Melco S100 Ethernet Switch Measurements

[Confused]

This month's edition (October?!) of Hifi News and Record Review features a review of the Melco S100 Ethernet switch.  Most Hifi News reviews feature a "Lab Report" section which details measurements of the product reviewed performed by Paul Miller's Miller Audio Research. (AVTech)

 

A copy of the review per the link below, see the "Lab Report" box on the right.

 

https://www.melco-audio-masters.com/uploads/1/0/1/5/101505220/hfn_oct_melco_s100_reprint-low.pdf

 

It is interesting that the measurements do actually show that the Melco provides a reduction in jitter at the DAC, at least when the Melco is used with a Lumin D2, not so much when used with a Mytek or Arcam DAC.  (no measurable difference for the Mytek DAC, 1psec reduction for the Arcam DAC)

 

So this seems to back up the view that if your DAC is "properly engineered", then stuff like audiophile switches should not matter, however, it does go against the view that it is absolutely impossible for a switch to make any difference in any circumstances. 

 

Of course, how much of an audible difference you will get from a reduction in jitter from 15psec to 10psec is yet another topic for debate.

[R1200CL]

Is that price without a 12V supply ?

[plissken]

10 hours ago, Confused said:

So this seems to back up the view that if your DAC is "properly engineered", then stuff like audiophile switches should not matter, however, it does go against the view that it is absolutely impossible for a switch to make any difference in any circumstances. 

 

You are talking about .4 dB spread over 146dB range. I'd ask for show of hands how many here would pay $2200 for .4 dB over that range.

 

Or you could not purchase a $2300 Lumin D2, a $2200 'audiophile switch'. Spend $40 on a used Cisco 2960 and spend the remainder on a competent streamer.

 

Again, these are data and not audio devices. I thing the subjective reviewer can't cash the check his ears are writing. My offer always stands.

 

 

 

 

[jabbr]

It would be far more interesting to repeat the comparison against a 10G fiberoptic switch that costs a bit over $100

[R1200CL]

29 minutes ago, jabbr said:

a 10G fiberoptic switch that costs a bit over $100

Where can I find such a reasonable priced 10G fiber optic switch?

 

I see your point with a 10G switch, and could be well worth testing with measurements. 

[mecani]

I have had the melco s100 in a mytek manhattan and the difference is very great

[mecani]

measurements can tell some data but it is the ear that has to draw conclusions

[ericuco]

8 minutes ago, R1200CL said:

Where can I find such a reasonable priced 10G fiber optic switch?

 

I see your point with a 10G switch, and could be well worth testing with measurements. 

Mikrotik CRS305-1G-4S+IN

[R1200CL]

https://www.amazon.com/NETGEAR-Gigabit-Ethernet-Insight-Management/dp/B01NBBA1HQ/ref=sr_1_30?dchild=1&keywords=SFP%2B+10Gbps&qid=1600113626&sr=8-30
 

https://www.amazon.com/QNAP-QSW-308-1C-Switch-Gigabit-Unmanaged/dp/B07VC9T3WQ/ref=sr_1_21?dchild=1&keywords=SFP%2B+10Gbps&qid=1600114319&sr=8-21

 

So 3 possible candidates 😀

[Confused]

1 hour ago, jabbr said:

It would be far more interesting to repeat the comparison against a 10G fiberoptic switch that costs a bit over $100

I agree. 

 

To me, it is interesting that HFN managed to measure any kind of difference with the S100.  Don't misunderstand this, I do not think that the difference measured itself would enable me to pass an @plissken style blind test, I just find it interesting that anything was measured.  So yes, a similar test with fibre, maybe a similar test with other "audiophile" type switches or indeed a range of standard network switches.  Would there be a range of measurable results?

 

And why the different measurements for the Lumin versus the Mytek?  What actually caused the different results between these two?

 

I do not think a a reduction in jitter from 15psec to 10psec is going to create a difference I could pick up in a blind test, but I still think there might be something to learn here.

[plissken]

2 hours ago, Confused said:

Would there be a range of measurable results?

 

Would there be same sample to sample variance? Either switch or streamer?

 

Again here is the problem: If you start playback on the Lumin D2 and pull the plug how long does it continue to play for?

 

Let's say you get 30 seconds at redbook. That means the data was transferred and buffered up locally 30 seconds ago.

 

So I would like to see this repeated with 48Khz 60 second sample played and then measured while playing out of buffer with the cable pulled.

[Miska]

19 hours ago, plissken said:

Would there be same sample to sample variance? Either switch or streamer?

 

Again here is the problem: If you start playback on the Lumin D2 and pull the plug how long does it continue to play for?

 

Let's say you get 30 seconds at redbook. That means the data was transferred and buffered up locally 30 seconds ago.

 

So I would like to see this repeated with 48Khz 60 second sample played and then measured while playing out of buffer with the cable pulled.

 

I think it is likely some other channel than the data transfer.

 

As I suspected, in this case too, the ethernet isolation is spoiled by use of STP cable:

https://www.melco-audio.com/products/c1ae/

So likely source for the differences are ground currents through the cable shield. HFN doesn't tell what other equipment was connected to the switch and through what kind of cable, because now in this case it matters greatly.

 

I'd guess the measured differences would go away and performance improve by using much cheaper (at most $10) standard UTP6 cable... But of course they insist on using expensive wrong type cable...

