A Journey into Setting Up an Audiophile Desktop System (Part LXIII)
All good things are 63 .. so be it.
What changed? Last time, i was experiencing some rush by realizing that the phase distortions are minimal when i use basically almost no shelving for the PSI A17 path.
This is nice, but adds more complexity below 80Hz to the system. I am still not 100% sure, but currently more into something that looks like this:
This filtering has the advantage that the mode at around 43Hz is greatly managed, just by removing signal from the monitor path by -14db shelving, which means no need for deep cuts there, which in turn helps the overall deep end clarity of the deep end / spatial support. And without much phase interference, that is the point.
The tradeoff here is to distribute the phase effects balanced from 43Hz over 80Hz and 103Hz (somewhat critical zones).
What i also did is to reinstantiate the Supra EFF-ISL cable for the PSI A17 path, which is an amazing cable that used to work only suboptimal before due to the residues in the system, which is brutally passes over to your ears.
The gain in detail, clarity and precision is amazing, again. No harshness, no brightness. It is always funny when people say a cable would be brightish — i assume usually this means the system is somewhat not very matured then, the brightness then rather comes from some noise in the time domain and the frequency domain.
Then the keystone in this round i would say is the Supra Excalibur USB cable, which i got from another audiophile–there is a very subtle but also absolutely essential improvement that comes with this cable. One would not believe it if one could not clearly hear it:
- The sound becomes more compact and tight, things fall naturally into their correct place on the time domain; some people like an ‘elastic’ sound, what this often means is somewhat of a stochastic drift in the time domain, which can also lead to a somewhat 2nd order space impression if that tells you something
In essence this means that the time transmission via the usb path is more accurate. This will obviously only be heard when your system displays this level of precision.
What also made a difference is using the NIF filtering happening inside the Supra SPC md06 plug bar in a different way than before: usually, also manifacturers are recommending to use the ‘unfiltered’ plug for current-heavy devices like a subwoofer or a large amp.
Get Christian’s stories in your inbox
Join Medium for free to get updates from this writer.
This is probably the most intuitive way to think about it. However, in reality, this only holds if you do not consider other consumers in the rail:
what happens after plug 1, the unfiltered one? one puts the PSI A17, in my case, which are filtered then and get their supply somewhat AFTER the sub. This results in interferences that compromise the acoustic integrity of sub and monitors. The ideal solution is to put the sub somewhat late, at 5 or better 6 (in my case), and the PSIs on 2 and 3 such that they get NIF filtering, but as few as possible.
All those changes result in a definite improvement which is called Reference State XXXVIII.
Minor aspects are using a proper usb c -> a adapter, since supra does not offer a usb c -> b excalibur usb cable per default. If you need to do this, there is a product from ddhifi called tc01, which is the best converter i found, and it apparently had been built exclusively for audio. Is there a big difference to something like lets say a schwaiger device? I do not know. But every detail counts.
What i also did now is removing the iPurifier3 from the dac input. It somehow is not needed right now since the timing seems much better wich the excalibur usb cable, hence i do not really feel the need any more for ‘reclocking’ or ‘rebalancing’.
It still remains an open, but rather experimental question to what extent an intona 7055-c would actually improve the signature. It is amongst all options the only device that is really of interest right now, in the light of the overall attempt to create the perfectly neutral desktop setup. Logically, the only residues remaining are potentially the ones flowing on the usb rails. For this, i would probably get another short excalibur since the intona is a smallish box one puts in the usb signal via usb b, and output the signal via usb a.
I just wonder if the rebuild of the usb signal would then cause similar ‘issues’ as with the iPurifier3, which are very smallish smoothing artifacts. The only real measure here becomes more and more the ‘acoustic realism’ in different aspects, ratherthan hearing real errors. How did i come up with the conclusion that the iPurifier3 is a wacko element in the chain given the current setup? It was the texture and decays of sibilants, consonant stability in general, very subtle things. I recommend hearing to voice recordings like eg. Autumn in New York on an album from Diane Krall. You always have the complication of recording errors and reproduction errors, which are NOT the same thing, but will result in similar acoustic artifacts or residues. But it is possible to differentiate between both by cross-checking different samples.
I will give you the topology such that you can rebuild it yourself:
It is worth noting that the Neotech cable is as important as everything else — it is the details, and all of them. Some things to note:
- The small little red elements are the Phonosophie EII fuses; i put them into basically all devices i could; this will cost a bit of money, but rule out some of the ‘possible bottlenecks’
- Using the room correction / ‘xo’ filters @192khz is probably more relevant than the conversion itself; but one thing does not work without the other so to say; there is a clear quality improvement, but also more calculation effort, which the ucx ii can handle
- Some ideas to test our / for improvement:
- Using Audes DBC-1 or Supra DC filter instead of the iFi dc+ device; it is really small and i would expect some more improvements
- Testing intona 7055-c as mentioned; it may or may not improve smoothness even more
- Using a 2kva transformer from thalheimer, or even a more elaborate device, maybe like audes, and see how it alters the signature; i still live under the impression that the ert 4L does nothing but galvanic separation following a laboratory norm (EN 61558–2–4), which is a nice thought in the sense of the overall ‘purity’ (in contrast to ‘audiophiling’ the signature; supra comes closest in terms of acoustic philosophy)
- I cannot use a different DAC since i somewhat rely on TotalMix and also would like to maintain that usage since i believe this is by far the best software for DACs that is out there, also from a UX perspective; it is lightyears ahead. This means that i could only use another rme device (for now); the complexity would be increased in stupid ways if i would have to use something like a miniDSP device to do the xo / filtering instead, not knowing the implementation of the algos etc.
So rme just relased a new version of the adi (june 2026), which may or may not be a candidate to drive it to the max; but the tradeoff is that i loose the flexiblity of TotalMix
- Obviously using even much more refined power cables would probably result in even more openness etc; but i am just not sure about the ROI here; if i had 20k laying around, i would probably be tempted to try it out
- On the long run, ps audio AC regeneration will probably blow your ears, but again, very expensive
