Symptom index: where to start
Đội ngũ của chúng tôi đang thực hiện bản dịch. Bài viết này tạm thời chưa có bản tiếng Việt và đang được hiển thị bằng tiếng Anh.
You have a symptom, not a diagnosis. Find the row that matches what you see, run the first checks, and follow the link to the detailed material. Every row assumes the diagnostic ladder as the baseline: levels → clean → swap → OTDR.
Link state
| Symptom | First checks | Likely causes | Read |
|---|---|---|---|
| Port down, module shows unsupported / invalid / Mod_Inv | host log, identity on the bench | vendor lock; identity or compliance codes the host rejects | Vendor lock, No link checklist |
| Port down, Rx −40 dBm / LOS on both ends | is Tx enabled? fibre continuity? | fibre cut, Tx disabled by host, Tx/Rx crossed, wrong BiDi pair, wrong wavelength for the mux | No link checklist, Physical mismatches |
| Port down, Rx and Tx levels fine | FEC and speed on both ends; AN on DAC | FEC mismatch, rate/breakout config, SGMII vs 1000BASE-X, CDR at wrong rate | Speed & rate, FEC & link training |
| Port down, Rx very high (> −3 dBm) on a short link | long-reach optic on a short span? | receiver overload | Physical mismatches |
| Works in one direction / LOS on one end only | compare Rx on both ends | dirty connector or weak laser on the failing side; polarity | DDM levels, Physical mismatches |
| Link flaps up/down | levels, CRC counters, connectors | contamination, marginal budget, thermal throttling | Link flapping |
| 400G/800G module recognised but never ready, stuck LowPwr | CMIS module state, power class vs port | handshake / firmware / power class | CMIS issues, Power & thermal |
| DAC does not link, optic in the same port does | AN/link training, cable length, identity | AN disabled, unsupported cable, too long for the host | Copper & DAC, Speed & rate |
Link up, but…
| Symptom | First checks | Likely causes | Read |
|---|---|---|---|
| CRC / input errors climbing | Rx level vs baseline; FEC corrected counters; clean | dirt, marginal budget, dispersion at 10G+ on long CWDM | Link flapping, DDM levels |
| Slow throughput on a multi-lane link, no alarms | per-lane Rx/bias | one degraded lane hidden by LACP | Failures, Per-lane diagnostics |
| Works at 25G, dead at 10G (or 100G ok, 4×25G breakout not) | rate codes, CDR bypass, host breakout config | CDR range, module not breakout-capable, missing extended compliance code | Speed & rate, CDR |
| Absurd DDM values (0 °C, +10 dBm, 0.00 V) | byte 92 calibration bits | external calibration not applied, zeroed constants | Calibration |
| Tx power identical across a batch | compare modules | constant Tx written to memory | DDM levels |
| Module hot, port disables it, "power exceeds" log | power class byte vs port rating; airflow | class mismatch, thermal shutdown | Power & thermal |
| Everything fine for a year, then sudden failures | Tx bias trend, batch | laser ageing, bad batch, heat | Failures |
| Modules died after a switch firmware upgrade | host release notes; identity | stricter identity checks; clones failing authentication | Vendor lock, Failures |
On the bench (programmer / CodingBox)
| Symptom | First checks | Likely causes | Read |
|---|---|---|---|
| Module not detected on the programmer | seating, form-factor adapter, power-up time | not seated, wrong cradle, module in LowPwr | EEPROM read/write errors |
| Reads return FFh / 00h or garbage | address (A0h/A2h/page), bus speed | I²C NACK, wrong page, bus too fast | EEPROM read/write errors |
| WRITE FAIL | password, WP type, clock speed | protected module, hardware WP, timing | Write-protection types, EEPROM recoding |
| Write "succeeds" but reverts after power cycle | module type | temporary storage needs a save command | Write-protection types |
| Host rejects module after edit | checksums | stale CC_BASE / CC_EXT / CC_DMI | Memory map |
Access networks
| Symptom | First checks | Likely causes | Read |
|---|---|---|---|
| ONU does not register on the OLT | serial / LOID, PON standard, ONU Rx | whitelist, wrong wavelength plan (GPON vs XGS-PON), Rx out of class | PON problems |
| ONU shows LOS | SC/APC cleanliness, splitter, polish | dirt, UPC in an APC plant, fibre | PON problems |
If nothing matches
Work the no-link checklist top to bottom, then read the module on the bench with CodingBox: the Check transceiver screen tells you what the module claims to be, and DDM what it is actually doing.