 

One can use shielded STP cables only in very specific circumstances, where both sides of the connection are connected to the same common ground point by some other means.

 

Noting here, that the switch likely uses a floating PSU, although not specified anywhere what kind of PSU it has. So it doesn't likely have any ground (earth) connection of it's own. In case of Mytek Brooklyn Bridge this comes from the earthed IEC power connector it has, through the wrong use of shielded ethernet cable. While the Arcam CDS50 is floating (non-earthed) two pin IEC power connector, so in that case the mains earth is not connected through the ethernet cable. While the Lumin D2 is similar to Mytek that it has 3-pin earthed power connection and thus connects the switch ground to the mains earth through the ethernet cable and Lumin device. Then when measured, it depends if the XLR connections are used and how the ground pin of XLR is related to the measurement equipment ground and the IEC mains earth.

 

IOW, you only need an expensive cable to spoil large portion of the advantages Ethernet can offer. Of course with optical it would be much more challenging to screw up, but I wouldn't be surprised if some company decides to come up with optical cable that has metal shield connected to both ends of the cable...

 

[plissken]

27 minutes ago, Miska said:

As I suspected, in this case too, the ethernet isolation is spoiled by use of STP cable:

https://www.melco-audio.com/products/c1ae/

So likely source for the differences are ground currents through the cable shield. HFN doesn't tell what other equipment was connected to the switch and through what kind of cable, because now in this case it matters greatly.

 

Does the Lumin D2 have a shielded RJ45 port? Good catch. That's why I would like to see a write up of the entire testing regimen.

 

Critical thinking for the win 🙂

[mecani]

Has anyone had the opportunity to listen to the melco s100?

[R1200CL]

23 hours ago, plissken said:

Again here is the problem: If you start playback on the Lumin D2 and pull the plug how long does it continue to play for?

Please do your homework 😀

 

The signal from the switch with jitter, noise etc, has already been injected into the buffer. 
You should have learned this by now if you had paid attention to John Swenson’s post. 
 

Besides I think this also depend on clock in the Lumin as well. 

[R1200CL]

4 hours ago, Miska said:

One can use shielded STP cables only in very specific circumstances, where both sides of the connection are connected to the same common ground point by some other means.

This is so true, and is probably only relevant to a very few people here at AS.

The rest of us that has been discussed this here at AS the last 3 years, should keep on with un-shielded cables (like blue jeans), or use JSSG360. And fiber. 

 

If we where building professional network from scratch, indeed we could (or should) use network that applies with the official standards. 
I don’t know how this would affect etherRegen as an example in the loop, but I suppose this is when that earth screw comes handy.

I think first problem that would occur is created by the supplied modem from the ISP. They are normally not grounded. Of cause can be solved. 
 

(I read rest of your post after writing). You are so correct what you’re saying. And it can even be more complicated. If the PCB is connected to chassis etc. 

Thinking of this, does any streamer or streaming DAC comply with 10GB networking standards. Wouldn’t that one day be a requirement as well. 
 

We need to standardize on fiber (or wireless) in our audio network. 

[plissken]

3 hours ago, R1200CL said:

Please do your homework 😀

 

The signal from the switch with jitter, noise etc, has already been injected into the buffer. 

 

Jitter is time domain error. It can't exist in a buffer. John will tell you that. Buffers represent clock domain boundaries.

[R1200CL]

8 hours ago, plissken said:

Jitter is time domain error. It can't exist in a buffer. John will tell you that. Buffers represent clock domain boundaries.

That’s interesting, cause then clock phase noise (which I understand as jitter) should be removed by buffers. And good clocks wouldn’t matter. But they do. I probably need to do my homework as well then 😀

Maybe I’m mixing tings. Thanks for telling me. 
I noticed the new 1.5 version of microRendu has better buffer instead of better clock, so that statement can’t just be marketing BS. 
 

Do you have equipment available to measure switches ?

And would you be interested in doing some ?

 

I just got a Cisco 2960 G version. As an example, should I expect equal measurements from the RJ45 as the SFP port ?

 

[plissken]

6 hours ago, R1200CL said:

And good clocks wouldn’t matter. But they do. I probably need to do my homework as well then 😀

 

Sure clocks matter. The one on your DAC the most.

 

My JRiver system can buffer up to 1GB of data. My 10GBe connection maxes out at ~330MB/s.

 

I concatenated a ~45 minute album into one .wav file and flac'd it to 445MB.  I start playback and in the time it takes from clicking play to unplugging my network cable the entire song is in buffer.

 

What does any switch $20 or $2300 or any upstream clock have to do with it then?

[mecani]

I can be a bit heavy but has anyone here heard the melco s100?

[plissken]

2 hours ago, mecani said:

I can be a bit heavy but has anyone here heard the melco s100?

 

The Melco isn't an audio device. It produces only 25Mhz frequency. I can't hear that high. My best is 14.6KHz.

[mecani]

ok right, someone in this forum has heard a streaming dac connected to melco?

[plissken]

9 hours ago, R1200CL said:

I just got a Cisco 2960 G version. As an example, should I expect equal measurements from the RJ45 as the SFP port ?

 

Which measurements? Ping, and throughput should be the same.

[mecani]

they talk about a product without having tried it

[plissken]

1 minute ago, mecani said:

they talk about a product without having tried it

I heard that having your wall framing made of metal stud makes your sound system better vs wood framing...