Ethernet switch test

Michaels HiFi

Well-known member
Joined
Oct 3, 2022
Messages
1,403
Location
Texas
OK guys,

Anyone ever compare off-the-shelf ethernet switches for sound differences in streaming?

I've bought or already own 6 different switches from TP link, Netgear and Cisco and will be doing a comparison shootout.

Has anyone else tried this themselves?
 
I had a TP link at one time. My EE friend suggested I purchase a Netgear switch and he piggybacked glass resistors and capacitors along with addressing some grounding issues. He build a separate power supply / DC blocker for it also.

It was what I would describe as quieter and cleaner sound.

I have no idea how any of this works but he has built or rebuilt so many pieces of audio equipment I trust him to try something when he suggests it.

He is currently working on major upgrades to my ARC 610T amps.
 
I had a TP link at one time. My EE friend suggested I purchase a Netgear switch and he piggybacked glass resistors and capacitors along with addressing some grounding issues. He build a separate power supply / DC blocker for it also.

It was what I would describe as quieter and cleaner sound.

I have no idea how any of this works but he has built or rebuilt so many pieces of audio equipment I trust him to try something when he suggests it.

He is currently working on major upgrades to my ARC 610T amps.

Interesting. Upgrading the switches is definitely the way to go.

I will say I am trying them all in stock form first, then choosing the top two will upgrade from there.

I hear a lot of good things in the past about the NetGear's for audio, and in my testing I was surprised to see the stock Netgear finish a distant last place.
 
OK guys,

Anyone ever compare off-the-shelf ethernet switches for sound differences in streaming?

I've bought or already own 6 different switches from TP link, Netgear and Cisco and will be doing a comparison shootout.

Has anyone else tried this themselves?

There are a couple specific problems associated with using "consumer grade" Ethernet switches (e.g. TP link, Netgear and Cisco) for hi-res audio applications: a) low- and high-source leakage impedance current, b) threshold jitter, and c) phase noise. These cause problems that result in an audible degradation in audio quality when these are use in high-resolving audio applications (these products were not designed for this application). This is why there are now a number of products that've been developed for "high-end" audio applications, e.g. EtherRegen, SOtM sNH-10G, English Electric, Aqvox, Melco, etc., and the considerably more expensive switches from Nordost, etc.

My hypothesis is all the "consumer-grade" switches referenced above will like sound like cr*p compared to the ones designed for high-end audio, they may just sound like cr*p in different ways. Also, using these "consumer grade" switches with the SMPS they usually come with is a sure way to produce even poorer audio qualty, they all require a quality, quiet linear power supply. However, using them with a quiet LPS will do nothing for the amount of phase noise produced from their crap clocks that cost a coupla cents in a BOM.

Then, there's the whole subject of...Ethernet cables. Even the quality of the connectors matters here.

i've mentioned this before, but to understand this better, one really needs to read this white paper by John Swenson.
https://cdn.shopify.com/s/files/1/0660/6121/files/UpTone-J.Swenson_EtherREGEN_white_paper.pdf?v=1583429386
 
There are a couple specific problems associated with using "consumer grade" Ethernet switches (e.g. TP link, Netgear and Cisco) for hi-res audio applications: a) low- and high-source leakage impedance current, b) threshold jitter, and c) phase noise. These cause problems that result in an audible degradation audio quality when these are use in high-resolving audio applications (these products were not designed for this application). This is why there are now a number of products that've been developed for "high-end" audio applications, e.g. EtherRegen, SOtM sNH-10G, English Electric, Aqvox, Melco, etc., and the considerably more expensive switches from Nordost, etc.

My hypothesis is all the "consumer-grade" switches referenced above will like sound like cr*p compared to the ones designed for high-end audio, they may just sound like cr*p in different ways. Also, using these "consumer grade" with the SMPS they usually come with is a sure way to produce even poorer audio qualty, they all require a quiet linear power supply. However, using them with a quiet LPS will do nothing for the amount of phase noise produced from their crap clocks that cost a coupla cents in a BOM.

I totally agree.

Fiber optic is in the next round, then higher quality switches will be in the following round after that.

I'm trying to get a baseline first for what most people start with / will have as a reference piece in which to compare the next steps from.

I felt going right to the higher quality stuff doesn't really give me a baseline to compare to and understand what the improvements are.
 
