2.9 Troubleshooting Copper Performance

The Troubleshooting Method

Copper troubleshooting is diagnostic, not brute-force. The Technician reads the certification report, identifies the failing parameter and the failure location, and maps that pattern to a root cause. Re-terminating "just to try" wastes time and often fails to address the actual fault. The method:

1. Read the FAIL parameter (NEXT, return loss, length, wiremap, AXT).
2. Read the failure location from the HDTDR/HDTDX diagnostic (near end, far end, mid-cable).
3. Map the parameter + location pattern to a root cause (see table below).
4. Apply the correction (re-terminate, replace cable, separate cables, replace component).
5. Re-test.

Failure-Pattern Map

Split Pairs

A split pair occurs when the two conductors that should form a pair (e.g., white/orange and orange/white) are instead on two different pairs at the connector (e.g., white/orange on pin 1 and white/green on pin 2). The pins are paired correctly (1–2, 3–6, etc.) but the conductors are not the two halves of a single twisted pair.

Symptoms:

• Continuity tester passes (all pins connected end-to-end).
• Wiremap on a certification tester fails — the tester detects pair-to-pair NEXT is wrong.
• NEXT fails catastrophically at the connector — the un-twisted pair has no crosstalk cancellation.

Root cause: incorrect color-code pairing at termination. Correction: re-terminate the offending end with the correct pair colors.

Swapped (Crossed) Pairs

A crossed pair is the result of terminating one end of a channel T568A and the other end T568B. The orange and green pairs are swapped end-to-end. 1000BASE-T will not link; 100BASE-TX may link on pins 1–2/3–6 only if one pair is consistent, but the channel is non-compliant.

Diagnosis: wiremap shows the orange and green pairs crossed at the far end. Correction: re-terminate one end to match the other (use the project's documented scheme).

Length Anomalies

A TDR trace that shows the cable as shorter or longer than physically measured points to a configuration error:

• Wrong NVP — the tester's nominal velocity of propagation is set incorrectly for the cable. The fix is to set NVP per the cable datasheet or calibrate with a known-length sample.
• Excess cable in a ceiling or wall — coiled extra cable adds length; install cable in continuous runs without coiling the excess at the patch panel.
• Stub spurs — abandoned stubs off the main run reflect TDR energy and report false length.

A TDR trace that shows a reflection at a specific distance points to a physical defect:

• Connector (0 m or full length) — bad IDC, damaged contact.
• Mid-cable kink — sharp bend, crushed cable.
• Water ingress — outdoor or below-grade cable with compromised jacket.
• Splice or tap — not permitted in horizontal cabling; replace with a continuous run.

NEXT Failures

NEXT is the most common copper certification failure. The diagnostic approach depends on location:

• Near end (0 m): pair-untwist at the near connector. Re-terminate with ≤13 mm untwist.
• Far end: pair-untwist at the far connector, or wrong-category jack. Re-terminate or replace the jack.
• Mid-cable: physical damage. Run the HDTDX plot to locate the distance, inspect the cable at that point, and replace the damaged section.
• All pairs at the same distance: usually a connector or a transition point. Inspect and re-terminate.
• One pair only: specific pair damaged or untwisted. Re-terminate the affected pair, or replace the cable if the damage is mid-run.

A NEXT failure that moves when patch cords are swapped suggests a bad patch cord; a NEXT failure that stays on the same channel suggests the fixed infrastructure.

Return Loss Failures

Return loss is reflection caused by impedance discontinuity. Diagnose by location:

• 0 m or full length: connector. Re-terminate.
• Mid-cable: kink, crush, water, or a splice. Replace the cable section.
• Multiple reflection points along the cable: damaged jacket, possibly from a tight bend over a long run.

Return loss failures often coincide with NEXT failures at the same connector because both are sensitive to pair geometry. Fix the connector and both parameters usually recover.

Insertion Loss Failures

Insertion loss is length- and connector-driven. Diagnose:

• Length near 90 m (PL) or 100 m (CH) and small margin: the channel is at the limit; consider a shorter route or a higher-category cable with lower attenuation.
• Length well within limit but insertion loss fails: bad connector (high contact resistance), kink (local attenuation), or wrong-category component. Use HDTDR to locate.
• One pair fails insertion loss while others pass: damaged pair. Re-terminate both ends; if still failing, replace the cable.

Alien Crosstalk Mitigation

Alien crosstalk cannot be cancelled by the PHY; mitigation is mechanical. The options, in order of preference:

1. Physical separation — maintain ≥30 mm between bundles in pathways; route data and power in separate raceways.
2. Un-bundling — break dense bundles into smaller groups; use cable trays with dividers.
3. Shielded cable — replace UTP Cat 6A with F/UTP or U/FTP Cat 6A; requires shield bonding at every termination.
4. Internal spline cable — UTP Cat 6A with internal pair separator reduces alien crosstalk without shielding.
5. Reroute — move the worst-case victim cable to a less dense pathway.

A common retrofit error is to add shielding without bonding the shield. An un-bonded shield is worse than no shield (see Section 2.1). Always bond the shield to the TGB at every termination.

Intermittent and Thermal Failures

Not all failures are static. Intermittent failures have characteristic patterns:

• PoE resets under load — high contact resistance at a connector or marginal insertion loss on a PoE pair. Re-terminate.
• Speed drops during business hours — thermal expansion of a marginal connector. Replace the connector or upgrade the patch cord.
• Failures after a MAC — patch cord category mismatch or new bundle density introducing AXT. Re-certify after every MAC.

Documenting the Correction

Every troubleshooting event is documented in the project's as-built record:

• Original failure (parameter, location, margin)
• Root cause identified
• Correction applied
• Re-test result (PASS with margin)
• Date and technician

This record becomes the baseline for future re-certification and protects against repeating the same fault on similar runs.