Optimal lengths for USB and Ethernet cabling for digital audio
Hello, I have learned over the years that for coax digital cables, like those with RCA or BNC ends should be 1.5-2 meters in length to avoid some digital reflections being transmitted. Does this apply to USB cables and Ethernet cables that carry music in data form when a person uses those types of cables from network to DAC or streamer or from a streamer to DAC?
It seems to me the USB or Ethernet cables have more specific wires in them, and they carry power and computer communications data in those wires along with the digital music signal, but that may be, they still carry the music that will be decoded later.
USB 1.5m with seperate output tails, most DACs don't need USB power so tape over pin 1 in the output power tail, or pin 1 if it's an ordinary single cable. USB cables offer two way communication so length is an issue. My system uses Ethernet cables between 0.4m and 1.2m, shorter seems better but I haven't done any critical experiments. |
I don't recall reading anything about reflection issues with ethernet or usb cables. Makes sense ethernet cables should be shortest possible to minimize rfi, although can still get in through connectors. For usb, seems 1.5m is pretty much standard length used, I've never deviated from this length.
As far as ethernet vs cable provider coax, I've found moving cable modem close to system via longer coax/shorter ethernet far preferable to shorter coax/longer ethernet. Another advantage is one can afford much better quality ethernet cable, 1 or 1.5m AQ Vodka vs aprox. 18m price, forget about it. |
it seems curious to me that it is commonly accepted that spdif rca digital cabling should not be 1 m in length or shorter, 1.25-1.5 m better to reduce chance of reflections... but the voltages expressing 1's and 0's are still traveling down the wire pairs in usb or ethernet cabling... why are they immune from such reflections, if they exist? is there something about the 75 ohm nature of the spdif cabling that makes this matter? |
steve nugent of empirical audio, who used to post sometimes here, but no longer, is one of the experts in this area (digital audio, jitter, cables etc...) -- it was his writings in Enjoy The Music some time ago which started quite a bit of useful discussion on managing jitter and read errors in digital systems data transfer, cabling etc etc... https://www.enjoythemusic.com/magazine/manufacture/0509/ https://www.empiricalaudio.com/computer-audio/technical-papers/ https://positive-feedback.com/Issue14/spdif.htm https://positive-feedback.com/Issue43/jitter.htm i shot him an email this morning, asking this question - his answer is that he feels usb cables carrying digital music data are prone to the same potential reflection issues as spdif rca cables, and thus, he uses 1.5m USB cables just to be safe... |