I tried a pair of Startech fiber converters and a pair of Trendnet switches that had fiber ports in my system.

To my surprise, the less expensive Trendnet switches sounded much better. I had gone out of my way to not buy the cheapest fiber converter and the Startech I bought were about $75 each. The Trendnet switch was under $55.

With the Fiber converters connected to my streamer the sound lacked clarity and sounded congested for lack of a better description. It was a step backward compared to wired ethernet.

Then I tried the Trendnet switches and the sound issue was resolved and improved. Then I added iFi power supplies on the switches and thought I heard an additional improvement.

I should add that I used the same Startech SFP modules with fiber converters and switches.
 
I tried a pair of Startech fiber converters and a pair of Trendnet switches that had fiber ports in my system.

To my surprise, the less expensive Trendnet switches sounded much better. I had gone out of my way to not buy the cheapest fiber converter and the Startech I bought were about $75 each. The Trendnet switch was under $55.

With the Fiber converters connected to my streamer the sound lacked clarity and sounded congested for lack of a better description. It was a step backward compared to wired ethernet.

Then I tried the Trendnet switches and the sound issue was resolved and improved. Then I added iFi power supplies on the switches and thought I heard an additional improvement.

I should add that I used the same Startech SFP modules with fiber converters and switches.

Interesting. Thanks for sharing.

I have both Finisar (recommended by Puma Cat) converter and transceiver and TP link converter and star tech transceiver I will be trying in comparison with each other.
 
I've considered one of the recommended Finisar SFP modules but I'm skeptical if I'd really hear improvements - I think streaming sounds fantastic right now in my system. There is another recommended Finisar module that is "industrial grade" or something like that and sells for around $350 I think. It doesn't seem to be available anywhere at the moment, however.

It's worth noting I also use a Shunyata Omega ethernet cable from the switch to the streamer module built into my Vitus amp and that's the best ethernet cable I've tried. It was the icing on the cake. It does not and will not take the place of the ethernet set up I described above with the switches and fiber. The Omega isn't a short-cut that can take the place of other improvements is what I'm saying.

I'm interested in how the gear you got sounds in comparison to each other.
 
I've considered one of the recommended Finisar SFP modules but I'm skeptical if I'd really hear improvements - I think streaming sounds fantastic right now in my system. There is another recommended Finisar module that is "industrial grade" or something like that and sells for around $350 I think. It doesn't seem to be available anywhere at the moment, however.

It's worth noting I also use a Shunyata Omega ethernet cable from the switch to the streamer module built into my Vitus amp and that's the best ethernet cable I've tried. It was the icing on the cake. It does not and will not take the place of the ethernet set up I described above with the switches and fiber. The Omega isn't a short-cut that can take the place of other improvements is what I'm saying.

I'm interested in how the gear you got sounds in comparison to each other.

I'll keep you updated. Also debating on whether to add a base SOtM switch or AQVOX switch to the test. The loaded AQVOX is the same price as the base level SOtM.
 
I'll keep you updated. Also debating on whether to add a base SOtM switch or AQVOX switch to the test. The loaded AQVOX is the same price as the base level SOtM.

You must have a contact somewhere Michael. You are always going to wonder if you don't try one of those for a comparison.

I assure you we all want to know what you think about the differences in them.
 
You must have a contact somewhere Michael. You are always going to wonder if you don't try one of those for a comparison.

I assure you we all want to know what you think about the differences in them.

Thanks Brad. I know you are right - I will definitely always wonder "what if".

Even with trying out cheaper switches - I know they won't be great but I love to just try things for myself and discover what works and what doesn't. That is the real fun part of this hobby that some people on forums seem to forget about.

I'll be sure to keep you updated.
 
I'm also going to be testing which PORT sounds better (if any) on the switches for the output - the one closest to the input or the one furthest away.
 
I've considered one of the recommended Finisar SFP modules but I'm skeptical if I'd really hear improvements - I think streaming sounds fantastic right now in my system. There is another recommended Finisar module that is "industrial grade" or something like that and sells for around $350 I think. It doesn't seem to be available anywhere at the moment, however.

Just some terminology clarification for the gang in the interests of accuracy: "SFP" is an industry specification of a standard for a port that a specific type of optical transceiver can be used with: Small Form-factor, Pluggable (aka "SFP). The device actually doing the sending and receiving of the optical signals is an optical transceiver.

