2.3 Twisted-Pair Theory and Crosstalk

Why Twists Matter

A twisted pair is a balanced transmission line: the two conductors carry equal-and-opposite signals, and any external electromagnetic field induces nearly equal voltages in both conductors. Because the receiver responds to the difference between the two conductors, the induced common-mode noise cancels. This is the basis of twisted-pair noise rejection, and it is the reason UTP works in noisy commercial buildings without shielding.

The twist does two additional jobs:

1. Limits pair-to-pair crosstalk inside the cable. Each pair twists at a different lay length (turns per inch or turns per meter). When two pairs share the same average twist rate, energy from one couples efficiently into the other. Differing twist rates reduce the average coupling.
2. Limits radiation from the pair. Adjacent half-twists radiate in opposite polarity and cancel at a distance.

The TIA-568 specification defines pair color codes and minimum twist rates but does not specify a single twist density — manufacturers engineer lay lengths to meet the crosstalk limits. Field termination must preserve pair twists as close to the termination point as the standard allows; un-twisting more than ~13 mm (0.5 in) for Cat 5e/Cat 6 degrades NEXT and is a frequent certification failure.

NEXT (Near-End Crosstalk)

NEXT is crosstalk measured at the same end of the cable as the transmitter. Energy transmitted on one pair couples into an adjacent pair and is detected by the test set at the near end while the far-end transmitter on the disturbed pair is silent. NEXT is the dominant pair-to-pair noise mechanism at high frequency because transmitted power is highest at the source end.

NEXT is expressed in decibels (dB) as a loss — higher (more negative) values are better. A NEXT value of 35 dB means the coupled noise is 35 dB below the transmitted signal.

TIA-568 specifies NEXT limits per category and per frequency. The limit curve is not a flat number; it gets tighter (lower dB) as frequency rises. The certification tester compares the measured NEXT against the limit at each frequency and reports the worst-case margin. A NEXT failure at one frequency fails the entire certification, even if the rest of the spectrum passes.

PSNEXT (Power Sum NEXT)

PSNEXT sums the crosstalk power from all three disturbing pairs into a single victim pair at the near end. It is the relevant metric for full-duplex, four-pair operation (1000BASE-T and 10GBASE-T use all four pairs simultaneously). A channel that passes pair-to-pair NEXT can still fail PSNEXT if three pairs simultaneously couple into the fourth.

PSNEXT limits are stricter than NEXT limits by roughly 3 dB because the power sum of three NEXT contributors exceeds any single contributor. Field testers compute PSNEXT from the same set of NEXT pair measurements, so no extra test is needed — but the Technician must understand the result represents worst-case four-pair operation.

FEXT and ELFEXT

FEXT (Far-End Crosstalk) is crosstalk measured at the opposite end of the cable from the transmitter. Because FEXT is attenuated by the cable length, it is typically smaller than NEXT and was historically less critical. Equal-Level Far-End Crosstalk (ELFEXT), now often called Insertion Loss to Far-End Crosstalk Ratio (ACRF), normalizes FEXT by dividing out the cable's attenuation. ACRF allows FEXT limits to be expressed independent of length.

FEXT and its power-sum variant PSACRF matter for 1000BASE-T and 10GBASE-T, which use all four pairs bidirectionally. The certification tester reports PSACRF and passes or fails it against the same per-frequency limit table as the other parameters.

Return Loss

Return loss measures signal energy reflected back to the transmitter due to impedance mismatches along the channel. Sources include:

• Connector pair geometry that deviates from 100 Ω
• Pair untwist at terminations
• Kinked or stretched cable
• Mismatched component categories

Return loss is a critical parameter for full-duplex operation because reflected energy returns to the transmitter's own receiver. A channel can pass NEXT and FEXT and still fail return loss. Field testers report return loss against a per-frequency limit curve and flag the worst-case margin.

Alien Crosstalk (AXT)

Alien crosstalk is coupling between adjacent cables in a bundle rather than between pairs within a single cable. AXT is the dominant noise source for Cat 6A 10GBASE-T operation because at 500 MHz pair-to-pair crosstalk inside a well-engineered Cat 6A cable is already very low; the limiting noise is from neighboring cables in the same bundle or pathway.

AXT cannot be cancelled by the PHY because the disturber is a different cable whose signals are not under the same adaptive equalizer. Mitigation is mechanical:

• Use Cat 6A cables with internal separators or shielding
• Physically separate bundles (TIA recommends ≥30 mm where possible)
• Avoid tightly packed parallel runs over long distances
• Use shielded Cat 6A (F/UTP or U/FTP) in high-density pathways

Cat 6A certification includes an alien crosstalk test (PSANEXT and PSAACRF) that requires sampling surrounding cables in the bundle. This is one of the more complex field tests and a likely topic on the Technician hands-on exam.

Signal-to-Noise Ratio at the Receiver

The receiver's effective SNR is the difference between received signal power and the sum of all noise sources:

SNR ≈ Received signal − (NEXT + FEXT + return loss + AXT + external noise)

As frequency rises, received signal falls (insertion loss increases) and noise rises (crosstalk coupling increases). The SNR therefore drops with frequency. The certification tester's PASS/FAIL is fundamentally a check that SNR remains above the PHY's required level at every frequency in the band.

A Technician who understands this can interpret marginal certifications intelligently. A channel that fails NEXT at 250 MHz but passes everything else has a single pair-coupling problem (usually a termination). A channel that fails AXT only is a bundle problem (cable-to-cable). A channel that fails insertion loss alone is too long, has a bad connector, or has a kink.

Balanced Pair Operation

Twisted pair is a balanced medium: the two conductors are at equal impedance to ground. Balance is what makes common-mode noise rejection work. Anything that breaks balance — uneven twist at termination, a connector that couples one conductor more strongly to ground, a shield left un-bonded — converts common-mode noise to differential noise that the receiver cannot cancel.

This is why shielded cable that is not properly grounded can perform worse than UTP: the unbonded shield couples noise into one conductor more strongly than the other, breaking balance. The Technician must verify shield bonding at every termination, not just at the rack end.