FOA CFOS/A common mistakes
FOA CFOS/A · Understand the traps and how to avoid them.
Treating dB and dBm as interchangeable when computing a link budget, so a launch power of 0 dBm is written as '0 dB loss' and connector/splice losses are subtracted from the wrong quantity.
Candidates memorise the loss table (0.3 dB per connector, 0.1 dB per splice, 0.35 dB/km at 1310 nm) without internalising that dB is a ratio and dBm is an absolute power referred to 1 mW. Under exam time pressure the units blur and the arithmetic still 'looks right'.
What to do instead
Before every calculation, label each number as dBm (absolute power at a point) or dB (a loss, gain or margin). Write the budget as an equation in words: source power in dBm minus total loss in dB minus margin in dB equals received power in dBm. Sanity-check the answer: a received power of -3 dBm is plausible, a 'received loss' of -3 dB is not a power at all.
Confusing modal bandwidth and chromatic dispersion limits, assigning a 62.5/125 µm multimode link a reach based on 1300 nm singlemode distances, or assuming OM3 and OM4 differ in attenuation rather than in effective modal bandwidth.
The fibre-type tables in the standards are memorised as a flat list of distances. Candidates do not connect the distance limit to its physical cause, so they cannot tell which column (attenuation, EMB, dispersion) drives the number when the question changes the bit rate or wavelength.
What to do instead
Build a one-page matrix: for each fibre category (OM1-OM5, OS1/OS2) record core/clad size, attenuation at each window, minimum EMB or dispersion specification, and the Ethernet reach at 1G/10G/40G/100G. Learn the cause of each limit: multimode reach is set by modal bandwidth and differential mode delay, singlemode reach at 10G+ by chromatic dispersion and, for long haul, by PMD and OSNR.
Quoting a bend radius or pulling tension from memory without conditioning it on the cable type, installation state and manufacturer data, for example applying the 10x/20x rule to a loose-tube riser or assuming the installed radius equals the loaded radius.
The familiar '20 times the cable diameter during installation, 10 times after' rule is taught as universal. Candidates do not realise that the standard requires following the manufacturer's specification where it is more restrictive, and that figures vary with cable construction and fibre count.
What to do instead
State the rule as a hierarchy: manufacturer's published limits govern; absent those, use the standard's default (commonly 20x outside diameter under load, 10x installed for typical indoor cable). Note the exceptions explicitly: armoured, aerial and microduct cables often carry tighter or looser limits, and mid-span or indoor tight-buffer designs differ from loose tube. When a question supplies a manufacturer figure, that figure wins.
Reporting a test result without the conditions that make it valid, such as quoting a loss figure with no wavelength, no direction and no reference method, or presenting an OTDR trace with no launch/receive fibre and no index of refraction setting.
Candidates study the pass/fail threshold and skip the surrounding procedure. They also conflate one-jumper and three-jumper reference methods, so the reported loss silently includes or excludes a connector, shifting every value by about 0.3 dB.
What to do instead
Always report the full measurement context: wavelength, transmission direction, reference method (one-jumper or three-jumper), test standard used, and, for OTDR, index of refraction, pulse width, range, averaging time and whether launch and receive fibres were used. Learn that two one-jumper references (both directions) give an end-to-end loss comparable to a light-source-and-power-meter measurement, while a single one-jumper reference understates by one connector.
Reading an OTDR trace and calling a reflective event a splice or a non-reflective event a connector, or accepting a 'gain' at a splice as a real improvement rather than an artefact.
The trace is learned as a picture rather than as a physical record. Candidates do not link reflectivity to the event type (connectors and mechanical splices reflect; fusion splices and bends generally do not) or to the cause of a displayed gain (mismatched backscatter between two fibres, or a bad index setting).
