FOA CFOS/T cheat sheet

Inspect every connector endface before mating: a single 5 µm particle on a 9 µm core blocks roughly 30% of the core area and can permanently damage the glass.

Use IEC 61300-3-35 zone criteria: Zone A (core) must be free of any scratches or particles; Zone B (cladding, to 25 µm) allows defects ≤3 µm; Zone C (adhesive, to 120 µm) allows defects ≤10 µm; Zone D (contact, to 250 µm) allows defects ≤10 µm with no loose debris.

Magnification for endface inspection is typically 200× for multimode and 400× for single-mode; lower magnification cannot resolve core-zone defects.

Inspect both the connector ferrule and the mating adapter/sleeve; contamination hides in the alignment sleeve and transfers to a clean ferrule on first mating.

Dry cleaning with lint-free wipes is the default; wet cleaning uses 99% isopropyl alcohol or purpose-made fluid applied to the wipe, never poured into a port, and must be followed by a dry pass to remove residue.

One-click or push-to-clean tools are for adapters and receptacles; never reuse a cleaning tip, and never insert a dry tool into a dirty connector and then mate it.

Never use compressed shop air, canned air with propellant, or cotton swabs on polished endfaces; propellant residue and lint become new contamination.

After cleaning, re-inspect; if contamination persists after three cleaning cycles, replace the connector or patch cord rather than continuing to mate it.

Scratches in Zone A that intersect the core are a permanent defect; a scratched connector cannot be cleaned back into specification and should be replaced.

A visual fault locator (VFL) injects visible red light (typically 635–650 nm) into the fiber; breaks, tight bends, and cracked splices scatter light that is visible through the jacket or at the fault point.

VFL effective range is limited by fiber attenuation: roughly 2–5 km in single-mode and less in multimode; it is a continuity and fault-locating tool, not a loss measurement instrument.

Use the VFL in continuous-wave mode for tracing and in pulsed (blink) mode to distinguish the target fiber from adjacent fibers in a bundle.

VFL light escaping at a bend indicates a macrobend; bend radius violations show as a bright glow on the outside of the coil.

Pair the VFL with a tone generator and inductive probe to identify the far end of an unlabeled cable without disconnecting live traffic.

Never look into a fiber or connector while a VFL or any source is active; use a fiber identifier or optical power meter to confirm a fiber is dark before visual inspection.

A VFL cannot see through a connector that is not mated or through opaque jacketing; use it on bare fiber, at a break, or with an adapter that exposes the fiber.

Fault location workflow: confirm continuity with the VFL, then locate the break by moving the source along the cable and observing where the glow stops or the far end loses light.

For buried or long-haul cable, the VFL is a first-pass tool; the OTDR provides the distance-to-fault measurement when the fault is beyond VFL visibility.

Optical power is expressed in dBm: P(dBm) = 10 × log10(P(mW) / 1 mW); 1 mW = 0 dBm, 0.1 mW = −10 dBm, 10 µW = −20 dBm.

Choose a power meter calibrated at the source wavelength; a meter calibrated at 850 nm used at 1550 nm will read inaccurately because responsivity is wavelength-dependent.

Set the meter to the correct wavelength and modulation (CW or 270 Hz/1 kHz/2 kHz tone) before recording a reading; a mismatched tone setting can cause a low or unstable reading.

Zero the meter (dark current) with no light on the detector before each measurement session, and allow the source and meter to warm up per manufacturer specification.

Use a reference patch cord and adapter matched to the connector type; every mating introduces loss and air gaps between dissimilar connectors produce high, unstable readings.

Record absolute power at the transmitter output and at the receiver input; the difference is the link loss, and the receiver input must remain above its sensitivity and below its overload point.

Typical single-mode transmitter output: −3 to +2 dBm for Fabry-Perot/Laser sources; receiver sensitivity for 1G/10G links is roughly −20 to −28 dBm depending on optics.

Uncertainty sources for power measurement include calibration accuracy (±0.2–0.5 dB typical), connector repeatability (±0.1–0.2 dB), wavelength mismatch, and detector nonlinearity.

Never measure a high-power source without checking the meter's maximum input; amplified (EDFA) outputs can exceed +20 dBm and destroy a standard detector.

Insertion loss (IL) is measured by the reference method: record reference power P_ref with a launch cord, then insert the link under test and record P_test; IL(dB) = P_ref − P_test.

