Optical Network Configurations

[jabbr]

and also regarding SFP modules -- I have been happy with all the fiberoptic modules I have, and I have many. I particularly like the dual 10/1g modules made by Finisar. I pick up mine on ebay usually. The only modules I have ever had a problem with are the RJ-45 modules either 1g or 10g ... in that case I stick with the Mikrotik 10Gbase-T module for like $50-60 ... 

[Foggie]

41 minutes ago, jabbr said:

I particularly like the dual 10/1g modules made by Finisar.

You have a model number?  Would these be pretty interchangeable with existing GB switches?  However, I plan on getting a 10GB optical / managed switch sometime in the near future (HP, Ubiquiti dunno yet)

[jabbr]

So these are the Intel branded modules: https://www.intel.com/content/www/us/en/support/articles/000005528/network-and-i-o/ethernet-products.html

 

and you can see that FTLX1471D3BCV is the base Finisar LR module

 

check that your switch supports dual mode -- most should

[Tecnik1]

Having set up my optical network based on this very thread and a couple of others I'd like to share a cheap tweak I implemented. As I was reading through the list of modifications on the Aqvox SE switch and being a tweaker this one caught my attention.

 

- The lining of the metal housing with butyl rubber and thereby silenced also contributed to the sound improvement. A more black background.

 

This tweak does indeed improve the sound as described on my standard Aqvox switch along with a second basic Cisco switch and both sending and receiving FMC,s. I also installed teflon washers between the PCB boards and metal frame with the idea of removing any vibration.

 

Im pleasantly surprised with the SQ improvement of such a tweak that cost me nothing but a few hours of time.

[tgb]

Hi,

Which fiber "sounds" the best ? single-mode or multimode ? 🙂

 

My optical network is SM fiber based, from 2015.

At that time I've read somewhere that SM sounds better, so... like prices of SM or MM are equal => I installed SM fibers. No brainer.

 

In 2020, the hot threads are about modded switches / Buffalo switch / old Ciscos etc.

Some switches have SFP ports and people links 2 Meraki for instance with 2 SFPs & 1 fiber. They always choose MM fiber (the few I saw are only MM based). They go for MM because "why bother with a SM that can work over Kms when I need a 1m link ?" 🙂

 

Has anyone tested SM / MM recently & could give his feedback ? 

Thanks

BRgds

[Ultrarunner]

1 hour ago, tgb said:

Has anyone tested SM / MM recently & could give his feedback ? 

Thanks

BRgds

I tested this for the connection between my EtherRegen and opticalRendu. I tried both StarTech and Planet SM SFP modules. Both definitely sound better than MM. Planet better than StarTech. 

[kerisabe]

Just did a custom lpsu with swedish torodial and blackgate caps for my dual switch setup (c2960 48TC-L). Also did a fan delete on both 48TCs.

[tgb]

@Ultrarunner :

many thanks for your reply.

So, SM better MM, fine.

But SFPs sounds different too !? what a weird hobby we have...in fact, "weird" is not the proper word, "difficult" is more appropriate 🙂 

 

@ kerisabe :

nice work ! congrats !

In 2020, the audiophiles don't talk about speakers/amp/preamp etc anymore, only network & switches modds matters 🙂

So, as you shared I share 🙂

This is my modded TP-Link SG5412F, my basic 1Gb optical LAN is around it =>

https://ibb.co/yBWdf6f

This mod is not a "typical" mod based on LPS, I know. My point is that switches/FMCs use 3.3VDC max (3.3 / 1.8 / 1DC is this case, thus 3 external PS here), thus I use 5V power bank a main DC source (that's why you have USB-A plugs on the photo), and it works grreat.

I posted this mod, on this switch and on FMCs (TP-Link MC110CS & TrendNet TFC-1000MGA) on this French forum => http://forum-hifi.fr/thread-16709.html

This forum is a very active one regarding switch modds & others (see DIY part). Hope Google Translate will work if you need it.

BRgds

[plissken]

On 6/9/2020 at 2:36 AM, tgb said:

But SFPs sounds different too !? what a weird hobby we have...in fact, "weird" is not the proper word, "difficult" is more appropriate

 

I like to look at it from a critical thinking perspective. 

 

Fully realized 10G can ship 2 CD's worth of audio every second. My system can cue up a 16/44.1 track in about 300ms and then not use the connection any longer.

