New Generation Class D - AGD Audion Mkiii upgrade

Must admit too that I really like the looks of the tube thingy and was ready to pull the trigger.  I think though I’m going for the Atma-Sphere Class D instead.  I’ve always liked the sound of their amps, and the Class D will probably be a better match with my Atma-Sphere preamp.

@rocketiii - Put my MkIII's in this afternoon.  Listening to them as I respond.  Everything you said is true.  More detail in the "upper" midrange, cleaner / more separation and a little bigger sound stage.  I think the sound stage got deeper.  I'll keep you posted only have a hour on them.  

On a side note I have all three generations and at some point plan to compare.  Just not sure when I'll have.  Got to love the Pelican shipping boxes for the tubes.

@testrun I thought you’d be thrilled with the sound of the Mkiii tubes! I also loved that the tubes came in their own Pelican boxes. 
 

It would be fascinating to get your thoughts on the differences between each generation but enjoy your new tubes for the time being.

Hey @testrun how are you finding the Mkiii Audions now that you have more than an hour on them?

Incidentally, I’ve just removed them from my regenerator and plugged them straight into the wall and it’s improved the dynamics even more! I’m quite surprised at this because my speakers are 107dB/W so drawing very little current.

I should be receiving my MK III's on Monday.and am looking foreword to hearing them I talked with Alberto and he recommended a plugging them into a conditioner. I will try directly into the wall for an experiment. They do sound pretty sweet into the P.C., though.  

@giantsalami that is very interesting and I have been doing some testing and found my regenerator has stopped regenerating! I’d previously been running Atmasphere OTL and that definitely sounded better with the regenerator.

I’ll report back when my regenerator issue is sorted. Anyway no harm in trying wall vs conditioner.

I have a pair of the newly updated AGD Audions moving between my systems right now.  They are spectacular, full-stop.  I also have (currently) a Technics SU-R1000, a McGary SA1E, and a LTA Z40i to compare them to.

You can chalk up things to whatever gimmick you want, but I'm pretty sure that the GaN FETs that AGD is using are the only ones actually designed from the ground up for audio.  And the ability to simply upgrade the amps by updating the contents that go inside the KT88 shell is brilliant, you don't need to send the amps back for updates.

I'd love to directly compare the Atma-Sphere amps to the AGD, but I'd be surprised if the Atma-Sphere offered something that AGD did not.  I've always admired Atma-Spere amplifiers but in the interest of full disclosure, I owned OTL amps for years and was a Joule Electa guy (VZN-160s).

I'm pretty sure that the GaN FETs that AGD is using are the only ones actually designed from the ground up for audio. 

@badgerdms GaNFETs are designed as switching devices. To that end, they really can't be designed for audio other than to be switching power devices that happen to get used in a class D application.

@tonnesen 

Any update on your experience? Were you also able to try the Atmasphere class D?

I have the MK III for about 2 weeks now and am just impressed at the clarity and sound they produce. Immersive is a good description. I am thoroughly impressed and satisfied with these. 

@atmasphere I understand but before International Rectifier got bought by Infineon, they were working on an audio solution.  When that project got scrapped I believe Alberto bought all of the product produced to that point and began the process of designing and building his amplifiers from that stock--which had been produced with audio as the end game,

@atmasphere 

Sorry, but what you are stating may be applicable to the rest of the GaN MOSFET but that is not correct for the device we use in the Audion GaNTube88KT and I will be more than happy to describe to you why.

Without boring anyone with complex terms, in short, the reason is simple.

