You should post here. Everyone likes HiFi porn.
Touche! 
1 Like
Done. Not Hi-Fi porn but I think it is a pretty solid setup. It is in the vinyl rig thread since it has turntables.
1 Like
If you are looking at replacing the standard SMPS brick in an effort to improve the sound of the Nucleus / overall system, don’t bother - it does a perfectly adequate job of supplying a stable voltage and adequate current to the NUC. Yes, it generates an element of RFI because of its design and the unscreened plastic case, but that’s only of any issue if the PSU is sitting in a bunch of cabling at the back of your system, and only then if its cable carrying analogue data which is susceptible to RFI. Digital data is not, unless the RFI is so bad it causes packet loss, but this does not happen! Think of it this way - banks have been using enterprise network switches since Noah got his feet wet. Those switches invariably have multiple redundant SMPS supplies in them. If there was even the remote chance of data corruption / loss, those PSU’s would never be used. If you are so inclined, familiarise yourself with the 802.x Ethernet protocol and how data is actually transferred. While you’re at it, have a ganders at the ISO model and everything will become clear. Its hardly light bedtime reading, but it will, assuming you absorb it sufficiently, allow you to understand why changing your computers PSU, dabbling with exotic network cables, or dropping money on a so called “audiophile network switch” is simply not necessary. Transmitted data either arrives at its destination intact, or it doesn’t. Its a “yes/No” thing and unlike analogue, there are no in-between options. Unfortunately audiophile equipment manufacturers and reviewers would have you believe otherwise…
2 Likes
I’m pretty sure nobody has suggested that data integrity is the issue with SMPS use.
1 Like
No, people just make claims of “noise” affecting the resolution of their “high-end” gear.
I made a -120dBFS 1 kHz 50% square wave test track to test out my system after a (now suspended) user claimed my poor quality streamer (RPi4B running RoPieee) was compromising my DAC output and polluting it with noise. I can pull the test tone into audibility. So far, no one else who has tried it has been able to.
My system has a lot of very clean gain, exceptionally low noise and very low distortion. I have standard, enterprise grade network gear, with steel chassis (which BTW are better at minimising RFI than aluminium), bog standard Cat6 U/UTP cables, no fancy tweaks. The stuff banded around about noise from PSUs, ethernet etc. is nonsense. Pure audiophilia nervosa.
Knock yourself out.
5 Likes
Probably not data integrity but… I will mention here that the USB Audio protocol (which is NOT the same protocol used when you read/write to a USB hard drive!!!) is NOT error correcting, so communication between host and DAC is not guaranteed to be error free. If you had a lot of electrical noise, it is in principle possible that the data could be compromised.
2 Likes
You are right that audio data is not guaranteed error-free. However, if there were errors, those would manifest as clicks and pops, so they would be objectionable. If electrical noise was that high as to induce errors across a 6ft-or-so USB cable, then transferring data from a USB hard drive would also be affected - in the sense that many packets would have to be retransmitted and the data rate would be visibly reduced, and that would also be objectionable. What are the chances of that happening in a home listening environment?
1 Like
Please quote your source, since I think this is incorrect and refers to isosynchronous transfer.
My understanding is that USB audio (UAC2) is asyncronous, i.e., the PC sends bit-perfect data, and the DAC controls the clock. Error detection is provided by CRC. That is, errors are detected, dropped, and resent.
Yes, that’s right.
No, the USB Audio protocol, that part of it which involves sending the actual music frames, is indeed checksummed, so errors can be detected, but there’s no resend protocol. So, only detected and dropped.
2 Likes
To be more specific, “asynchronous” is one of the synchronization mechanisms used with isochronous USB mode.
Some fundamentals of USB and USB audio can be found here
2 Likes
Possibly. The frame rate is pretty short, on the order of a eighth of a millisecond or less. Burst errors would probably be audible, though. But it’s up to the DAC to decide what to do. It could, for instance, interpolate between the previous sample and the next one. Not sure what that would sound like.
1 Like
Yes I agree you’d probably hear ticks and pops, but not sure actually - if one sample got corrupted maybe you would not. For hard drive communication, there’s error checking. And indeed you can have errors with a bad USB cable or one that is not speced for the speed - hard drives transfer data much faster.
2 Likes
I am not sure about this. it is a serial protocol so you could have cases where there is corrupted signals but no pop. Unless there is some sort of verification followed by a mute signal, there could be no tick or pop.
PS: Just read the link above from XMOS. I would say an error must create a gap. What happens is the error-verification (which does exist in USB Audio) verifies that a bulk of bytes transferred is correct or not. So even though the DAC cannot tell the host “there’s an error, resend” it can know that the samples it received are somehow messed up and it’s probably a rule that this should result in a MUTE to the DAC creating a gap.
2 Likes
Yes, it depends on what the DAC does with a corrupted frame of samples.
However, I have experienced different sound qualities over USB while using either a USB repeater (eg USB Regen of Wyred4Sound’s USB regenerator or whatever it’s called). And also when switching to a microRendu from using the computer itself. I chuck this to ground plane noise and stuff like that, I don’t know, but you can hear it.
1 Like
USB is capable of transmitting data rates orders of magnitude greater than that required by audio. Let’s ignore USB 1.0 as it’s very much legacy. Here’s the data transfer rates for USB 2.0 and onwards:
Stereo DSD512 is only 45.16 Mb/s which is less than one tenth of the bandwidth capability of USB 2.0. If USB is dropping data, then something in your system is BROKEN!
Seriously, digital data transfer has been a solved problem for years. Audio data doesn’t even begin to tax the capabilities of modern digital data transport systems.
People need to stop worrying about stuff like this and focus on speakers, their placement, room interaction, room treatments and DSP. Everything else is small fry. Once you’ve sorted that stuff, if you’re of a mind to tinker and squander cash, then knock yourself out. BUT, until you do sort out your speaker and room interaction, anything else is wasted effort compared to the miniscule returns.
3 Likes
Have you any idea how much electrical noise is needed to compromise USB data integrity? Even the harshest of industrial environments manage fine with data integrity over USB. Seriously, I work in environments where very large variable frequency drives are used, electrical motors of > 1MW on HV (>= 3.3 kV) supplies are in service and I’ve never experienced USB data corruption. A typical home on a 220/230/240 V supply with a 100 Amp incomer can max at 22 - 24 kW, and you’d never be switching that sort of load. Industrially, motors of hundreds of kW are switched in an out with start-up currents of >10x running current, even with Star-Delta starters. The level of mains interference and EMI generated is huge, yet all of our instrumetation, PLCs, Devicenet, Modbus, Profibus, DH+, ethernet communication protocols work just fine. Data saved and backed up over USB is faultless.
We have instrumentation which is far more sensitive than audio equipment working just fine in these environments.
Engineers take care of this stuff. As I said before, worry about your listening room and how it interacts with your speakers first and foremost.
5 Likes