Adding a 12 km Spur and Upgrading to STM-4 on a Fibre Link
The SCADA backbone of a transmission utility uses a 46 km point-to-point single-mode fibre link at 1310 nm between a 220/66 kV grid substation and the load dispatch centre, carrying STM-1 (155.52 Mbps). Recorded design data: fibre attenuation 0.35 dB/km, ten fusion splices at 0.10 dB each, two connector pairs at 0.5 dB each, transmitter output +2 dBm from an FP laser with 2 nm spectral width, receiver sensitivity −30 dBm at 155 Mbps and −28 dBm at 622 Mbps, receiver overload at −8 dBm. Utility practice is to keep a 6 dB design margin plus 2 dB for ageing and repairs. A 12 km spur with four additional splices must now be spliced in at mid-span with no repeater site, and the link is to be upgraded to STM-4 (622 Mbps) for substation CCTV and IEC 61850 traffic. At 1550 nm the fibre attenuation is 0.22 dB/km and dispersion is 17 ps/(nm·km); EDFAs and APD receivers are available only at 1550 nm.
1. Compute the link margin for the existing link, for the 58 km link at 155 Mbps, and at 622 Mbps, and state pass or fail against the 8 dB design rule.
Read the worked answer
Budget rule: margin = P_tx − P_rx(min) − total loss; required margin = 6 dB design + 2 dB ageing/repair = 8 dB.
Existing 46 km at 155 Mbps Loss = 46 × 0.35 + 10 × 0.10 + 2 × 0.5 = 16.1 + 1.0 + 1.0 = 18.1 dB. Margin = 2 − (−30) − 18.1 = 13.9 dB → passes with 5.9 dB to spare. Received power is −16.1 dBm, well below the −8 dBm overload point.
With the spur, 58 km at 155 Mbps Loss = 58 × 0.35 + 14 × 0.10 + 1.0 = 20.3 + 1.4 + 1.0 = 22.7 dB. Margin = 2 + 30 − 22.7 = 9.3 dB → passes, but only 1.3 dB above the rule. Received power is −20.7 dBm: safe from overload.
Same 58 km at 622 Mbps Margin = 2 + 28 − 22.7 = 7.3 dB → fails the 8 dB rule by 0.7 dB. The higher-rate receiver is 2 dB less sensitive, and that alone sinks it.
Conclusion. The spur is acceptable at STM-1 but the STM-4 upgrade cannot be done on the present optical budget. Note that the 1.3 dB spare at 155 Mbps is thin: a few extra repair splices, a dirty connector or a patch-panel addition of 0.5 dB will erode it, so the spur should not be accepted without recording the as-built loss.
2. Evaluate the practical upgrade options with numbers and recommend one, considering the transmission physics as well as cost.
Read the worked answer
Option 1 – higher-power 1310 nm DFB at +5 dBm, keep the existing fibre and receiver. Margin at 622 Mbps = 5 + 28 − 22.7 = 10.3 dB → passes with 2.3 dB spare. Received power −17.7 dBm, no overload. Cheapest: only the transmit card changes, both ends terminate on the existing G.652 fibre, and 1310 nm is the zero-dispersion region so chromatic dispersion is not a concern at 622 Mbps.
Option 2 – move to 1550 nm with a +3 dBm DFB and an APD receiver at −32 dBm. Loss = 58 × 0.22 + 1.4 + 1.0 = 12.76 + 2.4 = 15.16 dB; margin = 3 + 32 − 15.16 = 19.8 dB. Dispersion check: 17 ps/(nm·km) × 58 km × 0.1 nm (DFB linewidth) ≈ 99 ps, against a 622 Mbps bit period of 1.6 ns - negligible, so no compensator needed. This buys the most headroom and is the only route to EDFA amplification later (1310 nm cannot be amplified by an EDFA), but both ends' optics must be replaced.
Option 3 – regenerator at the spur junction. Solves the budget at any bit rate but needs a powered, air-conditioned, battery-backed site with two extra port pairs and a remote alarm path - high capex and a new failure point on a radial SCADA link.
Recommendation. Option 1 if the requirement is 622 Mbps now and capital is tight; Option 2 if the utility expects STM-16 or DWDM later, because 2.3 dB of spare will not survive a decade of splices and repairs. Reject Option 3 unless a site is needed for another reason. Whichever is chosen, model the OSNR and dispersion budget formally before ordering.
3. List the commissioning tests you would insist on and say what each one protects against.
Read the worked answer
1. Bi-directional OTDR at the working wavelength. Confirms the as-built loss against the 22.7 dB budget, locates high-loss splices and macrobends (a tight bend at a joint pit can add 1–2 dB), and detects a bad mid-span splice in the new 12 km spur. Test from both ends because splice loss is direction-dependent.
2. Splice loss verification (target ≤ 0.10 dB) and connector inspection/cleaning. Prevents a single dirty ferrule (0.5–2 dB) from consuming the whole design margin.
3. End-to-end power meter reading at the receiver. Must match the computed value (−20.7 dBm at 155 Mbps, or the new value) and must sit below −8 dBm to avoid overload; a much stronger reading usually means a wrong attenuator or a shorted patch.
4. Optical return loss / reflectance check. Guards against reflections that destabilise the laser and raise the bit error rate.
5. BER / 24-hour error-free soak with the actual STM-4 traffic and the teleprotection channels. Protects against a link that passes power-budget checks but fails under real jitter and traffic. If Option 2 is chosen, add a dispersion penalty check.
6. Alarm and management verification. Confirm LOS, LOL, BER-threshold and fibre-break alarms reach the NMS, and record the as-built fibre route, splice table and loss budget in the documentation - the next engineer will need them, and future spur additions depend on that record.