GPON Subdivision: Splitter Placement Under a Tight Loss Budget
A developer asks you to design GPON service for a 384-home subdivision. The central office is 14.2 km from the farthest ONT by the proposed route. You are considering a single 1:32 splitter in the CO, 1490 nm downstream, with a 0.35 dB/km fiber loss, 16.5 dB splitter loss, four 0.3 dB connectors, six 0.1 dB splices, and a 3.0 dB design margin. The ITU-T Class B+ budget is 28 dB. The developer wants the lowest first-cost architecture but also wants the network to support a future 1:64 split or an XGS-PON upgrade without rebuilding the feeder. Homes will be built in three phases, with only 40 percent occupancy in year one. You must recommend a splitter architecture and state what would change if the measured loss exceeds the calculated value.
1. Calculate the worst-case downstream loss for the CO-based 1:32 design and determine whether it meets Class B+.
Read the worked answer
Fiber: 14.2 km × 0.35 dB/km = 4.97 dB. Splitter: 16.50 dB. Connectors: 4 × 0.3 dB = 1.20 dB. Splices: 6 × 0.1 dB = 0.60 dB. Design margin: 3.00 dB. Total: 26.27 dB. Class B+ is 28 dB, so the design passes by 1.73 dB. That margin is real but not generous: an extra patch-panel connector pair or a higher-loss splitter can consume most of it. The calculation also assumes only four connectors end to end; if the design adds an FDH patch field, the connector count must be increased and the margin recalculated. A passing loss budget alone does not prove the design is robust, so the recommendation should preserve as much margin as possible.
2. Recommend a splitter architecture for the three-phase build and explain the trade-off between first cost and upgrade headroom.
Read the worked answer
Keep a centralized 1:32 splitter but move it from the CO to a field cabinet about 6 to 7 km out. That cuts feeder loss by roughly 2.1 to 2.5 dB, raising the margin above 3.8 dB while retaining one splitter to manage. A fully distributed 1:8 plus 1:4 combination is more flexible for phased occupancy but has about 17.7 dB total splitter loss, which is worse than the 16.5 dB 1:32 and erodes the budget. The lowest first cost for year-one 40 percent occupancy is a 1:32 in the CO, but it leaves little reserve for a future 1:64 split, which adds roughly 3 dB. For upgrade headroom, the field cabinet is the better balance: it supports 1:32 today and allows a later 1:16 or higher-power optic without replacing the feeder.
3. The contractor's OTDR shows the farthest link at 25.1 dB, but the OLTS shows 29.2 dB at 1490 nm. What should you investigate, and what corrective action would you recommend?
Read the worked answer
The 4.1 dB disagreement is larger than normal measurement uncertainty, so treat it as a test-method problem before condemning the fiber. Check the OTDR launch and receive cables, the pulse width, and whether the OTDR was set to the correct wavelength and group index. Verify that the OLTS was zeroed with the correct reference method and that both meters used the same connector type and launch condition. Also inspect for a macrobend near a handhole or a dirty connector: a bend can be nearly invisible on a short-pulse OTDR trace but lossy at 1490 nm. If the OLTS result is confirmed, recheck the loss budget with the actual connector count. Corrective options in order of cost are cleaning and repolishing connectors, replacing a high-loss patch cord, rerouting a bend, and only then reducing the split ratio to 1:16 or relocating the splitter. Do not accept the link on the lower OTDR number alone.