What to do instead
Learn the signature of each event: a spike with a dead zone is a reflective event (connector, mechanical splice, crack); a step with no spike is a non-reflective loss (fusion splice, bend, taper). Treat any apparent gain as suspect and re-measure from the opposite direction; a genuine splice loss is the average of the two directional readings, and a bidirectional gain indicates a backscatter-coefficient mismatch.
Choosing a singlemode or multimode transceiver, or a wavelength, by cost or habit instead of by the existing fibre and reach, then discovering the link will not close because the wrong source type was assumed.
The candidate separates 'network design' from 'fibre types' in their head. In practice the two are one decision: source type, wavelength, fibre core size and reach are chosen together, and a 1310 nm singlemode optic cannot be used on multimode plant without a mode-conditioning cord and the correct reach.
What to do instead
Work the design as a chain: application and bit rate, then distance, then fibre available or to be installed, then source/wavelength, then connector and splice count, then loss budget, then margin. Check that the chosen optic's specified reach is achievable on the actual fibre category, and remember the mode-conditioning patch cord requirement when a 1300 nm singlemode optic is used on 62.5/125 µm multimode.
Designing a link that passes on paper with zero margin, or omitting the repair margin and the effect of temperature on attenuation, so a compliant calculation still fails in service.
Candidates add up the component losses, compare to the standard's maximum, and stop. They do not apply the design margin the standards expect, nor do they consider that installed cable loss rises at temperature extremes and that aged connectors and re-spliced fibres consume budget over the plant's life.
What to do instead
Always subtract a design margin (commonly at least 3 dB for indoor plant, more for outdoor or harsh environments) before comparing to the transceiver's sensitivity, and show the margin explicitly in the calculation. Include the number of splices and connectors at their worst-case values, use the cable's maximum specified attenuation rather than typical, and note any temperature range the plant will see.
Mixing up the polarity schemes for multi-fibre connectors and array connectors, wiring a duplex link A-to-A, or assuming a Method B harness works with a Method A patch panel.
Polarity is taught as three named methods without the physical picture of which fibre the transmit signal lands on at the far end. Candidates memorise the labels and cannot trace a signal through a cassette or harness under exam conditions.
What to do instead
Trace one signal path physically for each method: Method A (straight A-to-B), Method B (crossed, paired flip at each end), Method C (crossed pairs in the harness). For array connectors, remember that position 1 must reach position 12 at the far end and that a Type A/B/C harness or cassette is what performs the flip. Draw the connector key orientation at each end; the key-up/key-down rule is the thing most often missed.
Skipping the safety and handling fundamentals as 'obvious', for example disposing of cleaved fibre ends on the bench, looking into a connector to check for light, or using isopropyl alcohol on a connector that requires a dry clean.
These items look like common sense, so candidates do not rehearse them and answer with the plausible-sounding but wrong option (such as blowing dust off with compressed air, which can drive particles into the ferrule).
What to do instead
Learn the correct sequence for each task: inspect with a scope before and after cleaning, dry clean first and use wet cleaning only when specified, use lint-free wipes in one direction, and never look into a fibre or connector without confirming no source is energised (use a scope or an infrared viewer). Treat every fibre fragment as a sharps hazard: cap it, collect it, and dispose of it in a labelled container.
Answering troubleshooting questions by naming the most dramatic cause (a broken fibre) rather than following a fault-isolation sequence from the symptom, so a high-loss connector is misdiagnosed as a cable break.
Exam scenarios describe a symptom (link down, high BER, intermittent errors) and the candidate jumps to the catastrophic explanation. They also forget that a link can be 'up' with errors, which points to loss, contamination or reflectance rather than a complete break.
What to do instead
Use a symptom-to-cause table: total loss of signal suggests a break, disconnected connector or failed transmitter; high but nonzero loss suggests contamination, a tight bend, a poor splice or a damaged connector endface; intermittent errors suggest a loose connector, vibration or temperature-driven loss. Then name the confirming test: OTDR from both ends for location, light source and power meter for end-to-end loss, visual fault locator for a nearby break, and inspection scope for contamination.