Optical Network Configurations

[dbastin]

5 hours ago, jabbr said:

Future of networked SSDs

https://www.servethehome.com/ethernet-ssds-hands-on-with-the-kioxia-em6-nvmeof-ssd/

 

I wonder if this approach will have benefits for NAS for the consumer market.

[dbastin]

18 hours ago, Bertel said:

Thanks - my question is with which product. Which product provides 10Gbps SFP on one side and WiFi on the other as kind of a bridge?

I searched for a 10Gbe WiFi Access Point (WAP) but gave up.  My interest was not for speed, but better specs than 1Gbpe.  I did consider getting a SFP+ router with WiFi (eg. gaming ones) and setting it up as WAP only.  Still an option.  There is also 60Gbps wireless wire ... https://mikrotik.com/product/wireless_wire

 

Currently I use Mikrotik mAP lite with USB Power Bank and Akiko USB stick.  It is at the end of a Shunyata Sigma.   The mAP lite provides network connection to the endpoint, Devialet Pro.

 

In other words, I have minimised noise getting into the WAP, and try to defeat what little noise is left.  I could do the same with a gaming router as WAP, but it might not really be an improvement.given those are built for speed and may generate more noise and jitter than mAP lite which is 100Mbs.  However a SFP+ router would enable me to connect using fibre from the Mikrotik CRS 305 (Rouer OS) and guarantee galvanic isolation, and sell the Sigma to gain savings.

 

For now the mAP lite is producing a result that is very pleasing, so not really great incentive to change that except curiosity.

[dbastin]

9 hours ago, Jud said:

 

I know others have said "Well this isn't what you asked, but I have a different solution," and though I hate to be one more of those, here is my thought:

 

Fiber from Mikrotik to Fitlet3 with optional optical Ethernet. Get Fitlet with Linux Mint OS and install NAA and upmpdcli so you can use your choice of player software.

 

If this is more technical than you want to get, my apologies for the unproductive suggestion.

Not sure if this was intended for me or @Bertel.  I have a player in the Devialet Pro.  Wifi to the Devialet is exceptional and much much better than the last wired system I tried with it (Gigafoils and EtherRegens in series), although I will revisit that because I may not have optimised isolation of power supplies and cable shields.

 

9 hours ago, MartinT said:

Personally I would not introduce wi-fi into the link. It is very jitter prone. An optical link should give you lower overall noise (certainly lower jitter and phase noise). Just my opinion.

See above.  Until I tried Auralic Aries G1 and Devialet over wifi, I was avoiding wifi too.  It seems to me, even if wifi has shortcomings, it has way less problems for sound quality than wired connection to those two endpoints (which don't have fibre input).  Further upstream where users have more options, wifi may be more problematic than the other options.

[dbastin]

5 hours ago, Bertel said:

 

Let my try to clarify and visualise what I'm trying to achieve - please bear with me 😉

 

Current setup

My music server (PC with HQP), my NAS, my Signature Rendu SE all connect to the MikroTik switch with optical fiber. Vodafone is my ISP, the router they provide manages all the IP addresses, DHCP, DNS etc., also telephony. I don't stream, I don't use  Roon, but the network devices need the network management services of the router, the Rendu needs internet access for SW updates.
So when playing music, an audio file is taken from the NAS, played on the music server with HQPlayer, and sent to the SIgnature Rendu(and from there over USB to the DAC, either Holo May KTE or VMV D1se at the moment)

The bit I'm focusing on here in this topic is the upper right corner, i.e. the connection between the MikroTik switch and the internet router. Everything in the music path from NAS to the Rendu is optical fiber already, so 'light' not 'wire' in my naive view, galvanically and mechanically isolating the components, a thought which I like a lot. 
The only wire still in place is the one to the 'dirty internet' and the router with all its non-audio activities, and I want to get rid of that wire. Not because it was a technical issue of any sort, just because I can (hopefully) ;-) It shall be replaced by a wireless (Wifi) connection.

