Gavin’s Digital Audio 2022 roundup – ATX power supplies

[Gavin1977]

 

No photos, just words for this one!

 

Products focused on in this article:
 

• Micro-Audio SMPS 440-TK3 36vdc and Micra-P5 5v linear (SR7)
• Nenon V3 (see this thread for more information on how to build one)
• Sean Jacobs Unregulated ARC6 Power Supply, and DC3/4.
• JCAT Optima S ATX
• Network Acoustic Muon USB cable

 

This month I managed to listen to all four of these music server power supplies.  Admittedly, the Sean Jacobs powered music server was not my own, I popped over to @ASRMichael's to listen to it and brought the Micro-Audio with me since I was already familiar with it.

 

You can see ASRMichaels system here.  What was immediately apparent is that the Chord Qutest, as good as it may be, does not cut it at all against the TAD D1000 DAC… a really lovely piece.  So we used Michaels system as is and just switched music servers for comparison.  Still, a good reference point for me to know what the TAD D1000 was contributing.

 

My music server consists of ROG MAXIMUS XI APEX, 8086k, APacer RAM, JCAT USB XE, JCAT Optimo Nano, Solarflare, HDPlex DC-DC, Ghent Audio cables were used.  It even has a passive Nvidia 1030 graphics card in it!  I’ve found that it really doesn’t make too much difference if I leave it in, and its much more convenient to be able to see what you’re going!

 

I use Windows LTSC, ASRMichael uses Euphony v4.

 

Now the first point I can make is that if you are only doing PCM playback, then there appears to be no easily discernible difference between running a 8086k vs 10900k, at least in an afternoons listening session.  Of course what a 10900k does allow you to do is PGGB or HQPlayer, DSD Upsampling and convolution filters with greater ease, so if that is what you like, then you should of course opt for the greater flexibility of the 10900k.  Our listening was based on PCM, local playback and also via Tidal.

 

Second observation is that it’s not easy to discern any great difference between Taiko ATX and HDPlex 800w DC-DC converters in this scenario – I suspect it’d take a longer listening session to do so.  But equally, this is good news, because the lack of significant difference at least shows that you don’t need to be overly concerned about missing out on audio nirvana if you can only afford the HDPlex 800w DC-DC.  I will certainly be in no rush to change. 

 

So, what differences were observed?

 

Well, firstly I should state that all of the power supplies listed here are a very noteworthy jump up over using a regular ‘high performance/low ripple’ ATX/SFX Power supply (e.g. Corsair AX1600i, EVGA T2, Corsair SFX750, HDPlex 400w AC-DC, and indeed HDPlex 500w ATX Linear).

 

What was also apparent to me (general observation for all four power supplies listed here), is that they all seem to extend the frequency response – for example, if the music allows for it, you will notice more bass/sub-bass and crystal clear highs that were either absent, or severely compromised with as bog standard SMPS.

 

Switching out my music server with the Micro-Audio and over to ASRMichaels Sean Jacobs powered music server showed that the Sean Jacobs has a slightly darker tilt to the presentation compared to the reference neutral Micro-audio.  The Micro-Audio is more ‘lit up’, in a traditional HiFi sense, and focuses on detail retrieval and technicalities.  Both have a really low noise floor, and the Sean Jacobs has the same detail retrieval, but presents it in a more natural way.  The Sean Jacobs had slightly better soundstage depth (the Micro-Audio is more forward) and I really enjoyed the smooth and textured mid-highs that it produced.

 

But in all reality was there a big difference?  Not really.  Perhaps the TAD D1000 is also more source agnostic than most… but I think not, and I can tell you this because I also witnessed the same changes in performance in my system when switching between the other power supplies, using just the humble Chord Qutest.

 

Both of our setups produced music of the highest quality, the differences between the power supplies are subtle, and you have to have a very revealing system to hear them.  What was way more apparent was the DAC used (£15,000 vs £1,350), USB Cable (Sablon 2020 vs Muon) and type of audiophile USB card used and the quality of power it receives.  It’s logic really, final leg is the most important.  My recommendation to anyone would be get one of these power supplies, but prioritise (or balance) your spend on the best DAC and audiophile USB card you can afford… and a Network Acoustics Muon.  The Muon can make as large a difference as any of these power supplies…  I never thought I’d say that about a USB cable, but it’s true and I bought it.  I’ve also ditched using any external reclockers as a result of owning the Muon – it just removes the need for them.

 

Other things I learnt from ASRMichaels system.  Firstly, he runs optical into his EtherRegen and then copper into his music server – sounded good.  I might try this because it means that the hot running Solarflare card can potentially be removed from my music server.  It seems that so long as there is some form of isolation (e.g. Optical) prior to your music server then all appears to be good.  My second finding was that I am now on the market for a new DAC!  Other than that our music systems, to me, sounded more alike than different.  Both very, very good.