One-jumper reference: meter and source connected through one launch cord; two-jumper reference: an additional launch cord is added; three-jumper reference includes a receive cord - each method changes the reference by one or two connector losses.

The reference method must be documented because one-jumper, two-jumper, and three-jumper references produce results that differ by the connector losses included or excluded.

Use the same connector type and polish at the reference and the test point; a UPC-to-APC mismatch produces a large, non-repeatable loss that is not the link's fault.

Clean and inspect every connector before every reference and test measurement; contamination dominates repeatability at the 0.1–0.3 dB level.

Measure in both directions and average when connector loss asymmetry matters, because loss measured from one end includes the near-end connector's loss differently than from the other end.

Test at the operating wavelength (850/1300 nm multimode, 1310/1550 nm single-mode) and record both wavelengths when the link carries both.

Insertion loss test sets measure only the total link loss; they cannot locate a fault, so pair IL testing with OTDR testing for fault location.

Typical acceptance: connector loss ≤0.3 dB (UPC) or ≤0.5 dB (APC/field), splice loss ≤0.1–0.15 dB, and total link loss within the calculated loss budget.

An OTDR sends a pulse and plots backscattered and reflected power versus distance; the slope of the trace is fiber attenuation and spikes are reflective events (connectors, mechanical splices, breaks).

Dead zone is proportional to pulse width: a short pulse (e.g., 10 ns) gives fine resolution and short dead zones but limited range; a long pulse (e.g., 1 µs) reaches farther but hides close events.

Event dead zone is the minimum distance between two reflective events that can be resolved; attenuation dead zone is the minimum distance after a reflective event before backscatter can be measured.

Launch and receive fibers (typically 1–2 km) move the OTDR's first connector and dead zone away from the link under test so the near-end connector can be measured.

Set the refractive index (IOR) to the fiber's actual value (typically 1.467–1.470 at 1310 nm for single-mode); an IOR error scales all distance readings proportionally.

Set range, pulse width, and averaging time to fit the link: 10–30 s averaging is typical for field acceptance, longer averaging reduces noise at the trace tail.

Bidirectional OTDR testing is required for splice loss acceptance: the average of the two directions cancels the gainer effect that makes some splices read as negative loss in one direction.

Reflectance is reported in dB (negative); a clean PC connector is typically ≤−40 dB, APC ≤−60 dB, and a break or open end is near −14 dB.

The OTDR cannot measure loss through a live PON splitter accurately without a dedicated PON-optimized OTDR and wavelength-isolated test port; standard OTDR pulses are blocked or distorted by splitters.

Optical power meter: measures absolute power in dBm or µW; used for transmitter output, receiver input, and loss measurement with a source.

Stabilized light source: emits a known wavelength and power (CW or modulated tone); paired with a power meter for insertion loss testing.

Visual fault locator (VFL): red laser for continuity, polarity, and break location; not a measurement instrument.

Optical time-domain reflectometer (OTDR): measures loss, distance, reflectance, and event location along a fiber from one end.

Optical loss test set (OLTS): combines a source and power meter, often with dual wavelengths and auto-test features, for insertion loss and link certification.

Fiber identifier: clamps onto a fiber and detects traffic or a tone without breaking the connection; used for live-fiber identification.

Fiber cleaver, fusion splicer, and inspection scope: preparation and inspection tools, not measurement instruments, but they determine the loss that the instruments will measure.

Optical spectrum analyzer (OSA) and chromatic dispersion / PMD test sets: used in fiber characterization for DWDM and high-rate systems.

Return loss meter / reflectometer: measures reflected power at a connector or link, reported in dB.

Loss budget formula: total allowable loss = transmitter output power − receiver sensitivity (with margin); the installed link loss must be less than this value.

Installed loss = (number of connectors × connector loss) + (number of splices × splice loss) + (fiber length × attenuation coefficient) + (splice/connector margin).

Typical attenuation coefficients: 850 nm multimode ≈3.0–3.5 dB/km; 1300 nm multimode ≈1.0–1.5 dB/km; 1310 nm single-mode ≈0.35 dB/km; 1550 nm single-mode ≈0.20–0.25 dB/km.

