advertisement


Does connecting a streamer to Ethernet make much difference compared to wifi?

I’m not sure what this means. If you connect an Ethernet plug to a device, Unless your streamer is running on PoE, it is astronomically unlikely to be picking up electrical noise via an Ethernet connection unless something is actually broken.

When I have had electrical noise on an ethernet connected streamer, this has always been a ground loop via a bnc or rca SPDIF connector or usb plug, not via Ethernet.

It's UHF noise, you can't hear it. You can hear the effects of that noise in the reproduction though.

I suggest that you try putting a switch between the router and the audio devices, or a USB isolator like the Intona or the Topping.
 
I know only 2 systems that perform well enough without much attention to the ethernet stuff: Aurender N200 + Audio Note Dac 5 and Grimm Audio MU1 + Chord Dave.
I was well aware of the claim that the MU1 was pretty much immune to network optimising. So I borrowed one from a dealer for 3 weeks to try in my system with the Chord Dave.

The bottom line after the 3 weeks was that the claim is simply not true. The MU1 sounds much better with a good switch between it and the network. I used a PhoenixNET. I mentioned this to the dealer and he agreed absolutely with what I was hearing.

EDIT. Ps. This was not the main reason I wanted to hear the MU1, that was to do with the possibility that it makes Roon sound quality as good as other playback software. Unfortunately it turned out that claim was also unfounded.

A third reason I wanted to hear the MU1 was the claim by a youtube internet influencer (Hans Beekhuyzen) that the Mu1 made the Mscaler redundant when used with the Dave. That also turned out not to be the case. The MU1 4FS internal upsampling was good but did not replace what the Mscaler does.
 
Last edited:
Potentially all the noise on the World Wide Web can find its way into your hifi. To prove this you only need to change your DNS server from 8.8.8.8 to 8.8.8.0. I didn’t believe it would make a difference either, but unless you try first you should refrain from both criticism and scepticism. Bits are indeed bits. But some bits are Marks and Spencers’ bits.

It's UHF noise, you can't hear it. You can hear the effects of that noise in the reproduction though.

I suggest that you try putting a switch between the router and the audio devices, or a USB isolator like the Intona or the Topping.
Ethernet generates RFI, you will notice this if you put an unshielded RCA cable next to an unshielded Ethernet cable. So there may be a justification for using shielded Ethernet cables in and around your stereo. I fail to understand the argument that there is RFI coming from anywhere that can, in some way, be reduced by sticking another switch between the device and the switch in your router. It’s worth remembering that the only thing that differentiates a router and a switch is software.
 
Ethernet generates RFI, you will notice this if you put an unshielded RCA cable next to an unshielded Ethernet cable. So there may be a justification for using shielded Ethernet cables in and around your stereo. I fail to understand the argument that there is RFI coming from anywhere that can, in some way, be reduced by sticking another switch between the device and the switch in your router. It’s worth remembering that the only thing that differentiates a router and a switch is software.

I am very adverse to foo, I would not be raising this issue if there was no evidence (though the ASR church only measures the audio band so you won't go far though).
If you use shielded cables you will defeat Ethernet's galvanic isolation (by linking the grounds at both ends), it's the wrong thing to do.
RFI is generated by the equipments (DSP) and their power supplies.

There is some literature available, for example "The Effects and Reduction of Common-Mode Noise and Electromagnetic Interference in High-Resolution Digital Audio Transmission Systems" by
Jon D. Paul:

"Abstract
High-resolution digital audio systems are especially susceptible to sources of electromagnetic noise from the environment, for example, crosstalk from adjacent cables. The noise can induce errors and increase jitter in the recovered clock signal.
We discuss the most important noise sources and their characteristics. Next, we analyze the noise susceptibility of typical transmitter and receiver circuits. Test results are provided for a system with induced common-mode noise. The paper concludes with circuit design, component and application considerations."


Or some Application Notes by Analog Divices engineers, e.g. "Sampled Systems and the Effects of Clock Phase Noise and Jitter" by Brad Brannon or "Analyzing and Managing the Impact of Supply
Noise and Clock Jitter on High Speed DAC Phase Noise"
by Jarrah Bergeron.

It's cheap to try too, just buy a new or used small entreprise network switch, mine cost £35 new off eBay.
 
It’s worth remembering that the only thing that differentiates a router and a switch is software.
Oh, and their internal circuits, and their power supplies. Apart from that I guess you must be right and they are exactly the same. 😉

Whatever the reason I hear fewer noise artefacts in the music when I have a switch in the system and even fewer with the better switches (all are not created equal).
 
