|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
98.114.91.92
In Reply to: RE: Optical Ethernet posted by jad on February 28, 2021 at 12:03:30
Jad, you are feeding your dac by toslink/optical cable. My dac gets hooked to the server by Ethernet. Supposedly, by converting the standard Ethernet to Optical cable you break any noise that rides on the Ethernet cable. I'm skeptical over a short distance. I was inquiring what people more knowledgeable than me about networking, think about this concept.
Follow Ups:
Ethernet/optical transceivers are not isolation devices, filters, reclockers, or signal regenerators. They simply convert from voltage to light intensity and back again. That's it. So if the differential voltage at the input to the first transceiver contains noise, that noise is simply converted into light intensity by the first transceiver and back to voltage again by the other one. They preserve the noise just as well as the signal.
And they also need +5VDC power, usually from wall-wart SMPS, so they introduce two more possible noise sources that wouldn't be there if you just used straight Ethernet.
One area where optical may be beneficial is over long distances and in noisy environments. The optical cable itself does not pick or emit any EMI/RFI noise. This is more important over greater distances but hardly a concern in the typical listening room.
Thank you all for the education. Very helpful information. Makes me want to delve deeper and really learn more about networking.
On my work projects, we run optical between any two locations that don't share common bonding, which is mostly between buildings. You can't run shielded cable between systems that don't share the same ground, otherwise you'll suffer equipment failures. We also use optical for 10GbE to avoid using shielded twisted pair, just because we've had better luck with that. All 1GbE within a building is plain old unshielded Cat 6. I've never had a problem with network reliability using Cat 6 even in electrically noisy environments.
Given that twisted pair Ethernet is balanced, common mode noise picked up on the pairs won't get beyond the pulse transformers at the end. Differential noise is basically filtered (for lack of a better term) by the PHYs, since the data out of them is latched. And there's no transmitted clock to jitter up. So all I can see is that there's some very short traces on the board between the jack, pulse transformers, and PHYs that could potentially radiate RF that came in on the cable within the case. But 1GbE signal itself is 80 MHz in bandwidth, and higher in level than the noise. If you can keep that out of the analog audio circuits, you can keep the noise out too.
So I personally think this is all audiophilia nervosa. But if CraigS720 wants peach of mind, the EtherREGEN should do that.
So I take it that you're not into high-end audiophile approved CAT8 cables? ;-) There's lower hanging fruit where it makes sense to focus attention and hardwired Ethernet is not it.I have government customers who run media converters for Ethernet and not necessarily over great distances. They are paranoid about any RF emission and optical solves that completely. This was in the days of UTP CAT5 and 100Mb Ethernet. But we have seen over the years that media converters are not the most reliable devices. Maybe that's changed.
Edits: 03/03/21
Thanks for the explanation!
In' no expert but I would think that is you use good cables over a short distance in an RF free environment you should be relatively noise free. I think there would be merit in a system that relies on a remotely located NAS or similar storage system.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: