Fibre infrastructure and CWDM developments

The most interesting things I saw at IBC last week were not software updates or new 4k workflow tools (save us from "workflow tools"!) but some developments from our good friends at Barnfind. 

Their current CWDM products top-out at 18-wavelengths on a single fibre, and if you need to pack more signals (synchronous broadcast; HD/SDi, MADI, AES etc OR asynchronous data; ethernet, fibre-channel etc) onto a fibre then DWDM is the answer - with a big price tag!

• They showed me a prototype of their "pre-mux" product which can take six wavelengths and multiplex them onto a single SFP-wavelength; essentially it reduces the channel spacing down to 1.6nm but the upshot is that it is entirely compatible with the existing product range. You can use your current SFPs, Optical de/multiplex and single-mode cabling. You wind up with 106 usable channels on a single fibre.
• They also showed me Embryonix's ST2022 video encoder/decoder pair in SFP form; this allows you to use Barnfind as your video router in and out of an IP environment - I was blown away; a complete 2022-encoder with 10gig fibre i/o from a 3G source. 
• Barnfind are also taking their router up to 12Gig for single-wire UHD-TV routing; this will allow much greater penetration into the 4K market and although I never thought routing ten gig ethernet was a good idea (you don't want your edit assistant assigning the backbone network traffic!) it will be important for both baseband and IP 4k.

Whilst on the subject of Barnfind I have a 1350nm SFP which when in a demo chassis caused an intermittent yet repeatable fault; routing 1.5G video (1920x1080 @4:2:2) through it would cause CRC errors in only the AES packets; the signal was then sent to a BT facility line and the first MPEG-encoder it hit at the tower showed the fault by both video and audio disturbance every few seconds; I assume that is down to how that model of encoder extracted timing information. Swapping the SFP out for a new one (exactly the same model) drove the fault away but putting the existing one back re-introduced the trouble.

I am flummoxed how a format converter can reach into the SDi stream and corrupt the AES CRCs - it doesn't seem possible and so I was eager to get it back to the workshop to test it and get a bit more info. So - I have replicated the configuration as much as is possible - even down to the 1000m fibre drum we keep to hand (it's 250m of four core with the ends spliced-back on each other). 

However - the thing has been on 24-7 since last week with my trusty Tek WFM7120 monitoring it waiting for any errors to show and there has been nothing! Do I trust this SFP now? I suppose if I keep it in the demo kit it will be a good test.