 

Alternative 1 - MikroTik hAP ac

The MikroTik hAP ac is a router with 5 10/100/1000 Ethernet ports and 1 SFP port as well as dual band Wifi (4 and 5). According to MikroTik support, it is possible to bridge SFP and Wifi tigether, making it act like a switch, and they confirm that it'll work for my use case. I'm not sure yet though, as I see two potential hurdles:

• Connecting with my MikroTik CRS305-1G-4S+ switch via optical should be trivial, but the Wifi side for connecting with the internet router could be tricky. The hAP ac is designed to provide its own network as an extension of its Ethernet and SFP, not connect into an existing one. Yet MikroTik support says it can be done.
• The configuration is done in SwitchOS. Which I have on my CRS305-1G-4S+ as well, but it's hardly documented, so probably a lot of research on the MikroTik forum.

 

Alternative 2 - Fritz!Box 5530 Fiber

This is AVM's wifi router with WFI 6, 2 Gigabit Ethernet ports, 1 2.5 Gigabit Ethernet port, and 1 SFP port. While Wifi 6 is great and integration with the existing Wifi that the Fritz Box 6660 Cable as my internet router provides will be a breeze, using its Mesh functionality (I already have a couple of Fritz Repeaters in the network, thy're truly plug and play and work reliably 24/7), I have no clue yet how to get the optical side connected. Three standards are supported - AON, GPON and XGS-PON (the latter only for models provided by ISVs if I understand correctly). I have no clue what these standards mean for practical connectivity in my use case, will need to figure that out now. Here are the details for reference:

 

FRITZ!SFP AON
ITU-T G.652; IEEE 802.3ah-2004 1000BASE-BX10
LC/APC 8°
Wavelength: TX 1310 nm, RX 1480 to 1580 nm
Full duplex transmission
Transmitter power: -9 to -3 dBm
Reception range: -3 to -23 dBm
Range: 10 km
Support for SFF-8472
Class 1 laser

 

FRITZ!SFP GPON
Suitable for ITU-T G.984.2/984.5 (GPON)
LC/APC 8°
Wavelength: TX 1310 nm, RX 1490 nm (with filter)
Suitable for rogue detection
Transmitter power: 0.5 to 5 dBm
Reception range: -3 to -28 dBm
Range: 20 km
Support for SFF-8472
Class 1 laser

 

FRITZ!SFP XGS-PON
ITU-T G.9807 (10-Gigabit-capable symmetric passive optical network)
SC-UPC
Wavelength: TX 1270 nm, RX 1577 nm
Full duplex transmission
Transmitter power: +5 to +9 dBm
Reception range: -8 to -29 dBm
Range: 20 km
Support for SFF-8472
Class 1 laser

 

 

You still with me? Good, thank you 😉 I have ordered both devices, the MikroTik hAP ac and the Fritz!Box 5530 Fiber. If there's interest I'll be happy to let you know how it goes. 

I actually think your current set up is very good even for streaming from the internet, provided the Cat 6a is UTP.  It is similar to mine from the Fritz box to CRS 305, however I use the CRS 305 as router.  I am curious if strapping an Akiko tuning stick to the Cat 6a right where it connects to ER would kill off more noise than ER alone.  It is a very worthwhile approach with speaker cables right at the speaker binding posts.

 

But getting to your proposal, my first question is can you hear a difference if you swap the Cat 6a for a cable you know is terrible?  I suspect you can, but if not why not just leave it as is?

 

Or you could swap the ER with a FMC, or a low cost switch with SFP (CRS 106). Or connect a WAP to CES 305 via its single RJ45.

5 hours ago, audiophilac said:

I’d rather you use eR nearest to your streamer (signature Rendu)

I agree, that is if it is an improvement given the rendu's ethernet is probably a similar design to ER (except the moat).

[dbastin]

11 hours ago, Bertel said:

 

No, I have never listened to / compared Ethernet cables. My goal is to get rid of them, and of wired (copper) connections altogether 😉

Maybe you need to review why you have that goal.

 

If there is no difference between different wired cables in the connecting the router to the  CRS 305, going to all the effort of eliminating the copper might not be beneficial, and could actually be detrimental by adding another device into your network that influences noise and phase noise.

 

You might need to choose the lesser of 2 evils:  excellent performing copper wire -v- an extra device.

 

A simple alternative is to only connect the CRS 305 to the router when you need to.