 

So, thumbs up for both the Micro-Audio and Sean Jacobs.  One is DIY only, the other costs more and is built for you.  Your choice.

 

Back at home, now that we know that the Micro-Audio and Sean Jacobs are two slightly different flavours on the same performance tier what about the Nenon v3 and the JCAT Optimo S ATX?

 

The Optimo S ATX is a 27kg beast, fully linear with very nice power supply cables, each of which contain a dedicated grounding wire - nice.  Initially my music server wouldn’t boot – my fault, the rather substantial ATX power lead was pressing against my M2 SSD, which it didn’t like, and I’d also put the ground switch on the Optimo S ATX in the wrong position.  So, with some minor cable rerouting and a flick of the grounding switch it springed into life.  First impression – complete darkness and absolutely zero grain.  Upper registers are clearer and more detailed compared to the Micro-Audio.  It’s also not as forward as the Micro-Audio and has better layering and separation.  All classic traits of a linear regulated supply.  It is also a less fatiguing listen than the Micro-Audio, because it's less forward, however you shouldn’t take this as the Micro-Audio being some how inferior, other than the somewhat better layering/separation and the clearer/slightly more detailed upper registers there is very little difference.  The JCAT is better suited to a system with classical ‘solid state’ traits, that might lack a bit of dimensionality or the flow of the best tube gear.  The Micro-audio on the other hand will provide presence (read slightly forward) and clarity to a system which is perhaps a bit too smooth or warm (e.g. classic tube sound).  The Micro-Audio is fantastic with intonation – best of the bunch.  The JCAT doesn’t put a foot wrong – everything just sounds right and it is the more polished of the two.  So just pick the one which your budget, skills and system synergy dictates.  Without a back-to-back comparison such as this I very much doubt if anyone would be able to recall noteworthy differences between the two units – they’d simply be very happy and impressed with the improvement compared to using a bog standard ATX power supply.

 

So, final comparison.  Nenon V3.  Now this power supply can sound very different because it is very sensitive to power supply quality – you must give it a dedicated power outlet as it is current hungry.  Change your power lead or fuse and you will hear the difference.  This is not a power supply for someone who is not a true Hi-Fi enthusiast, you need to have an optimised setup to get the best from it.  The Micro-Audio is the best in this regards, it is not very sensitive to upstream power – it gives you clarity and impact with pretty much any system.  The Nenon V3 (and I guess the Sean Jacobs ARC6 unregulated by logical deduction) will sound soft and lifeless if starved of current.  I know it sounds wrong given the amount of capacitance in this thing, but it’s true.  But sort out your power and it has the same life and energy as the Micro-Audio.  What the Nenon V3 does is provide a very natural sound.  It has a more natural presentation, accurate decay and smooth top end.  Zero fatigue.  Easy to pick up on artificial reverb used in recordings.  The JCAT and Micro-Audio are more traditionally HiFi sounding, Nenon V3 is for connoisseurs of fine tubes and artisan power leads and natural flow where the details are readily apparent but don’t pop in your face with the wow factor of a HiFi demo room where the dealer wants to impress make a sale.  The Nenon V3 almost has a slight tubey sound to it in the mid-bass, perhaps down to the use of large caps as opposed to lots of smaller ones, but it’s not quite as obvious as the Sean Jacobs, which is closer in this regard.  Who knows, I’d need more time to evaluate this difference.  One thing that was clear however was that the Nenon V3 has a slightly more hallucinogenic soundstage compared to the others – I believed this in in part to the quality of components used and the ultra-low impedance hardwiring… no PCB’s here.  Hard wired audio components just sound more dimensional and open to me.

 

Conclusion

 

All of the power supplies worked flawlessly for me in this review and they are all a substantial uplift over a traditional ATX PC power supplies.

 

Which one am I keeping – the Nenon V3.  The differences between all of the units are minor compared to the quality of USB output and DAC that you use, but the Nenon V3 has a more natural presentation.  I also love DIY and am happy to engineer things, so this gives me the opportunity for further experiments (Taiko Rectifier on order).

 

Overall ‘most’ people would be better to choose the cheapest power supply of the bunch, the Micro-Audio, and spend your money on a Network Acoustics Muon and, the best Audiophile USB card, associated linear power supply and DAC that you can place your hands on… the last part of the chain has more impact than the differences between these four ATX power supplies.  Use a HDPlex DC-DC with the Micro-Audio and save a bunch of money compared to the Taiko ATX – the performance delta is really not that apparent.