Typical component losses: fusion splice 0.1 dB, mechanical splice 0.3 dB, UPC connector 0.3 dB, APC connector 0.5 dB, splitter 1:2 = 3.5 dB, 1:4 = 7 dB, 1:8 = 10.5 dB, 1:16 = 14 dB, 1:32 = 17.5 dB (excess loss).

Include a design margin (commonly 3 dB for premises, 3–6 dB for outside plant) for aging, repairs, and measurement uncertainty.

For a PON, the budget must include the splitter loss plus feeder, distribution, and drop fiber losses plus connector and splice losses at each end.

Example: 2 km single-mode at 0.35 dB/km = 0.7 dB; 2 connectors × 0.3 dB = 0.6 dB; 1 splice × 0.1 dB = 0.1 dB; total 1.4 dB, well inside a typical 20+ dB budget.

Calculate per wavelength and use the worst-case wavelength; 850 nm multimode links lose roughly three times as much per km as 1300 nm.

A measured loss above the calculated budget indicates a fault, contamination, a bad splice, or a wrong component - not a budget error to be adjusted.

Fiber characterization measures the properties that limit high-speed and DWDM transmission: attenuation, chromatic dispersion (CD), polarization mode dispersion (PMD), and for some systems, fiber nonlinearity.

Attenuation versus wavelength is measured with an OTDR or a spectral loss test set (cutback method); the attenuation profile shows water-peak absorption near 1383 nm in legacy single-mode fiber.

Chromatic dispersion (ps/nm·km) causes pulse broadening; it is compensated with dispersion-compensating fiber or modules, or by electronic dispersion compensation in coherent receivers.

PMD (ps/√km) arises from asymmetry and stress in the fiber; it is a statistical parameter and is specified as a link PMD value, not a per-km constant.

Zero-dispersion wavelength for standard single-mode fiber is near 1310 nm; dispersion-shifted fiber moves it to 1550 nm but creates four-wave-mixing risk in DWDM systems.

Fiber characterization is required for 10G and above over long spans and for all DWDM systems; short-reach premises links typically only need attenuation and insertion loss testing.

CD and PMD test sets use interferometric or Jones-matrix methods and report link values; results are compared against the system's dispersion tolerance budget.

Characterization is done on installed cable before turn-up and after any re-splice or reroute that changes the optical path.

A PON is a point-to-multipoint network with a passive splitter (typically 1:32 or 1:16); all testing must account for splitter loss and the shared feeder fiber.

Standard OTDR testing through a live PON is not possible without a PON-optimized OTDR and a wavelength-isolated test port, because the splitter distributes the test pulse to all subscribers and live traffic interferes.

PON test wavelengths: 1310 nm upstream, 1490 nm downstream, 1550 nm video overlay; test equipment must be filtered to the correct wavelength to avoid interference.

Insertion loss testing on a PON is done from the OLT side to each ONT with the splitter in place; the measured loss includes splitter loss and must be compared to the PON loss budget.

Splitter loss is the largest single contributor: 1:32 ≈17.5 dB, 1:16 ≈14 dB, 1:8 ≈10.5 dB, 1:4 ≈7 dB, 1:2 ≈3.5 dB, plus excess loss and connector losses.

Test each subscriber drop individually and record per-ONT loss; a single high-loss drop indicates a bad connector, dirty endface, or damaged drop cable.

Use a VFL and fiber identifier to verify splitter port assignment and polarity before testing; mis-assigned ports are a common source of false failures.

PON power meters measure upstream and downstream power simultaneously and must be set for the correct PON technology (GPON, XGS-PON, etc.) because burst-mode upstream power differs from continuous downstream power.

Pre-test: confirm the test plan, wavelength, reference method, acceptance criteria, and that all connectors are inspected and clean.

Reference: establish the reference (one-, two-, or three-jumper) and record it; re-reference if the source, meter, or launch cord changes.

Record for every measurement: date, technician, instrument model and serial number, calibration due date, wavelength, reference value, and test direction.

Verify polarity and labeling against the as-built documentation before testing; test the fiber that the documentation says you are testing.

Test both directions for insertion loss when connector asymmetry matters, and both directions for OTDR splice acceptance.

Compare results to the calculated loss budget and the project's acceptance criteria, not to a generic pass/fail number.

Document failures with the OTDR trace, the measured value, the location, and the corrective action taken.

