A Red LOS Light on a New FTTH Drop
A new FTTH connection is being commissioned for a customer who teaches online and needs service within the hour. The OLT port transmits +3 dBm at 1490 nm into a 1:32 splitter in a street cabinet, with about 2.5 dB of feeder and distribution loss before the splitter. The drop is 95 m of 2 mm indoor cable, run by the previous contractor using a staple gun at roughly 40 cm intervals, including a corner turned with an 8 mm radius. At the customer wall box there is one field-installable mechanical splice measured at 0.4 dB, and the ONT pigtail is SC/UPC (blue) plugged into an SC/APC (green) adapter. The ONT shows a red LOS LED and reports receive power of -30.8 dBm. The technician has one pre-connectorised 120 m SC/APC drop cable in the van and a VFL, power meter and cleaning kit.
1. Calculate the expected receive power and quantify the gap. Then rank the possible causes and state the order in which you would test them.
Read the worked answer
Expected receive power: +3 dBm at the OLT, minus about 17.5 dB for a 1:32 splitter, minus 2.5 dB feeder and distribution, minus about 1.5 dB for the connector pairs at the OLT, cabinet and wall box, minus about 0.03 dB for 95 m of drop cable, minus 0.4 dB for the mechanical splice. That gives roughly -19 dBm. The ONT reports -30.8 dBm, so about 11.8 dB is unaccounted for, far more than a single connector issue.
Ranked causes:
- Wrong or cascaded splitter port, or the customer provisioned on a 1:64 path (about 3.5 dB extra, and a second splitter stage could explain 15 dB or more).
- Drop cable damage from staples and the 8 mm corner, capable of several dB at 1490 nm.
- Dirty or damaged connector end faces, 0.5-5 dB.
- The APC-to-UPC mating, typically 0.5-2 dB plus high back reflection.
Test order: confirm the splitter port and stage count against the OLT record first, because it is free and can explain the whole gap; then measure at the wall box with a power meter to split the loss into network side and drop side; then inspect and clean connectors; then use the VFL on the drop with the far end disconnected to find the crushed point. Do the cheap, non-destructive checks before replacing 95 m of cable.
2. The APC-to-UPC mating has been described by a colleague as acceptable if the loss reads low. Explain why it is not acceptable and what the correct remedy is.
Read the worked answer
SC/APC ferrules are polished at an 8 degree angle; SC/UPC ferrules are flat. Mated together, the two end faces cannot achieve proper physical contact. An air gap remains at part of the interface, so insertion loss rises (typically 0.5-2 dB but variable and unstable) and back reflection climbs to around -35 dB or worse, against about -60 dB for a properly mated APC pair. That reflection travels back through the splitter and degrades the whole PON, not just this customer, and it can disturb the OLT receiver and other ONTs on the same port.
There is also a hardware risk: the angled ferrule can chip or crack the flat ferrule, and repeated mating damages both. So the loss reading is not the acceptance test; the geometry is.
Correct remedy: make the interface consistent. Replace the wall-box adapter and pigtail with SC/APC so the green APC connector mates with green APC, or if a hybrid is unavoidable for a short temporary restoration, use a purpose-made hybrid patch cord and record it as temporary. Never leave a hybrid adapter in a customer premises. Colour code is the control: green is APC, blue is UPC, and the two must not be mated.
3. Describe the restoration you would carry out and the customer-facing communication and documentation that go with it, given that the customer is waiting and has already been let down once.
Read the worked answer
Restoration: replace the stapled and kinked 95 m drop with the pre-connectorised 120 m SC/APC cable, which removes both the staple crush and the 8 mm bend in one action and gives a factory-terminated end. Route it with a minimum bend radius of 30 mm, use adhesive clips or hook-and-loop rather than staples, and coil the surplus as a loose 200 mm figure-eight or loop, never a tight coil. Replace the UPC adapter at the wall box with SC/APC and clean every end face before mating. Then measure at the ONT with the power meter and confirm the reading is at least -25 dBm and inside the Class B+ window of about -8 to -27 dBm, run a speed test, and clear the LOS alarm before leaving.
Communication: explain in plain language what was wrong (cable squeezed by staples and a sharp corner, plus mismatched connector types) without blaming the previous contractor by name, give a realistic completion time rather than optimistic guesses, and confirm the fix will not recur. Since the customer teaches online, agree a short downtime window if anything remains to be done.
Documentation: job card with ONT serial number, splitter port and stage count, measured receive power before and after, OTDR or VFL results, photographs of the new route and the wall box, the removed staples and UPC adapter, and customer signature. Update the as-built record so the next technician knows the drop was replaced and why.