Yes, SFPs do require break-in. It's a miniature circuit board, with traces, surface mount resistors, capacitors, and more.
Cheers, good to know.
Not sure if the cables you bought came with this, but it's pretty cool how my three ~$10 fibre patch cables came with measurements of
return loss and
insertion loss. And credit to
@OCC7N it seems there is a measurably better - if very slight - average return loss for 3m vs shorter lengths, although this is a mostly obscured by the cable to cable and cable-end to cable-end variation.
Here’s the average return loss across the 4 terminations of each cable pair:
Return loss - for those who don't know - is a measure of internal reflectivity, the difference in dB between the strength of the signal (incident) wave to the wave that is reflected back when the incident wave hits the end of the cable assembly (reflected wave). You want it as high as possible so the bounces/reflections from the teminations are negligible relative to the signal. This is a general rule for RF and optical comms/data, not audio specific. Insertion loss is negligible at these lengths, of less interest than return loss.
If I recall correctly when I looked at Minicircuits terminated RF cable range for use with master clocks a year back a mid-late 40s dB return loss at 10MHz was top end performance for BNC cables with >50dB only achieved by the very best - and expensive - SMA terminated test cables for lab use. So >50dB here (albeit at THz range for optical rather than 10MHz for clock signals) seems a pretty solid minimum beyond which any audible delta may be negligible. Not least as we're talking clocking of asynchronous network packet data rather than a master clock controlling a DAC or streamer's synchronous digital audio data timing. Though I guess with one directional RF clock signals one can double the measured RL as it takes a bounce at each end before the reflected wave heads back to hit the clock receiver, whereas network data is bi-directional so reflected waves impact at both ends so effective RL is the value measured at one end?
I admit I read the comments upthread re longer lengths being/sounding better with scepticism, I was thinking optical transmission, tiny wavelength of light relative to cable length etc - and that any differences in transmission or reflectivity would be negligible at lengths less than say 100s of metres or longer. And I’d expect the audibility of the tiny <1dB variance across my three cable pairs on top of a base of an excellent >50dB return loss would be subtle at best. But maybe the difference widens with longer lengths and/or different cable manufacturers/assemblers.
Contrast this with quoted return loss of >30/35dB for multi-mode cables, a big 15-20dB delta. No wonder the recommendation for single-mode in a hifi context. Maybe this has been discussed upthread - if so apologies.
I'd be interested to see how a 10m, 20m or 30m single mode cable would measure and perform. Or even 100m. Go on someone, take one for the team.
. UIP is a couple of weeks old. 1000 hours is 6 weeks of 24/7.