Since you know very well the difference between fixed D% (i.e Buck or Boost converter) and variable D% (i.e CLASS-D) operations, symmetrical HB and asymmetrical HB configurations, and among other things how important the stray inductance of the switching pair at the switching node is (we switch our power stages at ~800KHz but we could go much higher...), you should easily derive which parameters and physical elements of a semiconductor device (the 3 terminals switch) need to be optimized to make it operate at best in each different topology with the ultimate goal to obtain not just the best efficiency....but in our case, a great sound...and these parameters are not the same. I assure you ;)

well, that sure clarifies a few things...🤔😁

@agdproduction Your description simply sounds like a good switching device (i.e. GaNFET), combined with good engineering arranged to make an amplifier. I'm very aware of the difference between efficiency and good sound :)

@atmasphere 

Well, what can I say, I am very very glad you keep misunderstanding the explanation and the physics of the semiconductors applied to power electronic technology and applications :)

@agdproduction

Does the ability to build a GaNFET class D of our own design from the ground up, have it meet EU Directives and get nice comments about it and such, that I somehow I am

misunderstanding the explanation and the physics of the semiconductors applied to power electronic technology and applications

seem true? Not sure where this is coming from- its not something I'd say about you in the face of similar evidence. So I have to conclude you misunderstood my post.

@Atmasphere

1++++++++++

@atmasphere 

I don't think I misunderstood your post, you replied to our true statement about the fact we use a specific GaNFet HB device that was specifically designed and optimized for CLASS-D (Audio) applications as not true. That was already borderline offensive since your comment implied I was making up something or worse I was a liar. Nevertheless, I very gently tried to explain the reasons and the basis of my statement, and still, you ignored them and made a further simplification inherently calling me a liar or incompetent, by saying that all GaNFet as long as are "good switching devices" are good for Class-D anyway. They are not, and you can repeat that all the time you want, it will not make it right. In any case, I found this type of argumentation a total waste of time, if you wish to get a more comprehensive explanation just give us a call and I will be glad to re-explain my statement again, for the last time. 

@agdproduction The reason I responded the way I did is because your 'explanation' lacked anything that pointed to an audio application. So perhaps you could be a bit more specific. For example, having a lower gate capacitance as advantageous, but that would be a benefit to any switching device. So in order to support your position, you would need to be very specific about how the GaNFETs you use are somehow different from regular GaNFETs in a way that favors audio without favoring switching devices in general.

I had thought that was clear from my prior posts; I apologize for my lack of specificity in that regard.

Dear Ralph,

my position is supported by the quality of the sound of what we make (and what I designed from the ground-up...literally), and also those few decades spent in a much less fun environment than Hi-end audio...meaning the semiconductor sector.

Now, and with all due respect, the simple fact your last reply cites the "input capacitance", is an obvious demonstration of the gap in the fundamental know-how of semiconductors in general, and for sure in III-V material devices that exists on this matter from your side. For your benefit, you know, there are multiple FOM (figure of merit) indicators in every device that are application-specific and need to be optimized, together with the package and internal device layout, to reach the optimum performances (you can ask your friends at GaN System), and in the semiconductor industry, there are millions of dollars spent every year to gain small % edge vs. competition every year and with a new generation of parts to improve just a few of them. So it is fruitless to take this venue and continue an argument over a true statement, or transform it into a "Power Electronic 101 for dummies" type of training. Also, on the same token, and maybe I need to ask forgiveness for this ;)...well...I need to confess... I am definitely and categorically not interested in telling the "secrets" about how to make a better GaN for audio to my good competitor!  Right ?? ;)

So really, in all honestly that was not my goal, I simply wanted to re-iterate what I have said, written, and publically declared, sorry if you keep ignoring the evidence in a sort of "don't look up" style.

The results of the AGD products are the proofs and for the rest, well let's leave it as is, and let's enjoy the music again, in whichever form is reproduced :)

 Thank you and have a nice Thanksgiving!

All the best!

Ouch!  And then "Happy Thanksgiving"? I think that's what's called a sh*t sandwich.

@atmasphere 

I have never read a rude statement from you on this forum, including on this thread.

@roxy54 

1+++++

The LSA Voyager GAN with mods is one helluva amp. It’s error is by omission not commission. Beautiful mids, extended crystalline non-fatiguing highs. What it gives up to the Class A and AB behemoths to my ears is meat on the bones from the midbass on down. That’s about it.

