SOtM sNH-10G - SOtM’s first network switch

[FredericV]

An anecdotal story. Not scientific but I don't care. I cannot explain everything.

We borrowed the sNH-10G  + integrated clock board from the Belgian distributor who was also at the X-FI show in the Klipsch room. We had exactly the same system as in Munich Marriot's Hifi Deluxe show (432 EVO MASTER, Metrum Adagio, Vitus SS-025, Conquistar speakers), except for the switch, which was now the SOtM instead of an 8 port TP-Link easy smart.

The setup was mentioned in Hifi Pig's article, and the feedback on this show was much more positive compared to the Munich Hifi Deluxe show.

https://hifipig.com/birds-eye-view-of-the-xfi-premium-audio-show-2019/

Back home, we hooked everything up as before, but directly with the SOtM switch instead of our cheap metal TP-Link.  I had to wait one week as just after the X-FI the painter wanted to do all the woodwork in our dev / demo room.

On the Amphion Krypton 3 speakers (my own) instead of the Conquistar (our X-FI & Munich demo partner) it also sounded like never before: more refined. Even local file playback sounded better. That was really unexpected and having an IT background: even annoying that this was the case.

Then after a while, I switched back to my good old TP-Link easy smart switch, and both were powered with a 9V Sbooster BOtW MK2 in this shootout. On the X-FI we used SOtM's matching PSU and we did not do any shootouts.

This was very bizarre, as with the TP-Link the sound was more dull and less refined. I never expected a switch could do that, being a former network engineer.

But before the switch, 10m away, we have a fast Microtik RB1100AHx2 with lot's of VLANs and firewall rules, with a modified fan so it's virtually silent, but still with the stock PSU which sometimes makes PWM whine when you are very close to it. So this thing is probably very noisy. At the show we used a very simplistic Microtik haplite as wifi router powered by a 5V Sbooster BOTW MK1, connected to the SOtM, and the SOtM connected to our server.



Being a former network security engineer for companies such as Alcatel, I never expected a switch would or could even do that. But it did. Also note that a lot of data such as broadcasts and the data related to the control app, still go over the network, when audio is being played from local storage instead of Tidal / internet radio / a nas ....

It's a lot of money for a switch, but I ordered it as it also has two SFP slots. There was no way back going to the TP-Link.

I'm soon ordering the Microtik  RB4011iGS+RM as it's a compact 10 gigabit router with an SFP+ slot, so I can make an optical connection between the RB4011iGS and the SOtM. The Microtik is only 170 euro for such a capable router. The SOtM is more then 10 times the Microtik.

My good old RB1100AHx2 is now running on a modified DC inlet (directly soldered silver wire onto the DC input terminals), so I can use an external PSU instead of the cheap 3 amps @ 12V SMPS as I was tired of the PWM whine:



Next up is testing if the  noisy PSU vs the Sbooster BOtW MK2 makes a difference, with the SOtM one step closer to the music server, and the old RB1100AHx2 being my core internet router.

The 13 port microtik is drawing < 1 amps at 12V, which is nothing for the Sbooster.

Theoretically none of this should matter if async DAC's were designed idealistically and DAC designers did their job.

But the hifi industry is far from ideal as most async USB dacs are susceptible for whatever is feeding the signal even though that signal is already bitperfect. Even trying things such as a realtime kernel, it's kernel parameters and even the kernel version can have an influence on the sound. I recently found BFQ kernels to sound less refined than what we had before, but it's being hyped in some "audiophile" distributions.

Some kernel parameters make the sound even boring. Some make it sound more analog. Theoretically, none of this should matter for a bitperfect signal where the DAC is designed ideally.

But this idealistic audiophile world does not exists (yet). If it ever does, high-end music server brands & expensive USB cable brands and aftermarket tweaks related to digital audio will all go out of business as none of this will then matter.

But that will not very likely ever happen.

[FredericV]

The RB4011 looks like a good companion for the SOtM:

1. it has an SFP+ slot so it's future proof
2. it has the same cpu as their 13 port RB1100AHx4 model, which is the successor of my RB1100AHx2
3. maximum throughput should be similar to the RB1100AHx4 as they share the same cpu, so 7.5Gbit
4. it's completely passive cooled, so it tends to run a little hot - the whole case acts as heatsink
5. it has a DC input, so you can use any linear PSU with it, and most linear PSU's will handle it's amperage requirements
 

Max power consumption without attachments = 18 Watts

 

The only drawback is that you need network skills to get these up & running. It's not plug & play. The thing is wide open from factory, so you need to lock down the firewall rules when connecting this to the internet. It has a web gui, a windows gui and a command line mode so you don't need to learn all the commands from scratch.

If you have linux networking skills, including good knowledge of iptables, this thing is very easy to setup.

[FredericV]

For the Microtik, there's also a difference with and without a decent PSU, and this one is the router between my modem-only cable ISP setup (so just a docsis modem without wifi / NAT) and several internal VLANs.

So even when the SOtM switch is in between the Microtik and my music server, cleaning up the PSU of the router still has an influence on the final result.