There is an audible difference in various optical transceivers, and you don't need to spend $350 on a good one. While they're made by a number of manufacturers, the gang at Audiophile Style found that for ER, the Finisar FTLF1324P2BTV sounded the best for most guys in most applications, and this is consistent with my experience. There's also a $10 one from G-Tek that sounds good, as well, but not quite as good as the Finisar. This does not mean you may find another optical transceiver that works best in yourspecific system. Systems, components, and parts can...interact (just like Time*Temperature interact in many chemical or "physics"-based systems).

It's worth noting I also use a Shunyata Omega ethernet cable from the switch to the streamer module built into my Vitus amp and that's the best ethernet cable I've tried. It was the icing on the cake. It does not and will not take the place of the ethernet set up I described above with the switches and fiber. The Omega isn't a short-cut that can take the place of other improvements is what I'm saying.

I'm interested in how the gear you got sounds in comparison to each other.

Agreed, 100% The Omega Ethernet cable is frickin' amazing, and you're correct that it does not take the place of the Ethernet set-up you've described regarding switchces and fiber.
 
Thanks Brad. I know you are right - I will definitely always wonder "what if".

Even with trying out cheaper switches - I know they won't be great but I love to just try things for myself and discover what works and what doesn't. That is the real fun part of this hobby that some people on forums seem to forget about.

I'll be sure to keep you updated.

As someone who worked as a scientist his entire career, if I may provide some advice: Just be systematic in your evaluations. Create a structured experimental test plan, and designate a "datum" for reference. That will the "standard" that you compare other switches (optical transceivers, power supplies, whatever) to. Everything else should stay the same in the system, including the range of musical content, when doing your evaluations. It would also be ideal to pick the same time of day when you're doing listening evaluations. Also, bear in mind that every time you "swap something out", you need to allow time for the new device to "settle". Particularly for "100% digital-domain only devices", e.g. switches, optical transceivers, FMCs, etc. For example, for digital devices, e.g. switches and FMCs, any motion or movement induces vibration in he clocks, and these devices don't like vibration at all. This is particularly true for the crap clocks in consumer-grade electronics. The clocks also don't like fluctations in room temperature, so be sure the rooom is the same temperature for all your evaluations. Also, if you switch out cables, e.g. Ethernet, USB, S/PDIF cables, etc., they need time to settle, also.

Have fun!
 
As someone who worked as a scientist his entire career, if I may provide some advice: Just be systematic in your evaluations. Create a structured experimental test plan, and designate a "datum" for reference. That will the "standard" that you compare other switches (optical transceivers, power supplies, whatever) to. Everything else should stay the same in the system, including the range of musical content, when doing your evaluations. It would also be ideal to pick the same time of day when you're doing listening evaluations. Also, bear in mind that every time you "swap something out", you need to allow time for the new device to "settle". Particularly for "100% digital-domain only devices", e.g. switches, optical transceivers, FMCs, etc. For example, for digital devices, e.g. switches and FMCs, any motion or movement induces vibration in he clocks, and these devices don't like vibration at all. This is particularly true for the crap clocks in consumer-grade electronics. The clocks also don't like fluctations in room temperature, so be sure the rooom is the same temperature for all your evaluations. Also, if you switch out cables, e.g. Ethernet, USB, S/PDIF cables, etc., they need time to settle, also.

Have fun!

Great advice as always and I appreciate it.

I'm just finishing up about 3 hours of testing and my ears need a rest!

I'm going to "reset" myself overnight and start up again tomorrow listening to the current version as the new baseline to move forward from.

Some pretty interesting findings so far!!
 
Great advice as always and I appreciate it.

I'm just finishing up about 3 hours of testing and my ears need a rest!

I'm going to "reset" myself overnight and start up again tomorrow listening to the current version as the new baseline to move forward from.

Some pretty interesting findings so far!!

Cool beans.
 
fwiw: here are two threads documenting my research and implementation of an audio transport and the associated network ... ie: everything upstream from my DAC ...enjoy

> a trip up the digital stream

> the last mile — sonore opticalRendu


and if you really want to get into the weeds:

> phase noise -- what is old is new again?

___________________________________________________

btw: at this point i have moved on to analog -- and pretty much only use digital for streaming titles that i am interested in purchasing in vinyl format. in terms of system and musical enjoyment, could no be happier!!
 
Back
Top