Post-test: verify all connectors are capped, all ports are labeled, and the test report includes uncertainty or margin so the result can be interpreted.

Uncertainty is the range within which the true value lies; a measured loss of 1.0 dB with ±0.3 dB uncertainty means the true loss is likely between 0.7 and 1.3 dB.

Power meter calibration uncertainty is typically ±0.2–0.5 dB; source stability and wavelength accuracy add further contributions.

Connector repeatability contributes roughly ±0.1–0.2 dB per mating; contamination and endface condition can add much more and are not captured by a clean-connector specification.

Reference method choice changes the result by the connector losses included or excluded; report the reference method so results are comparable.

OTDR distance uncertainty comes from IOR error and cursor placement; a 0.1% IOR error on a 10 km link is 10 m of distance error.

OTDR loss uncertainty comes from backscatter variation, pulse width, averaging, and the gainer effect at splices; bidirectional averaging reduces it.

Combine independent uncertainty contributions in root-sum-square (RSS): U_total = √(U1² + U2² + …), not by simple addition.

A measurement result should be reported with its uncertainty and compared to the acceptance limit with that uncertainty in mind; a result within uncertainty of the limit is inconclusive.

Outside plant testing emphasizes attenuation, splice loss, and distance-to-fault over long spans; premises testing emphasizes connector loss, polarity, and short-link insertion loss.

Outside plant: use long OTDR pulses and launch/receive fibers for long spans; premises: use short pulses and short launch fibers to resolve closely spaced events.

Premises links are short, so connector loss dominates; a 100 m link with four connectors has more loss from connectors than from fiber attenuation.

Test outside plant cable after installation and before splicing, after splicing, and after termination; each stage has its own acceptance criteria.

Verify polarity in premises cabling (straight-through vs. crossover) with a VFL or fiber identifier before certifying the link.

Outside plant spans require bidirectional OTDR testing and splice loss averaging; premises links typically require OLTS insertion loss at both operating wavelengths.

Buried and aerial cable testing must account for temperature-driven attenuation changes; test at ambient and compare to the design budget.

Record cable plant documentation (as-builts, splice schedules, test reports) at turnover; the test report is part of the deliverable, not an optional extra.

Inspect every connector endface before mating: a single 5 µm particle on a 9 µm core blocks roughly 30% of the core area and can permanently damage the glass.

Use IEC 61300-3-35 zone criteria: Zone A (core) must be free of any scratches or particles; Zone B (cladding, to 25 µm) allows defects ≤3 µm; Zone C (adhesive, to 120 µm) allows defects ≤10 µm; Zone D (contact, to 250 µm) allows defects ≤10 µm with no loose debris.

Magnification for endface inspection is typically 200× for multimode and 400× for single-mode; lower magnification cannot resolve core-zone defects.

Inspect both the connector ferrule and the mating adapter/sleeve; contamination hides in the alignment sleeve and transfers to a clean ferrule on first mating.

Dry cleaning with lint-free wipes is the default; wet cleaning uses 99% isopropyl alcohol or purpose-made fluid applied to the wipe, never poured into a port, and must be followed by a dry pass to remove residue.

One-click or push-to-clean tools are for adapters and receptacles; never reuse a cleaning tip, and never insert a dry tool into a dirty connector and then mate it.

Never use compressed shop air, canned air with propellant, or cotton swabs on polished endfaces; propellant residue and lint become new contamination.

After cleaning, re-inspect; if contamination persists after three cleaning cycles, replace the connector or patch cord rather than continuing to mate it.

Scratches in Zone A that intersect the core are a permanent defect; a scratched connector cannot be cleaned back into specification and should be replaced.

A visual fault locator (VFL) injects visible red light (typically 635–650 nm) into the fiber; breaks, tight bends, and cracked splices scatter light that is visible through the jacket or at the fault point.

VFL effective range is limited by fiber attenuation: roughly 2–5 km in single-mode and less in multimode; it is a continuity and fault-locating tool, not a loss measurement instrument.

Use the VFL in continuous-wave mode for tracing and in pulsed (blink) mode to distinguish the target fiber from adjacent fibers in a bundle.

VFL light escaping at a bend indicates a macrobend; bend radius violations show as a bright glow on the outside of the coil.

Pair the VFL with a tone generator and inductive probe to identify the far end of an unlabeled cable without disconnecting live traffic.

