
     
    jusbe, HiFi Guy, Bill Mac and 3 others like this.
  25. Mike70

    Mike70 Forum Resident

    Surely can be a very good amplifier ... but it's about the dynamic range capability?

    If you like measurements, check the dynamic range in your recording ... with a digital recording it's pretty easy.
     
    head_unit likes this.

Share This Page

molar-endocrine