For people not comfortable with DIY, but slightly deeper pockets and an already excellent system with good power, I think the Sean Jacobs ARC6 unregulated is the way to go.  But if you want to venture into DIY and have the funds, then go the whole hog and get the Taiko DIY kit and build yourself and unregulated power supply – I don’t think you can go far wrong.

 

So, as you can see, with regards to ATX power supplies, what I have learnt is that its more about matching the supply to your budget and use scenario.  I’d genuinely be happy with any of them.   Everyone likes a winner - but I can’t give you one here.  My advice to you, is not to worry too much about which audiophile ATX power supply you have, so long as you have one of these then you’re good to go.  Focus on other areas of your system.  Have a balanced approach to putting your system together and consider synergy.

 

Shameless plug:  I can’t keep all of these power supplies, the Micro-Audio is up for grabs in the classified section.
 

[Gavin1977]

Apologies for the typos, have asked @The Computer Audiophile to help me amend.

[The Computer Audiophile]

Unfortunately there’s no easy way to edit typos after the 30 minute post-publishing period is over. People have to let me know what they want changed and I have to change it. 

[Exocer]

Thanks for the write-up Gavin!

 

Very excited to hear your feedback on going from the Saligny standard to the Taiko rectifier. I just made this change myself. I'll let you experience it or yourself :).

 

Using the Paul Hynes SR7T at one point, switching from the HDPlex 800w DC-ATX to Taiko DC-ATX was transformative for me. This was only feeding ATX initially that this difference was noticeable, CPU still directly fed by the SR7T 12v 12a rated rail. As a sanity check I did switch back HDPlex before eventually selling it and the delta between the two had grown, feeding CPU and ATX both from the Taiko DC-ATX. That being said, the HDPlex was certainly a reliable, good sounding and cost effective solution. So I do agree that for some, money may be better spent elsewhere but for the point I was at in this journey, the Taiko (with the SR7T) was a worthwhile step in the direct I wanted to go. So you're absolutely right about prioritization and focus. I will admit that my initial dac upgrade was slightly underwhelming because the other aspects of my system were far behind. It was only after some time with optimization and switching back did I realize the delta between the two. An eye opening experience!

 

I encourage you to compare the HDPLEX to the Taiko DC-ATX while powering your server with the Nenon ULPS V3, all else equal (not comparing the two in different servers with man different variables at play).

 

Thanks for the write-up. I really enjoyed it! Glad you'll be keeping V3! Excited for your journey. 😎

[Gavin1977]

Thanks for the feedback Rob - I will try the Taiko ATX in my system at some point.

DAC is now definitely a bottleneck in my system, so once I unblock that then I might hear benefits elsewhere.  What’s your DAC?

Good point on having a dedicated rail for EPS, many have already attested to the benefits of this, might well be worth it… a configuration worth trying.

[Exocer]

6 hours ago, Gavin1977 said:

What’s your DAC?

Yggy LIM. You’ll probably like it better than the A2 you tested. Noticeably improved noise floor. Even with A2 (I swapped boards to LIM when they were released) many of these experiments were worthwhile. 

As for a dedicated rail, that sounded better to me when I first received the Taiko DC-ATX but I ended up using 1 rail (no dedicated cpu rail) in the end since that sounded best. The gains in clarity, detail retrieval and transparency were immediately noticeable compared to the hdplex, using the Taiko only for the ATX section. I felt it to be a worthwhile addition and a very impactful purchase, even with Yggy A2 and a tube amp (many others here have spoken about this noticeable improvement over the hdplex). A month or so into owning the Taiko dc-ATX I tried feeding both cpu and ATX with 1 rail (19v sr7t) and the initial thinness I had experienced was gone. Now, I use 1 rail from the ULPS to feed ATX and EPS since this ended up sounding best in the end.

 

As for pure PCM playback, I revisited using my original i7 8700T processor earlier this year due to an unfortunate mishap, and it most definitely sounded thinner, smaller and less dynamic than the i9 9900k I normally use although it sounded pretty good overall. This was with Euphony V3.

 

Excited for you to try the Taiko in your system while keeping all variables the same. 

[Gavin1977]

Indeed - I think you’re right, Schitt LiM is the only DAC I’m really interested in at the moment.

[ASRMichael]

20 hours ago, Gavin1977 said:

Indeed - I think you’re right, Schitt LiM is the only DAC I’m really interested in at the moment.

Gavin I can send you the Taiko ATX to compare if you like? In September mind you. 

[Gavin1977]

8 minutes ago, ASRMichael said:

Gavin I can send you the Taiko ATX to compare if you like? In September mind you. 

Cool - no rush as I’m on leave anyway.  Thanks