It's cheap to try too, just buy a new or used small entreprise network switch, mine cost £35 new off eBay.
Yes, I tried the Zyxel 5 port and added a LPS. I think the Zyxel was £17 from Amazon when I bought it but now is £20.61 so still cheap enough to experiment.
 
I am very adverse to foo, I would not be raising this issue if there was no evidence (though the ASR church only measures the audio band so you won't go far though).
If you use shielded cables you will defeat Ethernet's galvanic isolation (by linking the grounds at both ends), it's the wrong thing to do.
RFI is generated by the equipments (DSP) and their power supplies.

There is some literature available, for example "The Effects and Reduction of Common-Mode Noise and Electromagnetic Interference in High-Resolution Digital Audio Transmission Systems" by
Jon D. Paul:

"Abstract
High-resolution digital audio systems are especially susceptible to sources of electromagnetic noise from the environment, for example, crosstalk from adjacent cables. The noise can induce errors and increase jitter in the recovered clock signal.
We discuss the most important noise sources and their characteristics. Next, we analyze the noise susceptibility of typical transmitter and receiver circuits. Test results are provided for a system with induced common-mode noise. The paper concludes with circuit design, component and application considerations."


Or some Application Notes by Analog Divices engineers, e.g. "Sampled Systems and the Effects of Clock Phase Noise and Jitter" by Brad Brannon or "Analyzing and Managing the Impact of Supply
Noise and Clock Jitter on High Speed DAC Phase Noise"
by Jarrah Bergeron.

It's cheap to try too, just buy a new or used small entreprise network switch, mine cost £35 new off eBay.
Hmm. I’ve got 40-odd brand new Cisco 8300 and 8200 x’s and an assortment on NIMs and SFPs sitting in storage at work. Maybe I should borrow one of those? Or maybe I won’t bother.

I already have a consumer switch sitting between my Asus router and my streamer and I have tried it without and wireless. It makes no difference.
 
If you use shielded cables you will defeat Ethernet's galvanic isolation (by linking the grounds at both ends), it's the wrong thing to do.
Remember that ethernet galvanic isolation is primarily a DC offset isolation feature (that is why it is rated in terms of DC voltage) and therefore with noise it is perhaps best thought of as being mitigation as opposed to isolation.
 
If you use shielded cables you will defeat Ethernet's galvanic isolation (by linking the grounds at both ends), it's the wrong thing to do.
RFI is generated by the equipments (DSP) and their power supplies.
+1

Use normal standard (read - not expensive) unshielded patch cables (Belkin/Lindy/Etc.) Cat5e or Cat6. There is no reason to use Cat6a, 7, 8 etc. By using the unshielded type of patch cable you avoid linking ground as @tuga points out.

I've experimented with a lot of IT related kit, imho unless your system is absolutely squared away the money is usually better spent elsewhere.

I personally think a separate dedicated mains supply with its own DB and sockets is worth doing and I’d spend funds here first (get a professional electrician to do the work). I plug the networking kit and associated power bricks into the normal mains and hifi into the dedicated supply

Old IT hardware costs peanuts used, adding a switch from a big name like Cisco is only a few £tens.

A popular option is using 2 switches connected by a fibre patch. You can buy 2 used Cisco 8 port switches, 2 fibre sfp modules and a fibre patch cable to split the network will only cost about £100 if you’re patient.

Whether it makes any difference all I’d say is it didn’t change things for the worse but it didn’t change things in a really noticeable way for the better, for the cost judged against a lot of folks system cost it’s not much to spend on ticking a box. I gave them to a mate and now use a £20 Fibre Media Converter because I work from home so my house has a proper comms cabinet with corporate grade switches with built in fibre modules, whether it makes a difference…

 
It's UHF noise, you can't hear it. You can hear the effects of that noise in the reproduction though.

I suggest that you try putting a switch between the router and the audio devices, or a USB isolator like the Intona or the Topping.

How does that noise transit a network device

There is some literature available, for example "The Effects and Reduction of Common-Mode Noise and Electromagnetic Interference in High-Resolution Digital Audio Transmission Systems" by
Jon D. Paul:

That paper is 20 yrs old, much has changed since then.
 
A popular option is using 2 switches connected by a fibre patch. You can buy 2 used Cisco 8 port switches, 2 fibre sfp modules and a fibre patch cable to split the network will only cost about £100 if you’re patient.

