FOA CFOS/W common mistakes
FOA CFOS/W · Understand the traps and how to avoid them.
Treating the optical splitter as an active device and assigning it a power budget or monitoring role in architecture questions, instead of recognizing it as a passive power-division element.
Candidates carry over Ethernet/access-network mental models where every network node is powered and managed. FTTH PON architecture is passive from the OLT to the ONT, and the splitter only divides optical power, so its insertion loss (roughly 3.5 dB for 1:2, and about 3 dB more per doubling, plus excess loss) must be counted in the link budget rather than any electronics.
What to do instead
Redraw the PON chain from memory: OLT, feeder fiber, ODF, splitter, distribution fiber, drop cable, ONT. For each element write only its loss contribution. Drill splitter loss values and confirm that no active element exists between OLT and ONT.
Choosing the wrong splitter ratio for a given scenario because the candidate optimizes only for cost or only for reach, ignoring the combined loss budget, distance, and subscriber count constraints.
Splitter ratio trades reach against subscriber count: a 1:32 splitter adds about 15 dB of splitting loss versus about 10.5 dB for 1:8. Candidates who memorize a single 'standard' ratio fail scenario questions that specify a long feeder and a small subscriber group, or a short feeder and dense subscribers.
What to do instead
Practice a three-step method: sum all losses (fiber at about 0.35 dB/km at 1310 nm or 0.25 dB/km at 1490/1550 nm, splitters, connectors at about 0.3 to 0.5 dB each, splices at about 0.1 dB each), add margin, then select the largest ratio that still fits the class B+ or C+ budget. Work at least five numeric scenarios.
Confusing OLT, ONT/ONU, and ONT roles in GPON/XGS-PON descriptions, or placing the OLT at the customer premises and the ONT in the central office.
The acronyms are visually similar and vendor marketing uses them loosely. The OLT sits in the central office or headend and manages the PON; the ONT/ONU sits at or near the subscriber and terminates the PON. Reversing them invalidates every downstream architecture, provisioning, and troubleshooting answer.
What to do instead
Anchor the terms with a physical picture: OLT in the rack at the CO, ONT on the wall in the home, splitter in the street cabinet or access point. Write one sentence defining each and check every scenario question against that picture before answering.
Mixing up single-mode and multimode fiber characteristics, such as quoting 50/125 µm core sizes or 850 nm windows for FTTH access links.
Candidates with data-center or LAN backgrounds default to multimode. FTTH access networks use single-mode fiber (about 9 µm core, 125 µm cladding) at 1310, 1490, and 1550 nm, and applying multimode assumptions leads to wrong connector, loss, and distance answers.
What to do instead
Memorize a small comparison table: core/cladding, typical wavelengths, attenuation per km, and typical reach for OM3/OM4 versus G.652.D single-mode. State explicitly in each answer which fiber type the scenario implies.
Selecting connector types and polishing geometries without matching them to the field condition, for example specifying APC connectors where flat-polish SC/UPC is required, or mating APC to UPC.
APC (8-degree angled) ferrules have lower back reflection but must never mate with UPC; doing so causes high loss and damage. Candidates often pick APC because it sounds superior without checking the mating plan or the equipment port type.
What to do instead
Learn the color code (UPC blue, APC green) and the rule that APC mates only with APC. For each scenario, identify the existing port type on the OLT or ONT and select the matching connector; note that APC is preferred where high return loss matters, such as RF video overlay or certain PON wavelengths.
Ignoring bend radius and mechanical protection during cable installation, such as pulling drop cable around a sharp corner or exceeding tensile load, and then attributing later failures to electronics.
Microbend and macrobend loss appears only after installation and often only at 1625 nm or under temperature cycling. Candidates who have not done field work underestimate how easily a tight bend or a staple through the jacket degrades a link.
What to do instead
Memorize the practical limits: minimum bend radius typically 10 times the cable diameter under load and 20 times unloaded for standard cable, with bend-insensitive fiber allowing tighter radii. Know that drop cable should be stapled only with the correct staple gun, and that pulling tension must stay within the manufacturer's rating.
Skipping or misordering the OTDR test setup, for example using the wrong launch condition, the wrong range or pulse width, or forgetting a launch fiber for short drops, and then misreading the trace.
OTDR traces are only valid with correct refractive index, pulse width, and range settings, and short drop links need a launch fiber to see the first connector. Candidates who have only read about OTDRs answer from theory and misidentify the near-end dead zone or a ghost reflection as a fault.
What to do instead
Learn the standard procedure: set the refractive index to the fiber's value, choose the shortest pulse that still reaches the far end, use a launch fiber (and receive fiber for the far connector), and always test at both 1310 and 1550 nm. Practice distinguishing a real event from a ghost by checking whether the event moves with a changed pulse width.
Reporting only insertion loss and omitting return loss or ORL when evaluating a link, or accepting a link that passes loss but fails reflectance.
PON systems, especially those with video overlay or high-order modulation, are sensitive to reflections. Candidates focus on the familiar loss number and forget that a dirty or mismatched connector can pass loss while producing harmful back reflection.
What to do instead
Build a two-column acceptance checklist for every link: insertion loss and ORL/return loss, each with its threshold. Note that APC connectors typically achieve better than -60 dB return loss while UPC is around -50 dB, and that a failed ORL reading usually points to a dirty, damaged, or mismatched connector.
Diagnosing a customer outage by replacing the ONT first, without first checking whether the fault is at the subscriber drop, the splitter, or the feeder, and without confirming the OLT-side status.
The ONT is the most visible element, so it becomes the default suspect. In practice a large share of faults are dirty connectors, a cut drop, a bad splice, or a splitter port issue, and swapping the ONT wastes a truck roll and leaves the real fault in place.
What to do instead
Use a layered isolation method: check OLT port and PON status, then test at the splitter output, then at the subscriber drop, then at the ONT. Clean and inspect every connector with a one-click cleaner and scope before declaring a component failed, and record the loss at each layer.
Treating certification as a pass/fail number only, submitting a link that meets loss but has undocumented test conditions, missing wavelength data, or untraceable test equipment calibration.
Candidates study thresholds but not the documentation requirements of acceptance testing. Certification packages are rejected when the report lacks the wavelength, the test method, the equipment serial and calibration date, or bidirectional results.
What to do instead
Memorize the required report fields: link ID, fiber type, test wavelengths, measured loss and ORL per direction, test standard referenced, instrument model and calibration due date, technician and date. Before submitting, run a checklist against these fields and re-test any link with missing or inconsistent data.