Never look into a fiber or connector while a VFL or any source is active; use a fiber identifier or optical power meter to confirm a fiber is dark before visual inspection.

A VFL cannot see through a connector that is not mated or through opaque jacketing; use it on bare fiber, at a break, or with an adapter that exposes the fiber.

Fault location workflow: confirm continuity with the VFL, then locate the break by moving the source along the cable and observing where the glow stops or the far end loses light.

For buried or long-haul cable, the VFL is a first-pass tool; the OTDR provides the distance-to-fault measurement when the fault is beyond VFL visibility.

Optical power is expressed in dBm: P(dBm) = 10 × log10(P(mW) / 1 mW); 1 mW = 0 dBm, 0.1 mW = −10 dBm, 10 µW = −20 dBm.

Choose a power meter calibrated at the source wavelength; a meter calibrated at 850 nm used at 1550 nm will read inaccurately because responsivity is wavelength-dependent.

Set the meter to the correct wavelength and modulation (CW or 270 Hz/1 kHz/2 kHz tone) before recording a reading; a mismatched tone setting can cause a low or unstable reading.

Zero the meter (dark current) with no light on the detector before each measurement session, and allow the source and meter to warm up per manufacturer specification.

Use a reference patch cord and adapter matched to the connector type; every mating introduces loss and air gaps between dissimilar connectors produce high, unstable readings.

Record absolute power at the transmitter output and at the receiver input; the difference is the link loss, and the receiver input must remain above its sensitivity and below its overload point.

Typical single-mode transmitter output: −3 to +2 dBm for Fabry-Perot/Laser sources; receiver sensitivity for 1G/10G links is roughly −20 to −28 dBm depending on optics.

Uncertainty sources for power measurement include calibration accuracy (±0.2–0.5 dB typical), connector repeatability (±0.1–0.2 dB), wavelength mismatch, and detector nonlinearity.

Never measure a high-power source without checking the meter's maximum input; amplified (EDFA) outputs can exceed +20 dBm and destroy a standard detector.

Insertion loss (IL) is measured by the reference method: record reference power P_ref with a launch cord, then insert the link under test and record P_test; IL(dB) = P_ref − P_test.

One-jumper reference: meter and source connected through one launch cord; two-jumper reference: an additional launch cord is added; three-jumper reference includes a receive cord - each method changes the reference by one or two connector losses.

The reference method must be documented because one-jumper, two-jumper, and three-jumper references produce results that differ by the connector losses included or excluded.

Use the same connector type and polish at the reference and the test point; a UPC-to-APC mismatch produces a large, non-repeatable loss that is not the link's fault.

Clean and inspect every connector before every reference and test measurement; contamination dominates repeatability at the 0.1–0.3 dB level.

Measure in both directions and average when connector loss asymmetry matters, because loss measured from one end includes the near-end connector's loss differently than from the other end.

Test at the operating wavelength (850/1300 nm multimode, 1310/1550 nm single-mode) and record both wavelengths when the link carries both.

Insertion loss test sets measure only the total link loss; they cannot locate a fault, so pair IL testing with OTDR testing for fault location.

Typical acceptance: connector loss ≤0.3 dB (UPC) or ≤0.5 dB (APC/field), splice loss ≤0.1–0.15 dB, and total link loss within the calculated loss budget.

An OTDR sends a pulse and plots backscattered and reflected power versus distance; the slope of the trace is fiber attenuation and spikes are reflective events (connectors, mechanical splices, breaks).

Dead zone is proportional to pulse width: a short pulse (e.g., 10 ns) gives fine resolution and short dead zones but limited range; a long pulse (e.g., 1 µs) reaches farther but hides close events.

Event dead zone is the minimum distance between two reflective events that can be resolved; attenuation dead zone is the minimum distance after a reflective event before backscatter can be measured.

Launch and receive fibers (typically 1–2 km) move the OTDR's first connector and dead zone away from the link under test so the near-end connector can be measured.

Set the refractive index (IOR) to the fiber's actual value (typically 1.467–1.470 at 1310 nm for single-mode); an IOR error scales all distance readings proportionally.

Set range, pulse width, and averaging time to fit the link: 10–30 s averaging is typical for field acceptance, longer averaging reduces noise at the trace tail.

