Sizing a Splitter for a Rural FTTH Build
A rural FTTH project will serve 96 homes from one OLT PON port. Homes are spread along 6 km of road; the farthest home is 5.8 km from the OLT. The design uses a 1:32 splitter at the OLT cabinet and a second-stage 1:8 splitter near the homes, giving 1:256 total - but the OLT port can only support 1:128. The planner wants to keep the 1:32 at the cabinet and change the second stage. Class B+ optics are used: 28 dB budget, OLT launch +2 dBm, ONU sensitivity -27 dBm. Estimated fiber loss is 0.35 dB/km at 1490 nm, connector loss 0.5 dB per mated pair (4 pairs), and splice loss 0.1 dB each (12 splices). Splitter losses: 1:32 = 17 dB, 1:16 = 14 dB, 1:8 = 10.5 dB.
1. What second-stage split ratio should be used, and does the resulting design close the power budget?
Read the worked answer
Second stage must be 1:4 (1:32 × 1:4 = 1:128, which matches the OLT port limit).
Power budget check (worst case, farthest home at 5.8 km):
- Fiber: 5.8 km × 0.35 dB/km = 2.03 dB
- Connectors: 4 × 0.5 = 2.0 dB
- Splices: 12 × 0.1 = 1.2 dB
- Splitter 1:32 = 17 dB
- Splitter 1:4 = 7.2 dB (use ~7.2 dB typical for 1:4)
- Total = 29.4 dB
That exceeds the 28 dB Class B+ budget by ~1.4 dB. So 1:32 + 1:4 does not close at 5.8 km with these assumptions.
Options: move the first splitter closer to the homes (reduces fiber loss), reduce connector/splice count, or use a higher-power optic class (e.g., Class C+).
2. The planner suggests moving the 1:32 splitter out to the field and using 1:4 at the cabinet. Does that help?
Read the worked answer
Yes, it can - because it changes where the loss sits relative to the fiber run.
- With 1:4 at the cabinet and 1:32 in the field, the high-loss 1:32 splitter is now closer to the homes, so the fiber loss is incurred on a shorter effective run for most subscribers.
- However, the total splitter loss is unchanged (17 + 7.2 = 24.2 dB either way). The only savings come from reduced fiber loss on the long feeder if the field splitter is placed partway out.
- If the 1:32 is placed at, say, 4 km, the feeder fiber loss drops to 4 × 0.35 = 1.4 dB, saving ~0.6 dB - still not enough to close the 1.4 dB gap by itself.
Conclusion: moving the splitter helps but is not sufficient; the design still needs a higher-power optic class or a shorter reach.
3. If the operator insists on 1:128 with Class B+ optics, what design change is most defensible?
Read the worked answer
The most defensible change is to reduce the optical loss in the field, in this priority order:
- Reduce splices and connectors - 12 splices and 4 mated pairs cost 3.2 dB. Consolidating splices and using pre-terminated cable can recover 1–2 dB.
- Move the splitter closer to the subscriber so the long feeder is on the low-loss side of the splitter.
- Use lower-loss splitter technology (planar waveguide with tighter uniformity) - modest gains.
- Only then consider a higher-power optic class (Class C+) if the operator will accept it.
What is not defensible: ignoring the 1.4 dB shortfall and hoping field conditions are better than worst case. A 25-year FTTH design must close on worst-case loss, not typical loss.