GPON Extension Exceeds Its Optical Power Budget
A regional ISP is extending GPON to a new housing estate. The design uses a class B+ OLT with a 28 dB optical budget at 1490 nm downstream, a 1:32 splitter with 17.5 dB insertion loss, 20 km of single-mode fiber at 0.35 dB/km, four connectors at 0.5 dB each, six fusion splices at 0.1 dB each, and a 3 dB design margin. The upstream direction uses 1310 nm and must also close. The field team wants to keep the 1:32 splitter for port density, but the first power meter readings at the ONT are marginal on longer drops. The ISP has class C+ optics available at higher cost, and a 1:16 splitter in stock. The estate developer is pressing for service activation in two weeks, and the ISP does not want to activate customers on a link that will fail in hot weather or after connector rework.
1. Calculate the passive loss and determine whether the B+ design closes. Show the reasoning.
Read the worked answer
Splitter loss is 17.5 dB. Fiber loss is 20 km × 0.35 dB/km = 7.0 dB. Connectors contribute 4 × 0.5 = 2.0 dB. Splices contribute 6 × 0.1 = 0.6 dB. Passive total = 17.5 + 7.0 + 2.0 + 0.6 = 27.1 dB. Adding the 3 dB design margin gives 30.1 dB, which exceeds the 28 dB B+ budget by 2.1 dB. The design does not close with guaranteed margin. It may appear to work on a short, clean drop, but connector loss, temperature, and repair splices can consume the remaining headroom. The upstream 1310 nm direction also needs a separate check because loss and ONT launch power differ.
2. Recommend a design change and discuss the trade-offs.
Read the worked answer
A 1:16 splitter reduces splitter loss by 3.0 dB to 14.5 dB, making the passive total 24.1 dB; with the 3 dB margin the result is 27.1 dB, fitting the 28 dB budget with only 0.9 dB spare. That is workable but tight after field repairs. Class C+ optics provide a 32 dB budget, so the original 1:32 design gives 32 - 30.1 = 1.9 dB spare and keeps port density, but costs more and may require matching ONT classes. A 1:8 splitter gives more margin but worsens economics. If port density is the priority and fiber loss is controlled, choose C+ with the 1:32; otherwise use 1:16 and re-splice. Do not remove the design margin to make the numbers fit.
3. What field validation and operational controls should be required before activation?
Read the worked answer
Perform OTDR tests at 1310, 1490, and 1550 nm on each feeder and drop, record splice and connector loss, and inspect connectors for contamination. Measure optical power at the ONT in both directions and compare with the calculated budget and receiver sensitivity. Check bend radius and splitter port assignments. Validate the worst-case drop, not only the shortest one, and leave the specified margin intact. Operational controls include locking the bill of materials to the approved splitter and optics, prohibiting unauthorized extra splices, and re-measuring after any rework. Hot-weather and repair scenarios are why a design that barely closes should not be activated.