 

I think that is what @R1200CL means by "Why don't you just pull a RJ45 from router to Microtik ?"

 

[dbastin]

5 hours ago, robi20064 said:

What does copper "add back" before it would be converted to fiber once again - right before the Rendu SE?

I gather what is suggested is:

 

CRS 305 > fibre > ER side A - Side B > copper > Rendu.

and

CRS 305 > fibre > Optical Module > copper > Rendu.

or

CRS 305 > fibre > ER side A - Side B > copper > Optical Module > fibre > Rendu.

or 

CRS 305 > fibre > Optical Module > copper > ER side A - Side B > copper > Rendu.

or 

CRS 305 > fibre > Optical Module > copper > ER Side B - Side A > fibre > Rendu.

 

So many possibilities, permutations and variables.

 

But it puts a better clock and design right before the Rendu, which could be a superior outcome compared to just fibre from CRS 305.  In other words, adding potential dis-benefits of copper is outweighed by the benefits.  This also leaves the only RJ45 of the CRS 305 to be used to connect to the router.

 

There is no ideal, every solution has 'cons', just some are better outcomes in any given situation.

 

7 hours ago, Bertel said:

Wait - convert fiber to copper Ethernet, and then back to fiber…? No way 😆

You could be assuming fibre is always best, which might not be true.

[dbastin]

59 minutes ago, R1200CL said:

The Rendu is fiber in. 

Oh, maybe there are different versions.  I read ...The Signature Rendu SE takes fiber optic Ethernet input or cable ethernet input here ... https://www.sonore.us/Signature-Rendu-SE.html

[dbastin]

23 hours ago, MartinT said:

It's an AOC

FYI, Afterdark has been offering these for some time ... 

 

https://www.adark.co/en-au/collections/project-constellation-fiber/products/copy-of-afterdark-project-clayx-constellation-sfp-finisar-active-optical-cable-aoc

 

I recall he can put carbon fibre sheath on the fibre to improve it further.

[dbastin]

10 hours ago, jabbr said:

Ok you realize that it is absurd to think that putting a carbon-fiber sheath on an optical fiber could improve it.

 

The modifications do not improve or detract from the Finisar products nor the underlying optical fiber that Finisar uses in their products.

Yeah, it sounded absurd when I read the website ... Since optical fiber are so sensitive to vibration that sensor on Finisar SFP+ module, vibration countermeasures are most effective to reduce jitter and improve sound quality. 

 

But hey!  Using ethernet for hi end audio playback people have found all sorts of absurd things to provide an improvement for them.

 

10 hours ago, jabbr said:

I would NOT say that Finisar AOC cables are "better" than, say, the same Finisar SFP modules connected by Corning fiberoptic cables.

AfterDark's Adrian seems to think they sound different to each other.  He also thinks 3m corning fibre sounds better than 1m.  And that using a high quality external master clock helps too which many other agree with.

 

Overall, the differences in using different approaches to fibre may make most difference when connecting to the endpoint and less difference further upstream.

 

For instance, the carbon fibre might just reduce microphony type thing being conveyed by the fibre cable from affecting the endpoint and/or its SFP module etc, rather than improving the performance of data transmission by the fibre.

 

7 hours ago, Nicholas_S said:

Just tested with a multimeter and there is no continuity between the transceiver shells of the AOC cable.

Phew, can anyone confirm this, so this gent doesn't get accused of perceiving what the multimeter says?

[dbastin]

2 hours ago, Superdad said:

Yes, that absolutely is the case—just as with most all SFP transceivers.

The AOC cable/transceiver assembly does NOT have any end to end wire/ground connection.

Presumably it is nothing like the new gen optical HDMI ...

[dbastin]

13 hours ago, juniorbudel said:

Thanks everyone for the reports.  
 

My set is completely isolated from the “dirty” side, in the listening room the fiber arrives at a Sotm switch, which feeds an Ultrarendu -> Matrix SPdif -> Direcstream DAC -> pre/power Electrocompaniet.  The sound for my intentions is spectacular, I haven't changed anything on the set for more than a year and I'm thinking about the possibilities without incurring huge costs.  