Whether it makes any difference all I’d say is it didn’t change things for the worse but it didn’t change things in a really noticeable way for the better, for the cost judged against a lot of folks system cost it’s not much to spend on ticking a box.
I have tried that as well. For me it consistently made the sound quality worse (worse as in more noise artefacts in the music such as over emphasised detail and percussion notes overly dominating the music). I preferred ethernet cable (Cat6 U/UTP) direct from my Cisco 2960-C compared to using fibre with the sfp modules.

I also tried so called audiophile versions of fibre media converters with more or less the same result. The Innuos PhoenixNET is still the best sounding way to optimise the network sound that I have heard so far.
 
I have tried that as well. For me it consistently made the sound quality worse (worse as in more noise artefacts in the music such as over emphasised detail and percussion notes overly dominating the music). I preferred ethernet cable (Cat6 U/UTP) direct from my Cisco 2960-C compared to using fibre with the sfp modules.
Easy for people to try themselves and make their own minds up.

I've never had an IT hardware change make me go wow the same way adding a TCap to a 202 did years ago, it's a very marginal change if any change detectable at all IME - as they say YMMV. I figure for the cost involved and the fact my system is squared away it's worth the few quid for me to put the network switches/fmc in place.

IT is a big market now streaming is so popular and has been embraced by the high end companies - there is also a lot of opportunity to rake in £s selling BS products. I was reading forum posts recently by guys who say they hear sq differences between types of SFP modules and fibre patch leads, IMHO that's pure bollox, the majority come off the same production lines - only differences are the OEM stickers.
 
Based on my experience, I'd simply avoid using a direct wired connection into an ISP all-in-one wireless router (esp. on copper broadband) or if using WiFi, older devices. Other than that, compare for yourself if you can or just use what's more convenient. I use WiFi these days and have sold off my audiophile switch and LPSU.
 
The bottom line after the 3 weeks was that the claim is simply not true. The MU1 sounds much better with a good switch between it and the network. I used a PhoenixNET. I mentioned this to the dealer and he agreed absolutely with what I was hearing.

I meant something different - not the fact it cannot be improved (idk, it possibly can be) but that standalone performance of MU+Dave was quite at the high level without tweaks. Hans Beekhuyzen used Network acoustics filters and also reported an improvement.

And when people saying that another 1k chinese dac majestically solved all the problems it’s just a joke.
 
it is astronomically unlikely to be picking up electrical noise via an Ethernet connection unless something is actually broken.

Please explain why (because it definitely carries noise, there’s no signal without noise in this world if you were not aware)
 
I am very adverse to foo, I would not be raising this issue if ...
oh dear. Where to begin. Look- you want to buy something, you buy something. This is after all basically a hobby for people who need to keep buying stuff in order to enjoy listening to music. But none of the selectively quoted crap has anything to do with needing a switch. Incidentally where did the idea even come from that a switch was a device designed to remove UHF from a wired network?
I suggest that you try putting a switch between the router and the audio devices, or a USB isolator like the Intona or the Topping.
These are very different things. Conflating them only adds to the general picture that people should keep buying stuff because it might just do something/because it is after all what this hobby is about.
Easy for people to try themselves and make their own minds up.
....
IT is a big market now streaming is so popular and has been embraced by the high end companies - there is also a lot of opportunity to rake in £s selling BS products. I was reading forum posts recently by guys who say they hear sq differences between types of SFP modules and fibre patch leads, IMHO that's pure bollox, the majority come off the same production lines - only differences are the OEM stickers.
Not really any different from so many accessories starting with analogue interconnects. As long as people continue to believe that the fluctuations in their listening experience are a reliable detector of minute differences in the electrical output of a hifi component then they will keep "proving" that these things work/sound great/sound horrible.
 
If anyone wants to try a Cisco switch for free I am happy to send it to you only for the price of postage.

It will be a fanless design - something like an 8 port switch.

It will need it to be returned when you have finished with it - or send it on to the next Fishy.

If I have a PoE switch I will run command 'power inline never' so that you don't worry about the power attempting to go over Ethernet. Anyway - I will pre-configure it first to be as optimised as possible for your audio gear (i.e. simple).
 
oh dear. Where to begin. Look- you want to buy something, you buy something. This is after all basically a hobby for people who need to keep buying stuff in order to enjoy listening to music. But none of the selectively quoted crap has anything to do with needing a switch. Incidentally where did the idea even come from that a switch was a device designed to remove UHF from a wired network?

Where have I claimed it was designed for that purpose?

These are very different things.

What, noise travelling over copper? Or UHF affecting clocks and D/A chips?


Starting a sentence with 'oh dear' is patronisingly unlike you, though...
 


advertisement


Back
Top