Bidirectional OTDR testing is required for splice loss acceptance: the average of the two directions cancels the gainer effect that makes some splices read as negative loss in one direction.

Reflectance is reported in dB (negative); a clean PC connector is typically ≤−40 dB, APC ≤−60 dB, and a break or open end is near −14 dB.

The OTDR cannot measure loss through a live PON splitter accurately without a dedicated PON-optimized OTDR and wavelength-isolated test port; standard OTDR pulses are blocked or distorted by splitters.

Optical power meter: measures absolute power in dBm or µW; used for transmitter output, receiver input, and loss measurement with a source.

Stabilized light source: emits a known wavelength and power (CW or modulated tone); paired with a power meter for insertion loss testing.

Visual fault locator (VFL): red laser for continuity, polarity, and break location; not a measurement instrument.

Optical time-domain reflectometer (OTDR): measures loss, distance, reflectance, and event location along a fiber from one end.

Optical loss test set (OLTS): combines a source and power meter, often with dual wavelengths and auto-test features, for insertion loss and link certification.

Fiber identifier: clamps onto a fiber and detects traffic or a tone without breaking the connection; used for live-fiber identification.

Fiber cleaver, fusion splicer, and inspection scope: preparation and inspection tools, not measurement instruments, but they determine the loss that the instruments will measure.

Optical spectrum analyzer (OSA) and chromatic dispersion / PMD test sets: used in fiber characterization for DWDM and high-rate systems.

Return loss meter / reflectometer: measures reflected power at a connector or link, reported in dB.

Loss budget formula: total allowable loss = transmitter output power − receiver sensitivity (with margin); the installed link loss must be less than this value.

Installed loss = (number of connectors × connector loss) + (number of splices × splice loss) + (fiber length × attenuation coefficient) + (splice/connector margin).

Typical attenuation coefficients: 850 nm multimode ≈3.0–3.5 dB/km; 1300 nm multimode ≈1.0–1.5 dB/km; 1310 nm single-mode ≈0.35 dB/km; 1550 nm single-mode ≈0.20–0.25 dB/km.

Typical component losses: fusion splice 0.1 dB, mechanical splice 0.3 dB, UPC connector 0.3 dB, APC connector 0.5 dB, splitter 1:2 = 3.5 dB, 1:4 = 7 dB, 1:8 = 10.5 dB, 1:16 = 14 dB, 1:32 = 17.5 dB (excess loss).

Include a design margin (commonly 3 dB for premises, 3–6 dB for outside plant) for aging, repairs, and measurement uncertainty.

For a PON, the budget must include the splitter loss plus feeder, distribution, and drop fiber losses plus connector and splice losses at each end.

Example: 2 km single-mode at 0.35 dB/km = 0.7 dB; 2 connectors × 0.3 dB = 0.6 dB; 1 splice × 0.1 dB = 0.1 dB; total 1.4 dB, well inside a typical 20+ dB budget.

Calculate per wavelength and use the worst-case wavelength; 850 nm multimode links lose roughly three times as much per km as 1300 nm.

A measured loss above the calculated budget indicates a fault, contamination, a bad splice, or a wrong component - not a budget error to be adjusted.

Fiber characterization measures the properties that limit high-speed and DWDM transmission: attenuation, chromatic dispersion (CD), polarization mode dispersion (PMD), and for some systems, fiber nonlinearity.

Attenuation versus wavelength is measured with an OTDR or a spectral loss test set (cutback method); the attenuation profile shows water-peak absorption near 1383 nm in legacy single-mode fiber.

Chromatic dispersion (ps/nm·km) causes pulse broadening; it is compensated with dispersion-compensating fiber or modules, or by electronic dispersion compensation in coherent receivers.

PMD (ps/√km) arises from asymmetry and stress in the fiber; it is a statistical parameter and is specified as a link PMD value, not a per-km constant.

Zero-dispersion wavelength for standard single-mode fiber is near 1310 nm; dispersion-shifted fiber moves it to 1550 nm but creates four-wave-mixing risk in DWDM systems.

Fiber characterization is required for 10G and above over long spans and for all DWDM systems; short-reach premises links typically only need attenuation and insertion loss testing.

CD and PMD test sets use interferometric or Jones-matrix methods and report link values; results are compared against the system's dispersion tolerance budget.