 

On the dirty side, I have the Roon Core (associated with HQPLAYER) which is actually an All In One HP that I use on a daily basis, I have no intention in the short and medium term of turning this server into a dedicated machine, much less mounting something fanless.  

 

In addition to the server, I have a router from the operator (any brand) and from this router the signal is sent to another Ubiquiti router that sends the fiber to the listening room.  

 

I was suggested a DC Filter Kit by FX Audio - Petit Susie and Petit Tank, could you try them on the Ubiquiti router and on the server, does anyone know?  Recommend?

If you change a cable or power supply on your dirty side, can you hear any difference?

[dbastin]

9 hours ago, The Computer Audiophile said:

Makes me wanna try 100Gbe

Me too!

I mentioned previously I am using Mikrotik CRS 305 (10GBe) as my router and it is easily better than the Ubiquiti EdgeRouter X SFP (1GBe) it replaced.  But ... when I installed a EtherRegen as FMC instead of a generic FMC immediately upstream of the CRS 305 that provided a further improvement.  So, it suggests perhaps even the CRS's 10GBe low jitter does not create immunity to upstream changes.

 

I am also curious if a low jitter NIC can be so beneficial that the performance of the server (in terms of noise etc) is not relevant.

 

My setup in simple terms is ...

 

Broadband service > EtherRegen > fibre > CRS 305 > Server (Antipodes) > WAP > Devialet

 

Maybe there should be a new thread about the benefits of 5Gbe, 10GBe, 100GBe and how best to gain the most benefits from these.

[dbastin]

8 hours ago, juniorbudel said:

I still haven't made an AB, today I have a Wireworld Chroma 7 between the server and the router.  I will do this test.  

 

The problem is that all the other network cables on the dirty side are common, they will certainly narrow the differences, if any.

How about compare UTP cables with STP cables.

[dbastin]

21 hours ago, juniorbudel said:

I hadn't thought of that possibility, has anyone reported differences between shielded or unshielded network cable on the “dirty” side of the system?

I haven't 'reported' this on AS before, and you might not believe it ...

 

My network is ...

 

broadband service > cable 1 > ER Side b - side A > fibre > Mikrotik CRS 305 (router) > cable 2 > Antipodes server > cable 3 > WAP > wifi > Devialet

 

Nothing else is connected to the ER, it is basically a FMC.

 

Cables 2 and 3 are audiophile circa $1k each, super reveiling.  CRS 305 is 10Gbe SFP+.

 

For cable 1 I tried generic Cat 6 stp, Cat 8 stp, Afterdark (optimised for ER - ie sheild disconnected at one end), Wireworld Platinum (cat 8 sheild connected at both ends), commercial grade Cat 6a STP, but felt a commercial grade Cat 6a UTP is most accurate.  I could hear differences between all these cable despite the fibre, there's moat,  jitter rejection of the CRS 305 and wifi.  The differences are small that far removed from the endpoint but still enough to distinguish and judge.

 

But it probably depends what the cable is connecting, so worth experimenting.

 

I will revisit when I get around to improving the SFP transceivers.  Maybe of I hack the Cat 6a STP to disconnect the sheild at the downstream end that will be best in this location in my network.

[dbastin]

16 hours ago, audiophilac said:

@dbastin is your devialet connected wirelessly to the network?

Yes.  I haven't tried wired for quite some time.  Wifi sounded great with a regular WAP (tp link), better with LPS but I now use a dedicated WAP (Mikrotik mAP lite) powered by a USB battery, with Akiko tuning stick, and each of these made a small improvement.

 

Go figure?!

16 hours ago, barrows said:

of course advocates for using hardware capable of even higher speeds, as that hardware is built to higher standards, but of course one can always use that 10 Gb hardware at 1 Gb, etc...

I run my CRS 305 10Gbe st 100MB.  But I think that may be causing more interruptions when TV and music streaming data download is patchy.

 

My broad band delivers 20G peak ... if the data is there.

[dbastin]

6 hours ago, jabbr said:

Get these while they are still available ... Intel 100G-CWDM4

 

100G at <3.5w -- low energy/bit! That's why the pulls are green 😝

I am curious how 100Gbe SFP would be an improvement in SQ compared to 10Gbe.  Didnt someone say they could not detect a difference in SQ between 10 and 100 Gbe switches?

