Testing PONs
This domain covers passive optical networks: wavelengths, splitter losses, what can and cannot be measured through a splitter, and how to test without taking subscribers offline. The exam tests wavelength assignment and the practical limits of OTDR and power measurement in a PON.
Wavelengths and architecture
A PON runs 1490 nm downstream and 1310 nm upstream, with 1550 nm carrying an RF video overlay on many networks and 1625 nm reserved for testing. The OLT feeds a feeder fiber, a splitter (commonly 1:32), distribution fiber, and the ONT. Splitter excess loss is roughly 3.5 dB for 1:2 and about 17.5 dB for 1:32, which dominates the budget. PON power classes are commonly Class B+ at 28 dB and Class C+ at 32 dB.
What you can measure
- OLT transmit power: with a power meter and, if needed, an attenuator; live-network work requires authorization and care.
- Downstream power at the ONT: the standard verification. Set the meter to 1490 nm, not 1310 nm.
- Upstream: transmission is burst-mode, so an ordinary meter averages the bursts and returns a meaningless value. Use a burst-mode meter or the OLT's own diagnostics.
- Insertion loss per leg: measure power at 1490 nm, or use an OTDR at 1625 nm with a PON filter.
OTDR through a splitter
Firing at 1310 or 1550 nm from the OLT produces one composite trace: the splitter appears as a single high-loss event and the individual legs beyond it cannot be resolved. To test a leg, work from the splitter or the ONT end, or use a PON-optimized OTDR that identifies ports. Never expect 32 separate traces from one shot.
Worked example
An OLT transmits +3.0 dBm. Losses: 1:32 splitter 17.5 dB; 2 km feeder at 0.4 dB/km = 0.8 dB; 0.5 km distribution at 0.4 dB/km = 0.2 dB; four connector pairs at 0.5 dB = 2.0 dB; two splices at 0.1 dB = 0.2 dB. Total 20.7 dB, so expected ONT power = 3.0 − 20.7 = −17.7 dBm. With a −27 dBm sensitivity, the margin is 9.3 dB. The field reading is −22.0 dBm, about 4.3 dB worse than predicted, so the technician inspects and cleans the connectors and finds a contaminated adapter at the splitter.
Traps that cost marks
Reading 1490 nm traffic with the meter set to 1310 nm; using an OTDR through a splitter and expecting per-ONU results; leaving the splitter out of the budget; disconnecting a live PON instead of using a fiber identifier; judging upstream power with an ordinary meter.
Key points
- PON wavelengths: 1490 nm downstream, 1310 nm upstream, 1550 nm video, 1625 nm for testing.
- A 1:32 splitter costs about 17.5 dB and dominates the loss budget.
- Upstream is burst-mode; an ordinary power meter cannot measure it meaningfully.
- An OTDR through a splitter gives one composite trace, not per-leg detail.
- Always set the power meter to the wavelength actually present in the fiber.