Anyone Running Fiber to their Streamer or Network Bridge?

[W9TR]

To answer the question about WiFi- it is potentially way dirtier than an Ethernet connection due to its relatively high RF power output of up to 100mW @ 2.4 GHz / 5 GHz. It can penetrate grounds and cable shields and induce currents there. It can cause audio rectification that shows up as in band audio noise.

Ethernet by comparison is pretty quiet. It is transformer isolated already by design. It runs approx 5V differential and the currents are relatively small. The twisted pair design limits unwanted radiation. The transformers can couple noise across component interfaces, so isolation at the point the signal enters your system is probably a good investment. Analog supplies in the last interface before the signal enters your system also make sense from an engineering perspective. Elsewhere not so much.

[crwilli]

Perfect! Thanks!!!

[IM3CPO]

Hmmm...

According to your diagram, the only cable (in your diagram) that should have any impact is the one running to your USB DAC. No actual conversion into anything close to "audio" is happening until you hit your USB DAC and the portion where jitter/timing comes into play (in your specific setup) is at the computer.

Ironically, your power over ethernet portion adds A TON of noise to your signal. That's one of the reasons those types of links are limited to such low speeds (comparatively). The fact that makes no impact to the sound quality again proves the above statement.

[wklie]

Originally Posted by crwilli

If there is noise in Ethernet, why not simply use WiFi?

SQ from some systems are affected by WiFi RFI. And not everyone can get their WiFi network to stream Hi-Res perfectly.

[Formerly YB-2]

Originally Posted by crwilli

IF fiber optics does indeed not pass on noise and does not generate its own or any other sort of artifacts, the concept seems to make a lot of sense to me.

Craig.......... do you have a link to some treatise/testing that this might be true? Seems to defy the laws of physics to my small brain.

[crwilli]

Glenn, The concept is raised in the first post. I know very little (like nothing) about the types of ‘noise’ in our Ethernet or WiFi or in our power cables etc connections. Since I am a devout Digital only follower, I am keenly interested in things that might provide me a better SQ.

[nicoff]

Originally Posted by crwilli

A quick, although topical, side trip. If there is noise in Ethernet, why not simply use WiFi?

I for one cannot tell a difference in audio quality when connected via WiFi or Ethernet. Except of course where there are problems with Wi-Fi (weak or no signal).

[clpetersen]

Originally Posted by Formerly YB-2

Craig.......... do you have a link to some treatise/testing that this might be true? Seems to defy the laws of physics to my small brain.

Ok, a short answer (hopefully). At audio rates, fibers can easily transmit "short" pulses of light, i.e. a true on/off situation*. Emitters/detectors at either end which can be electrically isolated, turn these back into electrical pulses.
The fiber itself is transmitting light, which is not affected by local electrical and magnetic fields. Nuclear reactor instrumentation tends to favor optical communication for this reason.
So, take a noisy input signal, detect as a '1' or '0', turn to a light pulse, transmit down the fiber, turn back to an (hopefully clean) electrical 1 or 0 at the other end.

* a typical fiber propagation speed is about 2/3 the speed of light. So, taking a 192kHz signal, a 'pulse' of light will be ~500 meters long. The entire fiber will 'light up' and turn off. Even at DSD512 speeds, the pulse will be 5-10 meters long.
This applies to the connection between a streamer and DAC. In most applications, the light pulses are part of a communications protocol, similar to ethernet. (ethernet is a protocol, Cat6 is a cable).

[crwilli]

Wow! The level of knowledge here.

[wklie]

The reason is to use fiber network is that there will not be any EMI, RFI, other electrical noise (or even ground loop due to shield) with the fiber. This is why FMC is used for Ethernet isolation. However, there is still conversion from optical back to electrical signal, that's why a clean power for the DAC-side FMC is essential.

This is well discussed in optical network threads elsewhere.