On an Ethernet cable, if only one side is grounded, and the other side left floating, it’ll become an antenna.
There are a few audiophile vendors who thought it’s a good idea to take that asymmetric grounding technique from (low-speed) analog signal world into Ethernet cables. They’re badly mistaken.
audioquest is among them; for example, their RJ/E Diamond Ethernet cable (I own one) is very well constructed, and measures quite well, BUT it’s “directional” meaning they ground the cable shield at just one end, which turns the whole cable effectively into an Antenna, not to mention violates Ethernet cable specifications.
Than it is strange that even a company like Melco (Buffalo) recently launched such a cable - Melco C100.
So …
If a CAT cable based on S/FTP would be used, than the only way to secure that noise - from transmitter to destination - is not entering a streamer/DAC - is to use scotch tape (RJ45 connector on the transmitter side is stripped with scotch tape)?
Or du you have another solution for a S/FTP cable setup?
If you do that, then your fancy asymmetric shielded twisted pair Ethernet cable becomes an UNshielded cable. Though with a metal foil around the twisted pairs - which then acts as a conduit for unwanted induced interference between the various pairs.
Best is to use an unshielded cable from the get-go - just as dCS advises.
Thanks Adam, I’ve recently moved to their Silver Diamond speaker cables and am very pleased with them….so it’s good to hear others also having good experiences with them.
clearly designed to pander to Audiophiles. That C100 violates standard Ethernet cable specifications. It’s not just how much EMI/RFI it attractions, but also how much it radiates out that could potentially affect the rest of your system.
Paul is exactly right. For home set-ups, just don’t use shielded Ethernet cables to your Audio gear.
For home use with audio kit, the Ethernet signal-to-(EMI/RFI-induced)-noise-ratio on unshielded cables is high-enough and usually out-weighs the potential problems associated with using badly constructed Ethernet cables.
I am currently using Belden Bonded Pair CAT6 (UTP) in my audio setup.
But from FMC to DAC/Streamer I use 50 cm Patch cable made from Cat.7 raw cable with Cat.6 connector - S/FTP - PiMF (paired in metal foil) - RJ45 connector on the transmitter side is stripped with scotch tape and with ferrite chokes:
@torbenrick out of interest what spec or part number are the ferrites? I am interested in what frequency noise you are targeting and whether you tried any others before deciding on those? Thanks.
Personally, you’re seriously complicating that connection for no reason - adding Ferrite cores to a Cat 7 cable is literally a contradiction; the former negates the benefit of the latter. Why would you do that?
dCS’ Ethernet Port magnetics have a proper common-mode choke that shunts noise to ground.
By adding ferrite chokes to your Ethernet cable, you’re negatively impacting the Ethernet physical layer digital coding signal transmission (PAM-5/GbE).
In other words, you’re going to see an increased bit-error-rate on your Ethernet connection because of those ferrite chokes.
Is the sole consideration in this assertion RFI/EMI-induced noise?
Are you suggesting that if Torben had added ferrites to an unshielded ethernet cable then there may (not will) be benefits, but doing so to a shielded cable will have the opposite effect?
If you could unpack this for the hard-of-understanding then the rest of us might learn something about the mechanisms at play here (FYI I use unshielded cables myself - and without ferrites).
This feels like an entirely different point. You appear to be suggesting that the addition of ferrites might mess up the voltages used in PAM-5 to the extent that the actual digital data errors might be introduced. Wow.
As above, if you could unpack this for the hard-of-understanding then the rest of us might learn something about the mechanisms at play here. Or not!
Well, first of all, I don’t think he really meant CAT 7, which was a proprietary development by a handful of companies that wasn’t originally part of the ANSI/TIA standard. It was for 10G Copper interface with a GG45 TERA connector, NOT the standard RJ45 connector.
The ANSI/TIA standard for Ethernet cables went from CAT 6A (10G) to CAT 8 (25G & 40G) for Copper interfaces. So, his CAT 7 is likely just CAT 6A disguised as a supposed “CAT 7”.
Secondly, the proprietary CAT 7 cable is a fully shielded cable designed to protect a 10Gbps/600MHz signal transmission over a distance of as much as 100m. Those speeds at those distances means the cable was designed to shield itself from EMI/RFI that could affect signal integrity.
Ferrite Beads are basically an inductive element that attenuates high-frequency signals. When used on things like unshielded AC power cables, or other unshielded low-speed digital cables, it works to attenuate high-frequency EMI/RFI.
But when used with a CAT cable that is supposed to carry a high-frequency digital signal itself, those ferrite bead will impede the actual digital signal itself, resulting in possible high rate of errors.
That’s a good explanation. It follows that if you’re going to use ferrite beads at all you’ll need to have (a) a good understanding of the range of frequencies you want to flow unimpeded - and by implication those frequencies outside this which you seek to stop in their tracks - and (b) the ability to design and implement with a fairly accurate (and ideally steep?) cut-off.
If there’s an overlap between the frequencies of the “music” (data) signal you want to travel unimpeded and of the RFI/EMI you want to mitigate then you’re stuffed (to coin a phrase). Might there be?
I believe CATn cables operate in the hundreds of MHz and RFI is broadly defined as low kHz to multiple GHz so in theory yes.
Yup. Plus, Ethernet signals are especially sensitive to impulse response impediments within the cable because unlike traditional audio carrying connections where there’s a continuous signal traversing the wire usually at a fixed frequency, whether or not there’s music playing, Ethernet signals on the wire - i.e. the high-frequency analog signals representing coded digital 1s and 0s - are sudden and bursty in nature, it’s not continuous (there’s a self-synchronising 7-octet preamble to initiate every new data stream to the receive side).
So, it’s generally a very bad idea to add anything onto an Ethernet cable that could affect the cable’s impulse response.
Those who believe RFI noise on ethernet cables is potentially an issue in the analogue domain further down the playback chain (and I know this doesn’t include you @Anupc !) might mitigate this noise through use of ferrites but in doing so not be aware that they may degrade the digital signal. They will perceive potential upsides and downsides, and presumably use their ears to determine if on balance they’ve caused an improvement or deterioration.
Those who believe there is no RFI issue to address will see adding ferrites as a pointless exercise in the analogue domain potentially inflicting pain in the digital domain: no upside and either no impact at all or downsides.
I don’t think there’s much to be gained by raising our hands to indicate (again!) which camp we’re in… though I don’t think I’ll be adding ferrites to my own short ethernet cable.
clearly designed to pander to Audiophiles. That C100 violates standard Ethernet cable specifications. It’s not just how much EMI/RFI it attractions, but also how much it radiates out that could potentially affect the rest of your system.
