Corrupted data from SFPs in a single mode system

Sounds like coherence fluctuation noise due to the back-reflections. Lasers generally have much worse phase noise than amplitude noise. Multiple paths with different delays turn the phase noise into amplitude noise, which can easily be 40 dB worse than the laser's intrinsic amplitude noise.

In long-haul fibre optic systems it's usually caused by double Rayleigh scattering, but discrete back-reflections do it too.

Cheers,

Phil Hobbs

Thanks for the response Phil. Can you tidy up some loose ends for me, and make sure I've understood what you've said

So dirt and grease and so on can cause significant enough back reflections to cause this kind of thing (I understand that you're saying that although the amplitude of the reflections may seem harmlessly small, they can phase problem which the real nasty)?

Am I correct to think that even so, most of the signal loss will be due to absorption of the signal into the dirt, and into the buffer where it is reflected in silly directions?

Incidentally, is double Rayleigh scattering simply light originating from Rayleigh scattering being scattered again, and ending up traveling in the same direction as the original signal? I say incidentally because we're only talking about 100m of cable as a maximum in this system, so I doubt we're dealing long haul issues.

The Rayleigh scattering can be in same or opposite direction. Direction is rather random really and reflection serves to exagerate any loss effects. I deal with dirty connections all the time at contract manufacturer's premises. Contaminents are typically dust or oils and tend to do just as you say. They'll absorb light like tiny black (gray?) bodies and rob light amplitude. To extend the analogy (and risk confuasing everyone again)... these contaminents have been known to actually burn up and cause disturbing smells and smoke in land terminal connection wher ouput power is boosted up to some really high number (a couple watts) of long haul systems.

It is possible that you are "over driving" the optics, therefor, when an attenuator is introduced, you don't get errors.

What kind of cable lengths are you talking about? 100's or 1000's of metres?

Regards

Dirt and connector problems cause worse reflections than attenuators do.

Cheers

Phil Hobbs

One reasonable guess would be multiple reflection effects -- portions of individual data pulses bouncing off one "incorrectly mated connector, etc.", traveling back to an earlier reflection, getting reflected back forward again. (Sometimes referred to as "multipath".)

Result: at the output end each data pulse is not only weakened a bit, but more important has one or more time-delayed ("multipathed") pulses arriving on top of it.