Ethernet to Fiber media converters

[Brice_Lang]

I decided to keep the fiber cable I already had (Tripp Lite Duplex Multimode 62.5/125 Fiber Patch Cable (SC/SC),(N306-006) and replace the TP-Link MC200CM (gigabit model) with the MC100CM (10/100mbps model). I installed them with the dip switches in the stock settings and they are working just fine with my DEQX PreMate Plus, which has a 10/100 ethernet board. It was harder than I expected to find compatible gear for my system, but I did it! Thanks to those on this forum who helped me.

[dabassgoesboomboom]

That’s all great… but how does it sound !?

Better, worse, same?

[Brice_Lang]

I installed the fiber converters just before I posted and the dishwasher has been running this whole time in the kitchen adjacent to the listening room…so I honestly can’t say yet. Ha! I can report back tomorrow. I have a sufficiently revealing system, but I’m honestly not expecting much difference. We’ll see.

[dabassgoesboomboom]

I installed the fiber converters just before I posted and the dishwasher has been running this whole time in the kitchen adjacent to the listening room…so I honestly can’t say yet. Ha!

Ha no worries. I never found it to be a night and day difference of course. But there was a bit more smoothness to the sound, just a little bit. And a little bit more bass. Something I noticed after extended listening and not immediately jumping out at me.

Not a large cost to try anyway.

[Brice_Lang]

Sounds a bit better. It is a marginal improvement. Hard to describe how exactly. I’d love to say I can hear a lower noise floor–that may in fact be the case–but the difference isn’t so great that I can hear the precise difference. I’d say I got what i paid for.

[dabassgoesboomboom]

ROON ROCK with a linear power supply?

Recently I have been developing radio enhancers and using data links with FMC’s, guess what, the noise introduced by FMC in the 812 - 813 MHZ bands is way more and powerful than ethernet

Hi zbyte. Which PSU was used with the most downstream (DAC side) FMC in your test? An SMPS?

I have a pair of TP-Links where I had the switching regulators removed and replaced with low noise linear regulators and I’m able to power the FMC’s with 5Vdc - including powerbanks… which means you are truely ‘off the grid’.

If using an SMPS (and even an un-grounded linear PSU) to power the last FMC, there is still a path for leakage currents to go through the DAC, according to John Swenson’s current personal research.

[zbyte]

Hi Sean:

The Fiber Media Converters are all powered by SMPS’s, assuming you can use a linear power supply to power these the leakage is not on the electric part of the circuit but on the oscillator / modulator needed and included within / inside the FMC to send the “data stream as laser” creating a big RF spike with enough power to interfere “within that frequency” with any device around in a 30 - 50 ft radio. Let me see if I can get some pictures.
Actually I have a video that maybe I can embed in this threat.
I tried several different manufacturers, with SFP’s and without SFP’s
Apologies for my accent, english is not my native language, second video is in spanish
Enciendelo means turn on - Apagalo means turn off

You’ll see the spike

Now if this affects “audio SQ” is yet to be seen and tested

I know that RFI will use the SMPS cable as an active antenna and will multiply the effect with any cables

Here it is straight from my dropbox

RFI

[zbyte]

My point is that if 812 MHZ RFI affects SQ then you may be better off with wired ethernet than fiber

If it doesn’t then fiber is good.

And I wrapped those transceivers in the video with copper tape, aluminum foil, several layers (people around me was asking if I was baking something), put a ton of ferrites on the cable. RFI still was strong, it laughs at solid aluminum foil and copper foil

I did not test tin braid which I believe because of the mesh structure may be way better to diminish the RFI effect (JSSG360?)

[zbyte]

One more thing, if the FMC uses 1310 for Tx and Rx on 2 strands the frequency is contained within a few Khz band

If you are using the single transmitters over just one strand BiDi which Tx is both 1310 and 1550 then the band is wider and RFI higher. That’s the second video I believe

[dabassgoesboomboom]

Thanks for sharing @zbyte !

Ok now it’s clear that you are only talking about airborne RFI, whereas I was talking about conducted (i.e. leakage currents and their associated RF pickup and radiated).

So when I was talking about the 5Vdc battery powered FMC’s, I have no leakage currents and therefore no RF pickup/radiated by the last ethernet cable near the DAC.

As Rob Watts says:

“battery mode operation should mean no ground loops, so no current flow into the ground planes, then no RF noise pick-up in the DAC and so no problem…”

As you probably know, John S also says if you can block the leakage currents getting into the cable, then shielding isn’t required (even JSSG), for the same reasons.

But yes, airborne is something else and definitely a big problem. You can use a long unshielded ethernet cable from the last FMC to avoid this issue but I’m surprised it needs at least 30 ft distance though.

Since my FMC’s clocks are powered by low noise linear regulators I wonder if this helps? Maybe not. The only way to know would be to test but I don’t have the tools you have.

It would be interesting to see if their theory that no leakage currents in the last ethernet cable really does mean no RF pickup / radiated. Are you able to power your last FMC with a battery and use a 30ft unshielded ethernet cable?

Of course if the last FMC is SMPS powered then you may still need proper shielding of the last ethernet cable, even if it’s 30 ft. If the theory is correct.

Thanks for sharing.

