09-Jan-2017, 14:38
Interesting comments guys, I'm really interested in your findings and am also following the CA topic closely. Can't wait for Romaz to test with his FMC's in the chain as this could also change my plans.
This really is a weird hobby!! I already called all of this 'counter intuitive' as it goes against everything I know/have learned this far. In this case I would surely have never expected things to become better, rather the opposite: removing the noisy computer and moving it as far away as possible from the system/adding proper isolation in between has always been good practice. This at least was true for USB but now with at least the SMS-200 people are reporting that a direct ethernet link between the music server and the SMS-200 is better. I really can't think of any sound reasons why this would be the case and am really puzzled by all of it. Why would any active ethernet components in between have an effect at all, especially when the last device in the chain provides full galvanic isolation and in effect buffers and rebuilds the complete (layer 1 and up) ethernet signal. If it would work in some cases/systems I could think of some valid reasons (ground loops, dirty/faulty switches or cables etc.) but if this shows to be a universal thing...
Besides using the fibre optic cable I realized I can easily test this direct connection using a copper ethernet as well by temporarily re-using a cable that's already running from my upstairs switch to the downstairs switch, I'll just have to extend it a bit and I believe I have a coupling block laying around somewhere to do so. Of course this is not optimal but should provide some idea of what's possible. Perhaps I'll do both tests this coming weekend.
This really is a weird hobby!! I already called all of this 'counter intuitive' as it goes against everything I know/have learned this far. In this case I would surely have never expected things to become better, rather the opposite: removing the noisy computer and moving it as far away as possible from the system/adding proper isolation in between has always been good practice. This at least was true for USB but now with at least the SMS-200 people are reporting that a direct ethernet link between the music server and the SMS-200 is better. I really can't think of any sound reasons why this would be the case and am really puzzled by all of it. Why would any active ethernet components in between have an effect at all, especially when the last device in the chain provides full galvanic isolation and in effect buffers and rebuilds the complete (layer 1 and up) ethernet signal. If it would work in some cases/systems I could think of some valid reasons (ground loops, dirty/faulty switches or cables etc.) but if this shows to be a universal thing...
Besides using the fibre optic cable I realized I can easily test this direct connection using a copper ethernet as well by temporarily re-using a cable that's already running from my upstairs switch to the downstairs switch, I'll just have to extend it a bit and I believe I have a coupling block laying around somewhere to do so. Of course this is not optimal but should provide some idea of what's possible. Perhaps I'll do both tests this coming weekend.
PS Audio P3, Shunyata ΞTRON Alpha Digital and HC/Furutech power cables, Paul Hynes SR7EHD-MR4, DIY Roon Server & Roon Endpoint running AudioLinux Headless, Phasure Lush^2 USB cable, Audioquest Diamond RJ/E ethernet, Uptone Audio etherREGEN, Mutec MC-3+ USB, Shunyata ΞTRON Anaconda Digital XLR AES/EBU, Devialet Expert 250 Pro CI, Nordost Tyr Reference LS cables, Von Schweikert VR-5 SE Anniversary Edition, Anti-Mode Dual Core 2.0, JL Audio Fathom F112. More detail here.
The Netherlands
The Netherlands