[c-nsp] Catalyst SX GBIC's and SMF

[Bill Wichers]

> * zardoz at hotblack.net (Tristan Gulyas) [Mon 10 Jan 2005, 11:21 CET]:
>> I have been told that running 1000baseLX with MM fibre + mode
>> conditioning patch cables gives you "slightly better" range than using
>> native SX, but there's no solid evidence for this, so I'd stick to the
>> 550 metre spec as per 1000baseSX to be safe.
>
> That's with 50/125 MM fibre but most I've seen is 62.5/125, which goes
> considerably less far.  With that kind of fibres LX is a definite win.

LX is 1310nm, which has considerably less problem with time smear than the
850nm wavelength used for SX, which allows proper data recovery over
longer distances. The difference is most noticeable in the 62.5 multimode
fiber where the MHz*km rating is usually lower. I have in the past used LX
GBICs to "cheat" and avoid the need to replace 62.5/125 multimode fiber in
some applications where the cost would be high (usually outside plant
cables for small office/educational complexes).

You should be safe using LX optics over multimode fiber. Mode conditioning
patch cords would be advisable, but are not always necassary.
Hewlett-Packard actually specs their gigabit LX Procurve GBICs for reach
over both singlemode *and* multimode cables, for example. In the HP
example, using their LX GBIC with 62.5/125 fiber gets you the same (550m)
reach that you can get with the best 50/125 fiber (500MHz*km) using their
SX GBIC.

     -Bill

*****************************
Waveform Technology
UNIX Systems Administrator