Product: OSN 6800 version (V100R004C01B01p)
the network is a 40 lambda ring characterized by the following description:
The ring consists of twelve FOADM sites without any OLA sites. (For confidentiality considerations any site name is denoted by first two letters of its name in attachments)
Attachment "Topology" shows the topology of sites, distance between sites and actual attenuation on links.
All the traffic on this ring was only two lambdas (80 and 76) operating normally with good performance (FEC_BEF_COR_ER= 0) and (FEC_AFT_COR_ER= 0).
Each of those two lambdas carried 10G SDH signal and was added/dropped at each site (LSX boards in all sites).
A 10GE traffic was applied on the middle lambda (78) which passes through six sites as follows:
(A/D) TH-JA-MU-DU-KU-ST-SE-SB (A/D)
The pass through connection was done very simply by directly connecting D40 to M40 in each site
The performance of this new lambda showed a very high error probability is "SB" site (FEC_BEF_COR_ER= 3) wherease in "TH" site the performance was very good (FEC_BEF_COR_ER= 0).
BEFFEC_EXC (on LSX of SB site)
1- All amplifiers inputs were reviewed and adjusted against standard value "standard single channel input power+10*log(3)".
2- Using OSA OSNR and flatness were reviewed along the path and it was accepted (+/- 1dB).
3- We interchanged OTUs of both sites i.e. TH site LSX was equipped in place of SB site LSX and SB site LSX was equipped in place of TH site LSX.
observation: (the situation didn't change which proves that both OTUs were OK)
4- We turned our attention to dispersion compensation modules and we started to review links lengths against designed DCMs and actually equipped DCMs.
Here the rrot cause is discovered. The designed DCM module for the span KU-->ST was DCM(A) "20km" but the actually equipped was DCM(D) "80km".
This didn't have a large effect on the previous two lambdas because they were regenerated by LSX boards at each site but for our pass through lambda the effect was very clear.
5- After replacing the wrong DCM module with the correct one, the performance became very good (FEC_BEF_COR_ER= 0) and (FEC_AFT_COR_ER= 0) and the warning alarm was cleared.
1- optical power "commissioning" along direction TH --> SB needs revision.
2- OSNR or flatness requirements along direction TH --> SB are not met.
2- Faulty LSX board.
3- Wrong dispersion compensation.