 

You are playing out of buffer, not off the wire. As a matter of fact, on well engineered systems, you can start playback and pull the plug. 

[The Computer Audiophile]

1 minute ago, plissken said:

 

I like to look at it from a critical thinking perspective. 

 

Fully realized 10G can ship 2 CD's worth of audio every second. My system can cue up a 16/44.1 track in about 300ms and then not use the connection any longer.

 

You are playing out of buffer, not off the wire. As a matter of fact, on well engineered systems, you can start playback and pull the plug. 

If I had to guess, I'd say their comments have nothing to do with the speed of the interface. Just a guess. 

[plissken]

Just now, The Computer Audiophile said:

If I had to guess, I'd say their comments have nothing to do with the speed of the interface. Just a guess. 

 

In a well engineered system, the faster the wire speed, the less the wire is used the less the transceiver is used. Hence talking about MM OM3/OM4/OM4/MPO 12 strand/LC BiDi or SM fiber or transceivers. 10/QFSP+/QFSP28/40/80/100G. 

 

One HAS to accept that we aren't playing off the wire. We are playing out of buffer. The quicker you can fill the buffer (regardless of size) the less the wire is of any issue. If I can get 5000MB/s wire rate (QFSP+) my wire time for a 75MB track is .015 of a second. It's also my tranceiver time. 

 

The only other argument that can be made is that somehow having a SM vs MM tranceiver installed, but playing back from local disk would still be affected, is one that could be effectively made. I'm all ears on that one. 

[ray-dude]

The correlation I have seen for SFPs seems to be more related to power consumption of the module.  I suspect the mechanism is some induced noise from the transceiver.  Generally speaking, the lower the power consumption, the better the sound.  

 

That being said, the different categories of SFPs (laser type, speed of interface, etc) seem to have different power profiles, so the correlation extends to those factors, but my spidey-sense (gut) says power consumption may be the key dimension for sound quality

[The Computer Audiophile]

1 minute ago, plissken said:

 

In a well engineered system, the faster the wire speed, the less the wire is used the less the transceiver is used. Hence talking about MM OM3/OM4/OM4/MPO 12 strand/LC BiDi or SM fiber or transceivers. 10/QFSP+/QFSP28/40/80/100G. 

 

One HAS to accept that we aren't playing off the wire. We are playing out of buffer. The quicker you can fill the buffer (regardless of size) the less the wire is of any issue. If I can get 5000MB/s wire rate (QFSP+) my wire time for a 75MB track is .015 of a second. It's also my tranceiver time. 

 

The only other argument that can be made is that somehow having a SM vs MM tranceiver installed, but playing back from local disk would still be affected, is one that could be effectively made. I'm all ears on that one. 

I hear you 100%. But, the faster the interface the less TIME it is used. I don’t know if faster interfaces have more noise due to higher speeds. No clue. 

[ray-dude]

5 minutes ago, The Computer Audiophile said:

I hear you 100%. But, the faster the interface the less TIME it is used. I don’t know if faster interfaces have more noise due to higher speeds. No clue. 

 

If the root cause is induced noise to power/ground plane, things are more insidious.  A faster interface would have faster switching components, which could be generating a different noise profile.  Depending on the sensitivity of the rest of the system, moving the noise could help or it could hurt SQ.  Other than the truism that less noise is better, noise related SQ impact is devilishly difficult to rationalize or connect the dots on (so system/environment dependent).

[barrows]

Spectral distribution of noise is going to make a difference on whether it matters, or not, maybe!  It is worth noting that higher noise frequencies are easier to filter out, but that higher noise frequencies also are more likely to go airborne...

 

With optical Ethernet I worry less about such things, as long as the noisy gear (commercial computer stuff) is away from the audio system in another part of the home.  Ethernet transformers should do a pretty good job with very high frequency noises (until they are airborne), but I am not a magnetics expert.

 

Then there is the latency issue, or not...speed equals less latency: some appear to believe that less latency in transmission equals better sound quality, although the reasoning behind this belief alludes me... 

[plissken]

7 minutes ago, ray-dude said:

The correlation I have seen for SFPs seems to be more related to power consumption of the module.  I suspect the mechanism is some induced noise from the transceiver.  Generally speaking, the lower the power consumption, the better the sound.

 

Power is consumed when the connection is in use. The faster the connection the less power is used. 

