There’s 3 possible things (and their combinations) that can ethernet system/cables sound different:
Noise going out from ethernet cable and interfering with nearby gear
External noise getting in
Different behaviour in the PHY as a result of the ethernet cable employed (signal integrity?)
I’ve suggested the ethernet pullout method before to see if it makes a difference. Unfortunately, that’s useless with Roon as the buffers are so small.
If one thinks it’s noise getting in and that’s why cable A sounds better than cable B or C, it’d be fun to do this in a professional shielded room and see if the results still hold up. In that case, one can’t claim external noise being a culprit.
Awhile back, I tried putting over 100 feet of ethernet cable directly on top of my interconnects and DAC to see if it could mess with things and then unplug it and whether it’d make a difference after.
I know these things can sound different from one another, but until a manufacturer shows measurements with a proper theory, I will not be buying a single thing from them. They need to show and convince me that they actually know what they are doing and know what’s really going on.
I’ve been giving wireless a shot again. It’s much simpler than ethernet and more easier to control the variables (router or switch / power supply/ cables). Better isolation. Cheaper.
I made a comment somewhere else about Lumin’s recent fiber input. If external noise getting in or noise getting out was the issue, then fiber should solve that. However, if one still hears difference by upstream differences, then you’d have to broaden the scope of investigation:
It’s good to see manufacturers giving the option for fiber inputs.
Since Lumin will have total control for the downstream fiber implementation, I’d now be curious to test out the hypothesis regarding noise in different domains. I’d be interested in hearing comments from the manufacturers and users whether changing the power supply of the upstream FMC can make a difference or not or if the length of cables make any difference, or on the clocks of those upstream FMC. Of course, the power supplies of those upstream FMC should be on other circuits and you could furthermore put the Lumin on a good isolation transformer to get separation, so one can’t claim that it’s the power supply interaction rather than the signal at hand. Batteries can be used for isolation as well if needed for true isolation.
If there is no difference heard, then problem solved, everything works as expected. End of discussion. However, if there are differences…
I live in a first floor apartment, next to a radio station, have wifi, an open kitchen in my living room, etc… I had fun waving around an AM radio in my living room at various times of the day, putting it up close to different equipment, turning them on and off, to identify “culprits”.
The interference you pick up through an AM radio is scary!
As a result, I changed a few cheap switch mode power supplies, got into the habit of turning off some equipment (TV, and others) when listening to music, replaced some led lights in my kitchen, and finally tested out this switch/cable.
None of this cost much, and I think it makes a difference… Only peace of mind perhaps?
In my offer of my $5000 to someones $1000 I stated that I could take their current cable, tone it out, and construct a like cable on the spot that would render it indistinguishable.
Also you have modified a network switch in addition to the cable. So I find your post enthusiastic, but also a bit disingenuous.
I’ve also recommended just getting a used Cisco or HP Procurve switch that has a 3 prong IEC.
Easy now… replying to me directly to say disparaging things like “disingenuous” is the sort of name calling that I believe is frowned on by the mods here.
Though it is interesting that anyone could find a post that has the following words…
as anything but the exact opposite of disingenuous. I think I was quite forthcoming with relevant detail. And I never attributed the change to anything specific, and made it clear that I did not know what was responsible for the difference I heard.
Some recent posts raise an issue that Admins (the devs) have raised with Mods in the past: the not quite infringing post that can draw an infringing response. “Disingenuous” has been modded previously when accompanied by an angry tone which drew an infringing response.
I recommend users keep an eye on the subjects of their posts:
Am I talking about gear or audio ? Usually not a problem.
Am I talking about myself ? Again, usually not a problem.
Am I talking about another person ? Now you’re in the arena.
Try and reframe things to be about gear/audio, the idea or your own experience.
I found CA’s solution a bit of a kludge. So I tried to improve on it a bit. I just got a simple ring loop terminal , bent it at a 90 degree angle, and soldered it to some green (because it’s a ground wire) insulated, stranded 18 AWG cable… stranded for flexibility, 18 AWG because the Computer Audiophile thread indicated wire in the 20+ AWG range was sufficient, so 18 was overkill. And soldering was not necessary. Crimping the connection point with pliers would work too. At the termination point I attached a simple 3 pronged plug, clearly only connected the green ground wire to the ground prong.