 

I am also wondering if a 10Gbe SFP is better than 1Gbe SFP for the same kinds of reasons why a 10Gbe switch is better than 1Gbe switch.

 

I am not seeking scientific proof, but are we speculating or making a well informed assumption?

[dbastin]

9 hours ago, SQFIRST said:

 

@dbastin - just came across your post and the details and wanted to share from my recent experience in case it helps:

A single CRS305, connected with your ER SFP(1000base-x), may not be benefiting from the 10gbe protocol as it is most likely using the 1000base-x circuit. Adding a second 10gbe fmc/switch, such as another CRS305, to connect with your current CRS305 Router will then be making a true 10gbe connection and should show effective jitter isolation. You might even find a better position your ER once you see the results.

 

Thanks,

 

I gathered SFP+/10Gbe spec is the big benefit, not the actual speed ... that is, the spec is the benefit even if the data rate is less than 10Gbe (mine is set to 100MB).  Perhaps I am not correct.

 

I can't quite remember how I arrived at my current set up. But for what it's worth in case it prompts someone to suggest something or benefits/inspires someone ...

 

In simple terms, the wifi performed better than 2 x ERs and a Gigafoil (although I since realised some of the cabling between these had shields connected at both ends).

 

Nevertheless, I was distracted/occupied optimising the WAP and learning how to set up CRS 305 as router, I then swapped Ubiquiti Edgerouter X SFP with CRS 305 and realised a good benefit from doing that.

 

At the time I was using a generic FMC (fs.com) and swapped that with 1 ER and realised another good benefit.  Then I was experimenting with different cables into that ER.

 

This became somewhat of a distraction that is very satisfying. 

 

I am intending to insert a 2nd ER ...

 

Broad band service > ER1 > fibre > CRS 305 > fibre > ER2 > server > WAP --- Devialet

 

... with a view to using an external clock on ER2, and mindful that ER1 is fantastic as FMC.

 

But perhaps this would be better ...

 

Broad band service > CRS 305 > fibre > ER1 > .. etc

 

... but this doesn't totally isolate the CRS 305 router with fibre in and fibre out.

 

My guess is the following would be inferior to the above, so I have tried it ...

 

Broad band service > CRS 305 > fibre > CRS 305 > Server > WAP > wifi > --- Devialet

 

Incidentally,,I run fibre from the CRS to the rest of the household network to isolate it from the CRS and hifi.

 

I am happy to consider feedback.

[dbastin]

Given there has been discussion here about SFP+, you may be interested to hear Taiko's new switch can be SFP+.

https://taikoaudio.com/taiko-2020/product/taiko-audio-extreme-switch/

 

They also have a PCIe network adapter/switch, and a router in the works.

 

Both are probably only called a 'switch' because (I presume)  they are fundamentally a switch (even if only 1 port in and 1 port out).

 

I have been wondering when someone was gonna 'hot rod' SFP+ with audiophile performance parts.  But I was hoping for a much lower price tag.

 

Given this comment ...

16 hours ago, Superdad said:

the ultra-low-jitter 10GHz-capable differential flip-flops we reclock with on both sides

... I am curious how EtherRegen would measure against SFP+ specs (is it stressed eye etc? ... I am not expecting a firm answer, just curious).

 

I also note people are preferring SFP+ modules when used with EtherRegen (eg. Afterdark).

 

So back to topic, I suppose EtherRegen might have even lower jitter, noise, etc than SFP+ so using SFP+ modules are complimentry icing even if the output of Side B is 100MB.

 

Hence, using EtherRegen as a FMC with fibre > SFP+ module > ER - moat - Side B (or vise versa) I presume has compounding benefits  (??).

[dbastin]

1 hour ago, Superdad said:

 

Only the EtherREGEN's 'B' side RJ45 is 100Mbps-only.  

The four 'A' side RJ45s will negotiate 10/100/Gigabit.

And the 'A' side's SFP cage is Gigabit only.

 

Remember, optical SFP and SFP+ transceivers are relatively "dumb" devices. No clocks and no "speed setting" themselves.