 

Also if we are talking power then you have an inherent problem with the SM vs MM position. 

[plissken]

27 minutes ago, The Computer Audiophile said:

I hear you 100%. But, the faster the interface the less TIME it is used. I don’t know if faster interfaces have more noise due to higher speeds. No clue. 

 

Faster interfaces typically have a clock that as speed increases clock increases away from our hearing threshold. 

 

Using my QSFP example: If we are only making 'noise' for 0.015 of a second for say a 9 minute track. That interval might as well zero.

 

The bottom line is that network connections are only using power when there is traffic flowing. These are all greened up from a power conservation standpoint. 

[plissken]

8 minutes ago, barrows said:

although the reasoning behind this belief alludes me... 

 

It's because it's not reasoning. It's something else entirely. 

[ray-dude]

I have not tested an MM transceivers yet.  Have you seen spec sheets where they are less than 500mW consumption? (best I've seen/heard for SM).  Alas, we're still in the empirical stages for these transceivers...try them and see if they sound better, then look at specs and try to figure out why things may be different.

 

Related to the network use comment, there is a definite SQ win in isolating my audio network from the rest of my home network.  Having my server on an isolated LAN (no broadcast or other traffic from my home network) definitely has a positive effect.  Back when I was fiddling with SqueezeLite endpoint buffers, there was also a huge win in increasing the input buffers so that tracks were preloaded into the buffer within a second or so.  These are all positive correlations (for me), but nothing conclusive to hang on yet.

 

 

 

[plissken]

24 minutes ago, ray-dude said:

 

If the root cause is induced noise to power/ground plane, things are more insidious.  A faster interface would have faster switching components, which could be generating a different noise profile.  Depending on the sensitivity of the rest of the system, moving the noise could help or it could hurt SQ.  Other than the truism that less noise is better, noise related SQ impact is devilishly difficult to rationalize or connect the dots on (so system/environment dependent).

 

With optical configurations there is no ground plane. It's optical. Faster is measured in Hz. The faster the higher the Hz the farther away from human hearing. 

 

Agreed the less noise the better: Faster connections spend less time making noise. 

 

I have a dual SFP+ Solar Flare NIC that I can 1 or 10G connections into. One link could be SM and the other MM. I'm keen on anyone being able to pick which is which even if I dialed back the buffering to 10 seconds. 

[plissken]

14 minutes ago, ray-dude said:

I have not tested an MM transceivers yet.  Have you seen spec sheets where they are less than 500mW consumption? (best I've seen/heard for SM).  Alas, we're still in the empirical stages for these transceivers...try them and see if they sound better, then look at specs and try to figure out why things may be different.

 

MM tranceivers have less Tx power than SM. SM is for long haul 10/40/80Km runs. At 40/80Km Optics I would have to use attenuators if going simply cross rack because the modules would burn out. 

 

Use MM for home environments. The spec is 980 feet. 

[ray-dude]

SFP+ is definitely on my "try it" list, but it is a big commitment to move to a 10G network end to end.  I strongly suspect there is a similar speed vs power/noise tradeoff with networks as a bunch of us have seen with CPU - faster you drive the system, the better, until you hit a point where it becomes worse.  If one puts in better power, that peak in SQ moves to higher performance.  Fortunately, it is a lot easier to power a NIC/switch/router than it is to power a server.

 

I do use MM to connect my two EdgeRouters (1m cable).  I'll pick up a longer MM fiber and give it a try vs my SM Finisars and Planet Techs.

 

 

[The Computer Audiophile]

1 hour ago, plissken said:

The bottom line is that network connections are only using power when there is traffic flowing. These are all greened up from a power conservation standpoint. 

I’d say they are only using more power when traffic is flowing. If there was no power use the switch would warn me that the device has dropped off the network. It may be a trickle compared to full blast at all times, but it has to be something. 

[plissken]

Just now, The Computer Audiophile said:

I’d say they are only using more power when traffic is flowing. If there was no power use the switch would warn me that the device has dropped off the network. It may be a trickle compared to full blast at all times, but it has to be something. 

 

Fair enough. What happens is the certain portions of the PHY will power down. Think of it like Magic Packet/WoL. 

[The Computer Audiophile]

1 minute ago, plissken said:

 

Fair enough. What happens is the certain portions of the PHY will power down. Think of it like Magic Packet/WoL. 

Very cool stuff.