And I wrapped those transceivers in the video with copper tape, aluminum foil, several layers (people around me was asking if I was baking something)

Hilarious

[zbyte]

Correct, airborne RFI sounds about right
RF is a complete different animal, if you look at the video you will see the spike 30 to 60 dbm if I’m not mistaken, 60 dbm in open space is HUGE
I have all the FMC’s here at home as I could not use them anymore for my RF installs, I do have a JS2 which I’m almost sure can power the FMC’s voltage and power wise and with the spectrum analyzer compare the results.
The more puzzling thing is we assume aluminum foil will block RF (thinking on the crazies with foil hats), it is used in cables to block EMI in general but from my testing RF still comes out but I remember at the time I could not completely confirm if the RF was still being emanated from the FMC or from the SMPS cable attached to it acting as an antenna.
I guess I can use some good old thick Reynolds foil and “bake” cable and FMC and take a video with the spectrum analyzer to compare results.

[dabassgoesboomboom]

The more puzzling thing is we assume aluminum foil will block RF (thinking on the crazies with foil hats), it is used in cables to block EMI in general but from my testing RF still comes out but I remember at the time I could not completely confirm if the RF was still being emanated from the FMC or from the SMPS cable attached to it acting as an antenna.

This is what would be interesting to test this - if you used the JS2 just on the downstream FMC only (but make sure you ground it’s DC plug). If you use JS2 to power both FMC’s you defeat the isolation.

You mentioned JSSG360 earlier but I don’t think JSSG360 is required for better shielding than the ‘original’ JSSG method. BJC Cat 6a is based on Belden 10GX series cable and features a floating shield. You just need to create a ‘loop’ by soldering silicone wire to the ends of the foil shield and running it externally and parallel to the Cat 6a cable, and you have the Swenson shielding technique. I used to do this but now prefer to block the leakage currents completely (for critical listening) - which allows me to stick with unshielded Cat 6 cable.

Correct, airborne RFI sounds about right
RF is a complete different animal

Agreed

[zbyte]

if you used the JS2 just on the downstream FMC only (but make sure you ground it’s DC plug)

FMC is 5v 1A so I will power one of the adapters with the JS2 which is currently grounded and the other FMC adapter will be connected using its original SMPS. We will be testing emissions (RFI) out of the FMC without foil and then with foil and will compare the results. Sounds valid?

You mentioned JSSG360 earlier

The only reason I mentioned this is because JSSG360 is an entire mesh so it seems the results are better that means the mesh provides better isolation than solid foil.

[dabassgoesboomboom]

FMC is 5v 1A so I will power one of the adapters with the JS2 which is currently grounded and the other FMC adapter will be connected using its original SMPS. We will be testing emissions (RFI) out of the FMC without foil and then with foil and will compare the results. Sounds valid?

Sounds good. But I just thought if it’s the original SMPS that’s the issue then maybe next test would be a linear PSU on each FMC.

Or better, just a 5Vdc USB powerbank… it’s easy to find USB to 5.5 x 2.1mm DC plug cables:

https://www.adafruit.com/product/2697

Then you are truely disconnected from mains RF

[zbyte]

I have a power bank that I can use indeed

[dabassgoesboomboom]

Their Vdc output is noisy but their benefit is free from mains RF / leakage currents etc. In theory That’s all I can go by since I have no means to test like you. Will be interesting to hear what you find

Btw, do these RF emissions you measured with stock SMPS’s exceed the standards, where you are? Or they are acceptable?

[zbyte]

Sean the emissions are off the charts, no standard at all and in my opinion has nothing to do with the mains so I believe whichever test I do it will have these emissions present

[dabassgoesboomboom]

Sean the emissions are off the charts, no standard

On the back of the units you have the relevant safety approvals etc, so these would be the applicable EMC standards, no?

I’m not an expert at all so apologies if these are silly questions.

[dabassgoesboomboom]

Hi @zbyte

Did you manage to do any more of the above testing? If so, what were you own conclusions / thoughts?

Like powering each FMC with a 5Vdc powerbank, to remove their stock switching PSU’s from the chain and area?

[dabassgoesboomboom]

@Magnus

In other words, my signal path is: Roon on computer -> switch -> ethernet -> fiber -> ethernet -> Allo USBridge -> DAC. I run my 5v DAC, the USB part of Allo USBridge and the last fiber converter on batteries (which made a noticeable improvement btw), which I charge when sleeping.

So now the question is: would I get a (noticeable) improvement on SQ if I moved Roon Core to its own fanless NUC (optionally LPS powered) running ROCK?

I thought I’d reply in the tinkering section

Assuming you don’t want to change DAC of course:

From experience (not expertise), having had a very similar chain to you (including FMC’s, USBridge and S2 DAC)… the best ‘bang for the buck’ won’t come upstream of the fiber optic cable…

Something a little over the budget of the NUC that helped the S2 DAC performance for me, was upgrading the USBridge to something like an ultraRendu…

Something well below the budget of the NUC (and given the USBridge is already very good) is what I did with my FMC’s… I had SOtM upgrade my FMC’s. They replaced the switching regulators for low noise linear regulators and upgraded some capacitors. This also allowed the FMC’s to run on 5Vdc (and you know what that means…).

This FMC upgrade cost under $150 from memory and helped… to my ears of course.