 

A reminder about how it works:

Optical does not do speed autonegotiation. What happens is here is a pin on the SFP and SFP+ modules that is used by the switch port to determine the speed. An SFP module grounds the pin. An SFP+ leaves it floating. A gig only port on the switch grounds the pin. A 10 gig port senses the pin. The result is that if you put a SFP module into and SFP port it runs gig. If you put an SFP+ module into the SFP port the port runs at gig. If you put a SFP module into an SFP+ port, the pin is grounded which runs the port at gig. If you put a SFP+ module into a SFP+ port the port runs at 10 gig.
 
The upshot is the SFP+ on both ends, the ports must also be SFP. If one port is SFP and the other SFP+ one end will run at gig and the other at 10 gig and no communication is done. So if you wave a switch with 10G port and ER (which is 1G) you must use SFP modules on both ends.
 
So it is not autonegotiation as we are used to in copper Ehernet, You have to understand how the Switch port speed is determined.
 

Excellent post, thank you.

 

If you could clarify this ... with a SFP module in a SFP+ port in a switch configured to run at 1G (or less), a SFP+ module at the other end in ER will be detected as SFP+ and therefore will not work? (ie. because the hardware pins dictate).

 

And for anyone who might know ...

 

is there really any or much difference in SQ between a SFP and SFP+ module of equivalent quality (eg. say both the best Finisar Industrial grade).

 

Does SFP+ modules have higher/better spec in terms of jitter and the like?

 

Will SFP+ module run cooler at 100MB or 1 G than 10G?  In other words, is the heat generated commensurate with the speed it is actually running at rather than its max capability?

 

These questions spring from me presuming there is a balance between 2 competing issues (both general assumptions, may be incorrect)....

 

1. high speed = high heat = damage to SQ

 

And

 

2. SFP+ module = lower jitter etc = improved SQ

 

I also imagine this is less relevant (impact) the further from the endpoint in the network the modules are. For example, way upstream of a few switches/links is a lot less impact than connection to an endpoint with fibre.

 

Cheers

[dbastin]

32 minutes ago, R1200CL said:

Jabbr claimes it must be so. This due to the eye pattern requirement and standards.

It  seems Jabbr addressed this on page 112 that implied (to me) the spec /standard might not fully apply to SFP+ modules.

 

"To be clear, there is a difference between an SFP(+) module and 10GBase-X (10 GbE). Many 1 GbE devices accept SFP(+) modules and run them at 1 GbE. In the very beginning of this thread I discussed using SFP(+) modules with 1 GbE equipment.

 

The "stressed eye pattern" isn't part of the SFP(+) specification, rather the 10GBase-X specification (802.3ae) see: IEEE 802.3 - Wikipedia"

 

So, to somewhat set aside the influence of the switch, if one is using a 10G compliant switch (or ER), is it possible or likely using a SFP+ module will provide better SQ than equivalent quality SFP (even running 1G or less)?

[dbastin]

5 minutes ago, R1200CL said:

You obviously didn’t read all links I provided. The answer is there. 

Um ... I was reading it while you posted this :) For those who don't read it, John S says:

 

is it worth using SFP+ with ER anyway? In many cases YES. As has been mentioned the circuitry and clocks in a SFP+ port do have significantly lower jitter than what is common for SFP ports. The SFP+ modules also have lower jitter, even though they do not do any clocking the conversion from electrical to optical and the other direction DOES add jitter, with SFP+ that added jitter is significantly less. So IF you have a switch with SFP+ ports that lets you set the port at 1Gb and you are willing to program it, going with SFP+ modules in both the switch and the ER will most likely give you lower jitter. If you do this any SFP+ modules can be used.

[dbastin]

3 hours ago, R1200CL said:

Instead of buying a new 10Gb switch, one could request Sonore and Uptone to make 10Gb products 😂

But ERs and OMs are by design (presumably) better than 10G switches that are available.  You gotta wonder if a 10G device designed specifically for audio would be much (if any) better.

 

I would think the starting point would be where audiophile switches started ... modifying and hacks to existing products to explore the improvements to be gained.  Perhaps someone like Renolabs will.

 

1 hour ago, Johnnydev said:

What do you think of these experiences and theories?

