Compare a splice estimate with acceptance test evidence

What you should be able to do

Sort three kinds of loss information correctly: the splicer's estimate, an OTDR event result, and an end-to-end insertion-loss measurement. Review the evidence against a defined acceptance plan without treating unlike numbers as interchangeable.

Start With The Question

Before comparing numbers, identify what each record describes. “Loss: 0.02 dB” is incomplete without an instrument, method, sample identity and scope. A small value may be useful process feedback while proving nothing about the rest of a link.

A splicer generally reports an estimate for the joint it has just made. Preserve that estimate with the splice record and any warnings. Do not rewrite its label as a certified link measurement. The FOA fusion-splicing reference explicitly distinguishes the machine estimate from subsequent testing.

An optical time-domain reflectometer, or OTDR, provides information along the fiber and can characterize a splice event. Its result depends on appropriate setup, event resolution and interpretation. A particular event may be difficult to separate from a nearby connector or other event. Do not invent a standalone splice value when the trace cannot resolve it.

An optical loss test set, or an appropriate light source and power meter, measures insertion loss over the defined test path. Its result depends on a valid reference and the specified method. This total does not identify how much loss belongs to each individual splice.

Check Identity Before Arithmetic

Confirm that each record belongs to the intended fiber. Record endpoints, wavelength, instrument and test method. For an OTDR comparison, locate the same physical event in both directions; its distance from the launch end changes when the test direction changes.

Do not average different wavelengths, different fibers or different events. Do not substitute a nearby trace because its number looks better. An unlabeled screenshot is weak evidence even if it displays a favorable result.

Directional Otdr Effects

Different backscatter characteristics across a joint can produce apparent gain in one direction and exaggerated loss in the other. Appropriate bidirectional testing and averaging of the matching event improve the estimate of that event's loss. A negative event reading does not mean the passive splice creates optical power.

This is not permission to average invalid traces. The event must be identifiable, measurements must be suitable, and the method must follow the equipment and project instructions. Preserve the underlying traces, not only the final arithmetic.

Original Worked Exercise

Assume a fictional training record for the same splice at 1310 nm contains: Splicer estimate: 0.02 dB. OTDR A-to-B event loss: -0.10 dB. OTDR B-to-A event loss: +0.30 dB. Assume both traces validly resolve the same event.

Average event result = (-0.10 + 0.30) / 2 = 0.10 dB. Do not average 0.02, -0.10 and +0.30 together. The splicer estimate is not a third directional OTDR measurement.

A separate valid insertion-loss report for the complete fictional test path at 1310 nm reads 1.80 dB. Its exercise limit is 2.00 dB. Arithmetic margin = 2.00 - 1.80 = 0.20 dB.

That meets this one stated numerical comparison. It does not prove all acceptance requirements were satisfied. The exercise supplies no splice-event acceptance limit, so the 0.10 dB event result cannot independently be declared passing. Required additional wavelengths, directions, polarity, length, reflectance or other project criteria still need their specified evidence.

Do not subtract 0.02 dB from 1.80 dB to “correct” the link measurement. Do not replace a measured result with the estimate. Both numbers retain their own meaning.

Review The Test Plan

Obtain the approved limits and required coverage before deciding acceptance. The plan should define the tested path, reference method, wavelengths, applicable directions and required deliverables. Confirm the selected limit matches the intended test scope.

A missing required result is an incomplete record, not a zero-loss result. An incorrect reference or mismatched test setup requires correction and valid retesting; a favorable displayed PASS does not cure the setup problem. Follow the applicable rules for borderline results and measurement uncertainty rather than inventing a rounding allowance.

Original Practice Decisions

1. Splicer estimate is low, but the required link report is missing: acceptance evidence incomplete.
2. End-to-end loss passes its stated limit, but a required event criterion fails: resolve the failed criterion.
3. One OTDR direction shows a gainer: obtain and evaluate the matching reverse-direction evidence.
4. Two screenshots have different fiber IDs: do not average them.
5. All required evidence is present: verify it against the plan and retain the authorized review record; do not claim approval merely from document count.

Worked through

Fiber ID / endpoints: Splice location / identifier: Splicer estimate and warnings: OTDR event IDs / wavelengths / directions / trace filenames: Event averaging method and result: Insertion-loss report / reference method / wavelength: Applicable limits and source: Missing or invalid evidence: Reviewer disposition / required corrective action:

Knowledge Check

Where beginners go wrong

Sources