Characterization is done on installed cable before turn-up and after any re-splice or reroute that changes the optical path.

A PON is a point-to-multipoint network with a passive splitter (typically 1:32 or 1:16); all testing must account for splitter loss and the shared feeder fiber.

Standard OTDR testing through a live PON is not possible without a PON-optimized OTDR and a wavelength-isolated test port, because the splitter distributes the test pulse to all subscribers and live traffic interferes.

PON test wavelengths: 1310 nm upstream, 1490 nm downstream, 1550 nm video overlay; test equipment must be filtered to the correct wavelength to avoid interference.

Insertion loss testing on a PON is done from the OLT side to each ONT with the splitter in place; the measured loss includes splitter loss and must be compared to the PON loss budget.

Splitter loss is the largest single contributor: 1:32 ≈17.5 dB, 1:16 ≈14 dB, 1:8 ≈10.5 dB, 1:4 ≈7 dB, 1:2 ≈3.5 dB, plus excess loss and connector losses.

Test each subscriber drop individually and record per-ONT loss; a single high-loss drop indicates a bad connector, dirty endface, or damaged drop cable.

Use a VFL and fiber identifier to verify splitter port assignment and polarity before testing; mis-assigned ports are a common source of false failures.

PON power meters measure upstream and downstream power simultaneously and must be set for the correct PON technology (GPON, XGS-PON, etc.) because burst-mode upstream power differs from continuous downstream power.

Pre-test: confirm the test plan, wavelength, reference method, acceptance criteria, and that all connectors are inspected and clean.

Reference: establish the reference (one-, two-, or three-jumper) and record it; re-reference if the source, meter, or launch cord changes.

Record for every measurement: date, technician, instrument model and serial number, calibration due date, wavelength, reference value, and test direction.

Verify polarity and labeling against the as-built documentation before testing; test the fiber that the documentation says you are testing.

Test both directions for insertion loss when connector asymmetry matters, and both directions for OTDR splice acceptance.

Compare results to the calculated loss budget and the project's acceptance criteria, not to a generic pass/fail number.

Document failures with the OTDR trace, the measured value, the location, and the corrective action taken.

Post-test: verify all connectors are capped, all ports are labeled, and the test report includes uncertainty or margin so the result can be interpreted.

Uncertainty is the range within which the true value lies; a measured loss of 1.0 dB with ±0.3 dB uncertainty means the true loss is likely between 0.7 and 1.3 dB.

Power meter calibration uncertainty is typically ±0.2–0.5 dB; source stability and wavelength accuracy add further contributions.

Connector repeatability contributes roughly ±0.1–0.2 dB per mating; contamination and endface condition can add much more and are not captured by a clean-connector specification.

Reference method choice changes the result by the connector losses included or excluded; report the reference method so results are comparable.

OTDR distance uncertainty comes from IOR error and cursor placement; a 0.1% IOR error on a 10 km link is 10 m of distance error.

OTDR loss uncertainty comes from backscatter variation, pulse width, averaging, and the gainer effect at splices; bidirectional averaging reduces it.

Combine independent uncertainty contributions in root-sum-square (RSS): U_total = √(U1² + U2² + …), not by simple addition.

A measurement result should be reported with its uncertainty and compared to the acceptance limit with that uncertainty in mind; a result within uncertainty of the limit is inconclusive.

Outside plant testing emphasizes attenuation, splice loss, and distance-to-fault over long spans; premises testing emphasizes connector loss, polarity, and short-link insertion loss.

Outside plant: use long OTDR pulses and launch/receive fibers for long spans; premises: use short pulses and short launch fibers to resolve closely spaced events.

Premises links are short, so connector loss dominates; a 100 m link with four connectors has more loss from connectors than from fiber attenuation.

Test outside plant cable after installation and before splicing, after splicing, and after termination; each stage has its own acceptance criteria.

Verify polarity in premises cabling (straight-through vs. crossover) with a VFL or fiber identifier before certifying the link.

Outside plant spans require bidirectional OTDR testing and splice loss averaging; premises links typically require OLTS insertion loss at both operating wavelengths.

Buried and aerial cable testing must account for temperature-driven attenuation changes; test at ambient and compare to the design budget.

Record cable plant documentation (as-builts, splice schedules, test reports) at turnover; the test report is part of the deliverable, not an optional extra.