Then, for cosmetics and ease of use, I pulled both the green wire and the power supply wire thru some plastic mesh wire covering, and slipped the ring loop connector over the power supply’s plug barrel. And then finished it off with 3/4" heat shrink tubes on the ends, But the typical 2:1 shrink ratio was problematic with the wire behind the loop connector. In only just barely shrinks enough. Final product below.
I should also mention that the ring loop terminal I had laying around had too small a hole to accommodate the power supply’s barrel. But momentary drilling to increase the hole size with one of these step drill bits made that problem disappear, and the barrel fits perfectly now, just a tiny bit snug for good contact.
Oh… I should note that I did not provide enough wire for the ground plug to be plugged in much more than a foot away from the power supply. That’s because John Swenson indicted that the two plugs should either go into the same socket, or be plugged into the same power strip.
That it does have a 3 prong power cable does not imply its IEC socket ground is connected to the ground of the PCB or the Ethernet ports. Many SMPS do not do this - most of the external SMPS do not, at least some of SMPS do not either. Please use a multimeter to test the AC ground is connected to the ground on the PCB or the Ethernet ports. If and only if this is tested, then yes, yours would be valid answer.
As for this particular brand, I’m sure it’s excellent for IT professionals or corporate use. However, too many times I see this brand in Roon network setup troubles, and I recall one of which was never solved without changing to another brand, because the particular model used by the user is not configurable.
Nice post BTW (delayed read). I would wager that the number of ‘golden ears’ you found in your tests can be explained by randomness alone.
Do you have any followup tests of the golden ears to see how consistent they are? Say, MQA vs CD? Just curious. Great topic for your dissertation. I hope you have plans to write a version for the masses and publish to a wider audience.
As CCNP I see users get into trouble all the time with L3 managed switches. You HAVE to know what you are doing since there is complexity with the switch: Port Fast, Etherchannel, LACP, Trunking, STP… The list goes on and on and on. I feel for Roon/Lumin/NAD/NAIM when some clueless end user calls up with a managed L3 switch and want help. You aren’t their network tech. Feel free to send them my way. My rates are reasonable
But with that said you can get the 2XXX series (layer 2 switch only) and that cuts out a lot of the headaches. 2690’s are available easily for less than $100 and absolutely have better build quality than your no name switches
I’m also a hobbyist speaker designer with everything from measurement mic system to a $1700 ADC. I’ve yet to see anyone produce any empirical data that shows Ethernet cables make a difference.
Simple networking is easy, advanced networking is hard. It gets harder by introducing managed switches and esoteric detours like bridging to the mix without rationale (or any clue of what’s going on).
Keep it simple. Router. Switch. Cable. Avoid WiFi and Powerline adapters. Run Roon Core on Linux (ROCK!) on an inexpensive but dedicated NUC. Keep storage local to the Core.
Enjoy stable Roon and the quiet in the Support Forum.
Testing if the high impedance leakage goes to ground with an ethernet switch is a little different and more complicated unfortunately. A simple continuity test isn’t enough. Best to stick with the Swenson recommended switches unless you really know how to test it.
Btw the new cheap Uptone branded SMPS (7.5Vdc and internally grounded) powers the Swenson tested and recommended Netgear GS105 and GS108 unmanaged switches.
Rene, Having started with Roon and “advanced audio” six months ago, I can tell you how valuable your advice is, and how fortunate I was to get similar advice from this community.
I did everything you suggested above and have never regretted it. While I over-specced my storage and RAM (I floated the term ‘JOCK’ but it didn’t take), all of the rest is, well I said it already, good advice.
And, I have noted that a large majority of support requests involve marginal setups. I quickly cried ‘uncle’ trying to use my old NAS and dated MacMinis. They worked…but just barely, then got worse as I added load to the system.
For the price of their SMPS and a NetGear GS105/108 switch and the technical mechanisms (leakage current loops) John Swenson discusses in detail on CA Forum, just give it a crack !
