In my pre NOS1/XXHE days I played with a linear psu and it did make a difference on the PC. Is it your position that it could make a difference? Or is it your position that if it does make a difference that something else is wrong?
Anthony,
Certainly not the latter (in my view).
I have been asked about the LPSU before and my answer was (and is) : I don't know. It is too complicated and too much is involved;
When selecting the SMPS psu for my XXHE PC I followed your recommendation for the BeQuiet L7 530w because you had measured it somehow and determined that it does not spew too much hash back into the mains (I read that on this forum somewhere). If we have a SMPS that works like that, where do you think the gains of linear supplies for a PC may lie?
Or is that a can of worms you would prefer not to open?
All we can do (indeed) is looking at it spitting back into the mains and next think how that by itself will influence the other devices. It may, it may not and that depends on those devices. Prevent the spitting back into the mains *is* something to better avoid of course. But whether really necessary ?
The can of worms emerges merely in the area Coen talked about. So, the PC will contain spades of voltage regulators and they all will be the most noisy. Now however, it is a matter of how dangerous that will be, thus now looking at the "output" of a PC, like USB is such an output (towards our further gear I mean).
Next thing though is that it is obvious to me that whatever noise is coming through there, it will be overwhelmed with what the DAC side creates for the very same connection/interface to operate. Thus, as long as that is in order, why to work on the better noise-free PC ? (but now in the voltage regulation department). And again next, when that is arranged for to some degree (like the NOS1 does this fairly well), NOW again the PC noise is in order. Sadly, and through my eyes, this should be so much less noise that at least I can not measure it. Well, not in-band, and it is not about in-band at all because in the end this is about jitter influence.
And then to think we can already influence sound by PC-processor (etc.) influence ? This should be way under the level of noise the regulators produce.
All remains relative, because when a NOS1's output noise of 8uV goes down to 0.5uV (theoretically possible) we might start to see PC noise (explicitly influenced or not) at the outputs. May, because that 0.5uV will be about rejection largely, and this all may happen only after the interface, that interface thus still implying jitter because of its noise there.
All in other words : All is one big pile of influencing matters, and it can be expected that when one is solved the other emerges because it becomes profound and which might not even be for the better (different jitter spectrum).
And in the mean time the by far most of us struggle with analogue noise at the output side which now from that angle kills the sound.
We could also say that we all go a bit far with this all. It is, however, fairly important to see through what we should work on FIRST. I personally don't think this is any LPSU for a PC. Some general rules to avoid that analogue noise would be far better to work on (first). Sadly, I do not see how these rules are or should be. Make it inaudible by whatever means (all trial and error) at least allows us to have a workable target. But it is only the first step.
Working on "complete systems" which at least inherently won't show any of this noise is one means (like actually Bert and me propose this today) but this is a bit of a rough way.
Can talk on, but it's all of not much use unless someone likes to know my general opinion. So, FWIW ...
Peter
PS: I for sure will try out a LPSU for the PC once I could obtain one for a "try out" little price. We could even produce it for you all, once that price could remain that "litte price". So don't be surprised when sooner or later such a thing emerges, and I think I have been talking about this before (but that project failed somewhat).