Although at some point I intend to send Ric my LSA Voyager (I am one of the lucky ones to have his dual mono EVS 1200), my stock Voyager is very livable

As someone who owned the first modded LSA Voyager and now the PeachTree GAN1 (also modded). My preference is the GAN1. I finally got the GAN1 back playing tunes (needed another streamer). It is such a good amp. I only use it now for my RAAL CA-1a headphones and you hear what an amp can do with the RAALs.

Only negative I had with the GAN1 was that it was not powerful enough for my Yamaha NS5000 (use a CODA #16 with it) and it was also not powerful enough for the Magnepan LRS+ (use a Sanders Magtech amp with it).

The modded GAN1 sounds as good (maybe better in some ways) than those 2 amps.

Not selling any of those 3 amps nor can I since I tossed the shipping boxes. 

@abd1 

I've had the AGD Duet's for about a month now - these are AGD's new monoblocks, 300 wpc at 4 ohms .  I'm using them with YG Hailey 2.2 speakers and Playback Designs MPS-8 DAC and MPX-X streamer (using the MPS-8 variable control with no preamp).   I'm very happy with the sound, the AGDs easily drive the speakers and work very well in this system.  I think they are a very good match with the YG speakers. 

 GaN FET’s don’t need 100 of hours of break in. I proved this at friends home a while ago.

@rajugsw You’re funny.......Any true audiophile knows that electronics Always improve after an initial burn in.....gear, cables....even your 4 K TV....you my friend must have a Sansui integrated with Bose 901’s.

Now, and with all due respect, the simple fact your last reply cites the "input capacitance", is an obvious demonstration of the gap in the fundamental know-how of semiconductors in general, and for sure in III-V material devices that exists on this matter from your side. For your benefit, you know, there are multiple FOM (figure of merit) indicators in every device that are application-specific and need to be optimized, together with the package and internal device layout, to reach the optimum performances (you can ask your friends at GaN System), and in the semiconductor industry, there are millions of dollars spent every year to gain small % edge vs. competition every year and with a new generation of parts to improve just a few of them.

Yeah, I have friends and customers that design semiconductors...

Boiled down, a US phrase is 'the proof is in the pudding'.

@atmasphere It sure is.  That's why I ordered the AGD Audion Mkiii's 

 Both AGD and Atmasphere can send me their amps for auditioning. 
I can provide an unbiased opinion based on what I hear. 
 

+1 @badgerdms   Always buy that which can be upgraded instead of the "one and done" inferior design that doesnt keep up with the newest technology and innovation possible.  Better to be a Master of 1 design than a mediocre "jack of all trades".

Always buy that which can be upgraded instead of the "one and done" inferior design that doesnt keep up with the newest technology and innovation possible. 

Yes, its always a good idea to make an amp easy to service and update!

I believe BOTH amps are easily upgradable.

Actually, I know this for a fact.

@dseltz I'm sure they are.  Ralph has certainly been successful for a long time and probably has a great plan in place.  I just love the idea of not having to send the amps back for updates and having the updates essentially delivered to you.

Good point, sir.

I am a fan of the AS Class D (I own a set)

I have not heard the AGD so can not comment on the sound

I believe BOTH companies designed a GanFET Module from the ground up for their own purposes. Building an amp with the sound characteristics that they felt was best.

Although an easy, at home upgrade is awesome, so is an upgrade that includes new Caps, resistors, and other parts I am not comfortable doing myself. Again, probably an individual preference so we should not be insulting one another over which one they prefer.

@arafiq - I, too, would be willing to do an at home audition between the two (if someone wants to let me audition their AGD lol

I can't afford Ralph's amps, much as I'd love to get one, or the AGD amps, but I have my eye on the March Audio amps. Owners seem to love them. 

