Nucleus SQ ethernet

We find that the SQ improves if we turn off the Nucleus, unplug the ethernet that goes to the network switch, plus the ethernet in again, and then turn it on again. It is not the turning off and on again that changes it but the unplugging and plugging back in of the ethernet cable. The effect lasts for some time–up to two weeks. Any idea why this should happen?

1 Like

So why do you see the need to power cycle the Nucleus at all?
What happens if you don’t?

Perhaps the DHCP lease is 2 weeks? Use a static address, and maybe the improvement won’t expire.

What does “SQ improves” mean? What are the differences in SQ between before and after performing the actions you specify?

Dodgy Ethernet cable or socket ?

You can find if the metal surface is tarnished (oxidized) re-plugging scrapes the copper oxide off and eposes a new clean surface of metal.

Do you get the same effect by changing the cable for a completely different or even New One

1 Like

I’m not good at such descriptions but more depth, clarity, subtlety.

I felt sound had gradually deteriorated and someone suggested turning the cable around, which worked. Someone else suggested turning the Nucleus off and on again.

Periodically we do a power cycle, sometimes partly by accident when we have a power outage. This last time we also unplugged the ethernet cable and plugged it in at the end. This had the most effect.

I am just curious about what it is.

Ethernet cable is one year old red Wireworld cables.

I know zilch about networking, but the address stays the same. If it changes the streamer doesn’t work, so I enter the new one. This time it moved to an old one.

The sound very gradually gets thinner and I begin to think there must be something wrong with my hearing. However in this last case we had a power outage.

Change the cable to a proper qualified standard cat5e ethernet cable and not the boutique Wireworld one. You might be getting a ground loop from unnecessary metal plugs and shielding these things have. See if it helps.


This is the only rational explanation. Otherwise, the Nucleus can’t have any impact on what you perceive to hear.


Let me start by saying that I really have a hard time imagining how that could be the case, given the technicalities of how ethernet and networks in general function.

It’s also hard to believe that a corroded connection would make a difference in the way described, but it’s cheap and easy enough to try.
As a side note, copper contacts are always plated to prevent corrosion, unless re-plugged as many times as to scrape off the plating.

:scream: I’ve found the problem!
Wireworld cables are worn off as soon as burn-in is completed, so they need be replaced quarterly!

Yep, that’s a great recommendation!

Just googled “red Wireworld” – Starlight 8 Ethernet Cable?
If so, that’s shielded and, contrary to wireworld’s prose, contra-productive in digital home-audio transmission.
As @CrystalGipsy says, standard unshielded ethernet cable might prevent audible noise due an existing ground loop, but that would usually “sound” different, and definitely not gradually build up with time as to correlate with your impression.

How could I possibly put that nicely?
There’s just no directionality in any cables, unless they have differing plugs at their ends.

Back to my former question, which you didn’t answer:
Did that make an audible difference?
It’s harmless to just re-plug the ethernet cable connection while the core’s still running when trying to isolate the root cause – better to just change one variable at a time.

I actually share @Martin_Webster’s conviction and that’s why…
I actually think it’s time for you to disclose all your system details:

Describe Your Setup:

  • Details on your Core machine (OS, Hardware specs, Roon build)
  • Details on your Remote(s) (OS, Hardware specs, Roon build)
  • Networking details (especially what hardware you’re using, how everything is connected, and anything notable about how it’s all configured)
  • Audio devices and how they’re connected

In many cases, screenshots of your audio settings or signal path may help.

Use the shortest possible cable between the Nucleus and the player. Cat 7.

This is poor advice. Use a good quality Cat 5e or 6a cable. If you’re installing cables use 6. That’s all you need in the home.


This is also poor advice. Clearly there’s no point in wasting cable, but a standard Cat6 ethernet cable will function correctly up to a distance of around 100m. I doubt you’ll get anywhere close to this in a typical domestic setting.


That’s true for any cable that’s 5e and up when the connection is 1gbe and slower.

Yes, I believe you. I don’t know what the protocol is between the NUC and the streamer. I don’t think it’s TCP. RAAT is used for this. I don’t know if there is packet authentication between the NUC and the streamer. The devices are back to back - the connectors are back to back and the connection is Cat7 and the cable length is 15 cm. SQ is better, especially in the high notes, it’s nice to hear the metal…

Crosstalk can occur between pairs, more interference will be induced on a longer cable…

I think it is TCP. We could use WireShark to find out for sure. Or we could just trust what Roon says about it.

What is “packet authentication”? Sound like a meaningless phrase in the context of Ethernet.

It’s digital. Crosstalk is not a problem.


So why are the individual pairs of Cat6, 7 cables shielded from each other, why isn’t only the outer shielding of the entire cable used? I apologize for the imperfect translation from Czech to English…

TCP/IP is needed to communicate over Ethernet. RAAT uses TCP.

The NUC and streamer will only function if both devices are connected to the router, so this makes no sense. And, it won’t be a “Cat7” connection since those devices have an RJ45 socket, and the cable does not. Thus, Cat7 will be downgraded to Cat6. Incidentally, the female GG45 connector is backward compatible, but a male connector is not, so don’t use them.

This is nonsense. Ethernet uses twisted pairs (differential signalling) for common mode noise rejection. There is no issue, so long as you don’t exceed the maximum cable length specified in the standard, and that’s unlikely in a home.

1 Like

If the RAAT protocol is used, does the music data go directly to the streamer without verification, so with possible errors? Yes, I certainly use a router connected to a switch. The switch is connected to the NUC and I go from it to the streamer via the switch with the shortest possible cable.