I don't think any one approach is generally best. Overall, my assumption is none is ideal for audio, we are having to determine the least damaging option/s.

 

Also, of course, have clear criteria for determining what is 'better'.  I may be hard to decide between what is accurate and what has a preferable flavour ... after all, what really is accurate? (No need to answer).  You may well know, I don't follow this thread closely enough to know your judgement, so please accept I am  not being critical.

 

My feeling is, explore each link in your own network and use the technology that gives the best cost benefit in each link. For example, Shunyata Omega might sound better than fibre, but over 15m not likely cost effective.  Then review the sound of the whole lot and tweak/adjust.

 

My network that leads to my hifi has fibre, wired ethernet and wifi to the endpoint (Devialet) via a low power WAP. The rest of the network is isolated by fibre > FMC > Wifi.  The ISP comes to the premises via proprietary fibre optic and termination device.

 

I would also say if your router is not SFP+/10G, then that is masking what could be achieved.  I have mentioned before my CRS 305 is my router running RouterOS.  It seems more accurate to me.  I wonder if RouterOS can force a SFP+ port to be 1G.  I have set my CRS 305 ports to run at 100MB, but that may actually not be effective as per the experience with SwitchOS noted earlier.

1 hour ago, Johnnydev said:

No, with or without the streamer, results where the same. Better SQ without the OM

What streamer? Maybe it is doing something  extraordinary?

[dbastin]

11 minutes ago, audiobomber said:

I suspect you've found an ethernet cable, or mix of cables, that are euphonic to your taste. What ethernet cable(s) are you using? How long is the longest run?

That is what I was thinking is a possibility but didnt want to offend.  But, we should not overlook that SFPs may reduce the neutrality of fibre.

[dbastin]

22 hours ago, jabbr said:

but part of switching a 10g port to 1g means using a dual rate 10g/1g module.

Firstly, I am not finding the compatibility table to help.

https://wiki.mikrotik.com/wiki/MikroTik_wired_interface_compatibility#10G_SFP+/25G_SFP28

 

In relation to SFP+ it doesn't refer to other manufacturers but refers to https://i.mt.lv/cdn/product_files/splitter_modules_180841.pdf

 

That document refers to (for example)  SFP+ CWDM module 10G SM 10km 1530nm Dual LC-connector DDM. 

 

Does the Dual mean dual rate?or Dual LC UPC?

 

The SFP+ modules listed in the document may all be discontinued.  Current SFP+

Modules = https://mikrotik.com/product/Splus85DLC03D (Units are tested and compatible with CCR1036-8G-2S+ and CCR1036-8G-2S+EM. Units are compatible with non-MikroTik SFP devices as well.)

 

AOC = https://mikrotik.com/product/s_ao0005

 

Note, both are dual rate 1G/10G, but no way of knowing which Mikrotik switch or router they are compatible with except the two mentioned.  And no way of knowing (from this info by Mikrotik) which non Mikrotik SFP+ modules are compatible with which Mikrotik switches and routers.

 

Am I correct assuming that not just ant dual rare 1G/10G module will work in any Mikrotik switch or router?

 

Is a Finisar or Cisco dual rate module likely to be better (spec, SQ) than Mikrotik modules?

 

For comparisin, from memory Ubiquiti community provides alot more more info about SFP compatibility.  Their range of SFP+ products seems more limited though ...

 

https://store.ui.com/collections/unifi-network-switching/products/switch-enterprise-8-poe

 

https://store.ui.com/collections/unifi-network-switching/products/switch-enterprise-8-poe

 

https://store.ui.com/collections/unifi-network-switching/products/us-xg-6poe

 

https://store.ui.com/collections/unifi-network-switching/products/unifi-switch-16-xg

 

And their DC voltage inputs are not terribly compatible with what psu we'd like to use.

[dbastin]

23 minutes ago, lasker98 said:

If someone is able to tell me how to get them working in my switch at 1G it would be much appreciated.

 

I am liasing with my local Mikrotik support agent to hopefully shed light on this issue regarding running 10G SFP+'s at 1G speed.

 

He suggested ... untick the 10G 'advertisement' ..  perhaps you know what this means in case you haven't tried it (I havent).