@roxy54 

You can score used LSA Voyager 350 GaN amp for ~ $1800, have it shipped to Ric Schultz (Tweak Audio) for mods and be beyond happy; You might even find a modded one ~$2500

HTH

@tweak1

Is there a reason why you think the Voyager Gan amp is superior to the March purifi?

OP

you would need to research it

I love the smell of napalm in the morning and especially so when the battles are likely to benefit us all in the long run.

@tweak1 

Were you addressing me? I'm not the OP.

Hi, I already have my MK3 upgrade from my Audion. I'm in the burn-in phase, it will take about 25 hours and I enjoy the sound of my Audions. I wouldn't know how to discern now between the sound of my Mk2s and this new Mk3. The memory of a fish has those things, and I think it is essential to wait for the burn time that Alberto recommends of 150 hours to make a serious comparison about the sound between both tubes. Beyond that, I'm really liking the sound of the Mk3, but I also really liked the Mk2. At this point, my question is which preamplifier would be recommended to use in my system. I try to explain the complexities of it: I have analog sources that require the preamplifier to have a line output and RCA and XLR inputs. Two digital sources that require XLR inputs. So it is necessary for the preamplifier to have various inputs and outputs

I am running a Primaluna 300 evo pre, which sounds fantastic. You would have to get the 400 for the balanced outputs.

@mbmi I have 2 pairs Bose 901 Series 1 & 2’s stacked for my Surround Channels powered (@4Ω) by a 3 Channel custom GaN FET made by me Orchard Audio Ultra (500WPC) Amplifier.

BTW....Yes GaN chips are used in power supplies as I humorously mentioned in my Video.

. Alberto recently got into a battle of words with Ralph of (I stayed out of it - I’m on Ralph’s Karsten’s side here). I’ve heard Ralph’s GaN Amps on a pair of PS Audio FR20’s and they sounded fantastic. I never got to swap in Leo’s Orchard Ultra Amp cause the dude sold the FR20’s !! At 40% of what he paid for them !!!

I think Jeff B still has Ralph’s Amps. Not sure. If anyone can make a GaN FET Amp sound like a tube Amp. Ralph is the Man !! But Leo’s Amps sound amazing with a Tube Preamp in front of it (just ask Zeos).

Giantsalami 

Thank you very much for the recommendation. I have not been lucky enough to hear Primaluna, I know that it has good reviews and the price is reasonable (just as the prices of HIFI equipment have been set). In addition to the Primaluna, have you paired the Audion with any other preamps?

After several weeks of use with the Audion updated to MK3, I am very happy with the sound, it has been a worthwhile update. Congratulations to Alberto for allowing this equipment update system. It is a pity that in Spain there is no representative from whom you can request proof of the previous Alto. Now my audio distributor has let me try a solid preamp from the German brand Trynity and it gives a good sound, but I think that before making any decision to purchase the definitive preamp for my system, try a good tube preamp. I think Jay's uses Allnic with his Audions.

@rajugsw Jay is correct.  The GaN FETS that Alberto is using were initially designed, they were designed with audio in mind.  Not claiming each company that uses the tech doesn't have all kinds of additional proprietary implementation secrets, but they are nearly all using chips designed for other things.  Whether you think that makes any difference is up to you, but the AGD amps are incredible devices,  They combine amazing design, fabulous sound, and easy upgradability in a way that no one else touches right now.

@nadimjaber , I have not tried them with any other pre, but will soon, as I recently purchased ( happy birthday to me!) an Ayon CD-35 that is also a pre amp. I am currently listening to it through the PL and will be trying out the pre amp side side soon. 

I had both the Audion II’s and the Atmasphere Class D’s at the same time and preferred the Audions. They are fast, neutral, delicate, well balanced and have a very black background.  I found the Atmas to have a bit of graininess in the upper mids.  They were a bit weightier.  There was a slight transformer hum, which might be abnormal, but that put me off a bit.