Fiber optic testing, inspection, cleaning and loss budgets
Testing, inspection, cleaning, and loss budgets test the field evidence that a fiber link works. Expect Tier 1/Tier 2 testing, OLTS and OTDR use, endface inspection, cleaning discipline, reference methods, and pass/fail calculation against a designed loss budget.
Test methods and references
Tier 1 testing uses an optical loss test set (light source and power meter) to measure insertion loss and length; it is the basis for acceptance. Tier 2 adds an OTDR to locate events, splices, connectors, and reflectance, and to measure optical return loss. TIA-568.3-D is the key premises fiber test standard. For multimode, use an encircled-flux-compliant launch to make loss measurements repeatable. Set the reference before testing using the specified one-jumper or three-jumper method, and never test through dirty ports. OTDR splice loss should be measured bidirectionally and averaged.
Inspection and cleaning
Inspect every endface with a scope before mating. IEC 61300-3-35 defines acceptance zones for scratches and contamination. Clean dry first with lint-free material; if needed, use an approved wet cleaning method and dry immediately. Wipe in one direction, never scrub back and forth. Inspect again after cleaning. Contamination causes insertion loss, backreflection, and permanent damage; a visually clean connector is not proof.
Loss budget
Allowed passive loss = fiber attenuation + connector pairs + splices + splitters. Add a design margin. Typical planning maxima are 0.75 dB per mated connector pair and 0.3 dB per splice, with singlemode attenuation near 0.4 dB/km at 1310 nm. Compare measured loss to the calculated maximum, not to a vague “it lights up.”
Worked example
A 2 km singlemode link at 1310 nm: 2 × 0.4 = 0.8 dB. Two connector pairs × 0.75 = 1.5 dB; one splice = 0.3 dB. Passive loss = 2.6 dB. Add 3 dB margin = 5.6 dB. A 20 dB link budget leaves ample margin. A measured OLTS loss of 2.4 dB passes; 3.1 dB exceeds the calculated maximum and requires investigation before acceptance.
Traps
Testing with dirty connectors; wrong reference method; using OTDR alone for insertion loss; ignoring EF launch; mismating APC/UPC; forgetting both directions for OTDR; and calling a link good because it has spare budget despite exceeding the standard limit.