Ninety-Six Links, One Wrong Adapter, and a Friday Deadline
Acceptance testing is underway on 96 new Cat 6A permanent links in a two-floor office fit-out. The technician is using a channel adapter on the field tester and has saved one autotest per floor rather than one per link, to save time. Current results: 88 links pass; four links show NEXT failures at the far end on all four pairs; two links report an electrical length of 94 m; one link shows a pass with an asterisk and a return loss margin of 0.4 dB; and one link fails with a wire map showing a split pair across pairs 3-6 and 4-5. The GC wants the test report at 5 p.m. today so the owner can accept the space.
1. Diagnose each of the four failure signatures and state your first action for each.
Read the worked answer
Far-end NEXT failing on all four pairs (4 links). A pattern that hits every pair at one end points at the termination at that end, not the cable: jacket stripped too far, pairs untwisted, punchdowns not fully seated, wrong jack, or a kink just behind the jack. First action: re-terminate the far end, inspect the pair geometry, and retest before condemning the cable. Only after a clean re-termination and a repeat failure do you replace the run.
Electrical length of 94 m (2 links). Over the 90 m permanent link limit. The tester's electrical length includes slack and service loops, so check for an excessive loop at the TR, a mislabeled link, a splice, or a run that took a longer route than drawn. Even if the link is electrically healthy, it is a design non-compliance and must be reported, not corrected on paper. First action: measure and document, and raise it as a change to the design or as-built.
Pass with 0.4 dB return loss margin (1 link). That is inside the tester's accuracy band, so it is effectively a coin flip. Re-terminate or inspect and retest; do not hand over a marginal link with a star next to it.
Split pair across 3-6 and 4-5 (1 link). A wiring error, not a performance issue - conductors from different pairs landed on the same pair's pins, typically one end punched to a different color code than the other. First action: re-terminate both ends to the same wiring scheme and retest.
2. Two testing-setup problems are visible in the description. Explain why each invalidates the results and what must be done.
Read the worked answer
1. Channel adapter on a permanent link test. The channel adapter includes the patch cords in the measurement; the permanent link adapter does not. Testing a permanent link with a channel adapter reports a different configuration from the one being handed over, with different limits - the numbers cannot be compared to the permanent link limit and the results do not demonstrate compliance with what was specified. The fix is to test with the correct permanent link adapters at both ends, matching the link type and the limit selected in the tester.
2. One saved autotest per floor. A saved test is a record for the link that was connected at the time. A single record per floor does not document 48 links, cannot be matched to cabling labels, and cannot support a warranty or a closeout. Every link needs its own saved record, identified by the same label used on the jack and the as-built drawing.
Also verify before re-running:
- Tester calibration is current and the adapters are undamaged.
- The correct test limit and cable type are selected.
- The tester's date and time are correct, since they stamp the records.
- The operator is trained on the link-type selection, not just the autotest button.
Assume every result so far has to be re-run.
3. The GC demands the report at 5 p.m. and asks you to sign the paper so the owner can accept the space. How do you handle it?
Read the worked answer
You cannot sign results you know were produced incorrectly. A signed test report is a statement that the links were tested to the specified limit by a valid method, and it is exactly what the owner will rely on when a link fails in service or when a warranty claim is filed.
What to do instead:
- Tell the GC today, not at 5 p.m., that the test configuration was wrong and the links must be re-run, with an honest estimate of the time required. A re-run of 96 links is hours, not weeks.
- Deliver what is valid. A short interim status - links tested, failures found, remediation under way - is defensible and keeps the project moving without a false signature.
- Fix the scope, not just the failures. Re-terminate the identified links, correct the split pair, resolve the two long links, and retest everything with the right adapters.
- Deliver a proper closeout package: per-link pass/fail records with link IDs matching labels and as-builts, the tester model, serial number and calibration date, the limit and standard used, the test date, and the operator. Include the failure list with the remediation performed.
- Do not initial or sign anything you did not test correctly. Escalate to your project manager if the pressure continues.
The exam-relevant judgment is that the deadline does not change what a passing test means, and that the correct response to pressure is an honest status plus a documented plan, not a signature.