10GBASE-SR Link Over OM3 Exceeds Loss Budget
A 200 m 10GBASE-SR link between two buildings uses a 12-fiber OM3 MPO trunk with LC cassettes at each end. The trunk includes one fusion splice in an outdoor splice closure. The link is tested at 850 nm with an OLTS using a one-jumper reference; one channel measures 2.9 dB insertion loss. The 10GBASE-SR channel budget is 2.6 dB. An OTDR trace shows a 1.3 dB event at the splice closure and normal loss elsewhere. The installer says the link will probably work because 10GBASE-SR equipment usually has margin and the 2.9 dB is only 0.3 dB over budget. The customer needs a standards-compliant acceptance result and wants to know whether to accept, re-splice, or replace the trunk.
1. Calculate the worst-case passive loss for this link and compare it with the measured result.
Read the worked answer
Use the 10GBASE-SR channel budget of 2.6 dB and worst-case component allowances: 0.2 km x 3.5 dB/km = 0.7 dB fiber; four LC connectors = two mated pairs x 0.75 dB = 1.5 dB; one splice x 0.3 dB = 0.3 dB. Worst-case total = 2.5 dB, leaving only 0.1 dB margin. That is a design check, not the acceptance result. The measured 2.9 dB is 0.3 dB above the 2.6 dB budget, so the link fails the standards-based acceptance test even though it may sometimes work with real equipment. 'Usually has margin' is not an acceptance criterion. Reject the result unless the measurement method is proven wrong.
2. How should the OLTS result and the OTDR trace be interpreted together?
Read the worked answer
The OLTS gives end-to-end insertion loss, which is the acceptance number for the channel. The OTDR localizes loss by distance; its 1.3 dB event at the splice closure is a huge red flag because a good fusion splice is typically 0.1-0.3 dB. That single event accounts for much of the 2.9 dB end-to-end loss. OTDR splice readings can be affected by launch conditions and fiber mismatch, so confirm with bidirectional OTDR, but a 1.3 dB event is not normal. The OLTS and OTDR are not competing tests: use the OLTS for pass/fail and the OTDR to find where the loss is.
3. What corrective action and retest sequence do you recommend?
Read the worked answer
Reject the link, reopen the closure, and inspect both fiber ends. Clean and re-cleave, then re-splice; a 1.3 dB splice usually indicates contamination, a poor cleave, a bad arc, or mismatched fiber. Clean and inspect the LC connectors as well, since contamination can add loss at any mating point. Retest with the OLTS using the same one-jumper reference method at 850 nm and record both directions if required. If the re-splice brings the channel below 2.6 dB, accept it. If it remains over budget, the design has almost no margin; consider OM4 fiber, fewer mated connections, or a different optical design.