Poll and chat: Fibre Media Converters (FMCs)

FWIW I think that's the 10/100 version of the 2960. The 2960cg is the gigabit version and may be more desirable in your network. I wish I could get the time back I spent reading a thread about nervosa surrounding 10/100 vs gigabit for audio applications.

 

Good catch, mine are CGs.

 

Never looked into this (and not sure if I want to), but here are my thoughts:

• The reason why I would do this is isolation from noise coming from the router. So my question would be if the receiving FMC is actually any better in terms of noise than a standard router. And which FMCs are better than which routers or switches? Arguably, the switch/router has to deal with noise itself coming in from a number of devices connected to it and thus might be inherently noisier.
• Does clean power on the receiving end make a difference? And/or on the switch/router?
• How much does cable EMI on the way influence the noise in the Ethernet case? Is that a factor at all, given that we’re talking about twisted pair cables that are rather well shielded, anyway?
• So RPis are rather susceptible to noise? But what about devices that are made to shield away the noise, like a Zen Stream or any other dedicated audiophile streamers? Has anybody tested if it makes a difference there? I’d hope that the difference is much smaller, at least?

 

post 24

 

Thanks, I had read over the paragraph with the Zen.

 

Here is my experience. Switching power supply wire of router run many times thru ferrite. (yes, audible). Then send ethernet to switch with that switcher PS wire also run through ferrite. (yes, audible). If I could get the right end connectors perhaps an iFi supply would make it sound even better, but alas my router and switch use tiny connectors.

Data comes from Ext Spinning drive shared via win 10 desktop. The switch is on the other side of the room with short cable to playing computer using jriver. I did try a second switch right after the router but that did not improve things. Finally, went to short CAT 8 Cables from desktop PC to router and at the switch to connect to the playing computer. Audible benefit vs the CAT6a that was being used. Still have a 30 foot run of CAT 6a from the router to the target switch that has not been replaced.

CAT 5e, CAT 6a are not shielded in their standard configs. There is a shielded version of CAT 6A though, but I never used it. CAT 7 is shielded but really is normally terminated with a different connector and is a bit of an outlier.

CAT 8 is shielded and the bandwidth is improved along with alien crosstalk, NEXT (near end crosstalk) and return loss. I did not think it would make a difference, but it did.

 

As a networking guy, and an ostensible audiophile, this is my opinion, so take that for what it is worth. Network equipment and audio equipment are primarily designed for two entirely different things. Network equipment at its core is designed to deliver a stream of ones and zeros quickly and without errors. That is all it has to do (setting aside routing and packet filtering which are just purposefully manipulating that stream). Audio equipment has to produce wanted sound, whether from a digital or analog source, and in doing so avoid the propagation of unwanted sound. You can use a cheap SMPS on your switch. It won't affect the stream of ones and zeroes, but it may create audible noise that would not matter otherwise. Likewise you can make your router part of your ground loop. The bitstream will still flow, but it may also add a signal you can hear. I can't speak to the efficacy of network equipment specifically targeted for audio, but it does make sense that things like cleaner power and opto-isolation, even if they don't matter for the pure networking aspect, would be beneficial in an audio application.

 

Your comments reflect my experience. In fact, I was considering a fiber run from my router to my best system, but that got me thinking more critically about clean power. I realized that there were a couple switched supplies on the same house circuit as the stereo. A quick look on the oscilloscope showed they were a lot noisier than my linear supplies. I was able to move things around and get the last of the switched supplies out of that system and did find a noticeable audio improvement. The thing about noise is that you have to get it all to see an improvement. The last step was building a Jean-Paul linear supply for the RPi under the Pi2AES and the sound improvement was a strong step forward. It pays to go after the noisiest part of your system first, if you can figure out what that is.

After this process, I'm less interested in fiber and more interested in building quiet power for everything I can.

 

Fiber may be an easier route for those less technically inclined

 

You can also try an iso transformer on switching supplies if you’re really set on eliminating them but it isn’t practical.

Keep in mind some LPS’es can measure like crap, too, especially some of the Chi-fi/AliExpress variety. I’ve cleaned those up in the past by putting LT3045 DC to DC regulator boxes in front of them - cheaper solution than buying some of the $$$ Audiophile LPSes.

 

That's been a hypotheses of mine for a long time. Your experience makes me want to follow in your footsteps a bit more aggressively.