Time-Domain Reflectometry (TDR) sends a short electrical pulse down a copper cable or optical fiber and measures the time delay and amplitude of the reflected signal. The reflection occurs at impedance discontinuities—connectors, splices, bends, or breaks—where the characteristic impedance Z_0 deviates from the nominal value (100 Ω ± 15% for Cat6a twisted pair, 50 Ω for coaxial cable, or 50 Ω for the "electrical equivalent" of single-mode fiber in OTDR terminology). The round-trip time of the reflected pulse is T_reflection = 2 × d / v_p, where d is the distance to the fault and v_p is the propagation velocity (the velocity factor, typically 0.65-0.70c for Cat6a STP, 0.75-0.85c for SMF-28 fiber). For a copper fault at 100 meters on a Cat6a cable with v_p = 0.67c = 2.01 × 10⁸ m/s, T_reflection = 2 × 100 / 2.01 × 10⁸ = 996 ns ≈ 1 μs. The OTDR or TDR instrument's timing resolution determines the distance accuracy: a 1 ns timing jitter at 1 μs total delay gives a distance uncertainty of Δd = v_p × ΔT / 2 = 2.01 × 10⁸ × 10⁻⁹ / 2 = 0.1 m—sufficient for most data center cable fault localization. However, when the fault is within 1 meter of the instrument (the "dead zone" for OTDRs), the reflected pulse arrives during the instrument's receiver recovery time (typically 1-5 μs for portable OTDRs, corresponding to a 100-500 meter dead zone), making near-end faults undetectable. Our theoretical RTT model bridges this gap by computing the expected round-trip time to any point in the cable and correlating it with the TDR trace, enabling the technician to identify whether a TDR reflection corresponds to a known connector (expected reflection from the theoretical RTT) or an unknown fault (unexpected reflection at an RTT that does not match any planned connector or splice location).

The OTDR dead zone limitation for fiber fault localization can be overcome by using the theoretical RTT model to identify the first valid reflection point. A standard OTDR with 1 μs pulse width has a dead zone of approximately 100-200 meters (the time for the OTDR's laser to recover from the Fresnel reflection at the launch connector). If the fiber break is at 50 meters from the OTDR (e.g., a damaged patch cord in the data center's fiber distribution frame), the reflection from the break arrives at 2 × 50 / (2.05 × 10⁸) = 488 ns, which is within the 1 μs dead zone—the OTDR cannot detect it. The technique to resolve sub-dead-zone faults is launch fiber compensation: the technician connects a known reference fiber (typically 500-1,000 meters of fiber on a spool, with a precisely measured length) between the OTDR and the cable under test. The known RTT to the far end of the launch fiber (computed as L_launch / v_p × 2) serves as a delay line, shifting the launch connector reflection from time 0 to time 2 × L_launch / v_p. Any reflection at an RTT shorter than the launch fiber's round-trip is inside the launch fiber itself and can be ignored by correlating with the theoretical RTT model of the launch fiber. Our theoretical RTT model includes a launch fiber configuration parameter that computes the expected reflection RTTs for each fusion splice and connector in the launch fiber, enabling the OTDR operator to subtract the launch fiber's known reflection signature from the measurement trace and isolate the fiber-under-test's sub-dead-zone faults.

The fiber bend detection through RTT anomaly correlation is an advanced fault localization technique that combines the theoretical RTT model with the OTDR's loss trace. A sharp bend in single-mode fiber (bend radius below 10 mm for G.657.A2, below 30 mm for G.652.D) causes a localized increase in attenuation that appears as a step-loss event on the OTDR trace. The bend loss at a given radius is: L_bend(dB) = A × exp(-R / R_0), where A = 500-1,000 dB for single-mode fiber at 1550 nm and R_0 = 3-5 mm for G.657.A2 bend-insensitive fiber. For a bend radius R = 6 mm (within the G.652.D minimum bend radius of 30 mm but below the G.657.A2 minimum of 10 mm), L_bend = 500 × exp(-6/4) = 500 × 0.223 = 111.5 dB per turn—effectively a total loss point (the OTDR trace drops to the noise floor immediately after the bend). The RTT to the bend location is used to identify which patch panel port or cable management point corresponds to the sharp bend. The theoretical RTT model predicts the RTT to each known bend point (cable manager, patch panel, cable tray direction change) in the structured cabling design, and any bend that does not correspond to a planned RTT is flagged as an installation defect. Our model processes the OTDR trace and identifies all step-loss events whose RTT does not match a planned infrastructure element, creating a list of potential installation defects that the technician can physically inspect.

The optical return loss (ORL) and Fresnel reflection amplitude correlation provides a third dimension of fault characterization beyond RTT and loss. When the OTDR pulse encounters a connector interface, the Fresnel reflection coefficient R = [(n₁ - n₂) / (n₁ + n₂)]² determines the amplitude of the reflection. For an APC (Angled Physical Contact) connector with 8° polish angle, n₁ = n₂ = 1.468 (both sides are silica), but the angle between the two fiber end-faces at the 8° polish creates an effective index mismatch: the return loss is -10 × log₁₀(R) = 60-65 dB for APC, compared to -50 to -55 dB for PC. A connector with ORL worse than -50 dB (indicating a damaged or contaminated end-face) produces a Fresnel reflection whose amplitude is 5-15 dB above the expected value. The theoretical RTT model predicts the RTT to each connector in the link, and the OTDR trace's amplitude at that RTT should match the expected ORL. An ORL anomaly (amplitude 5 dB above expected) at a connector's expected RTT indicates a high-return-loss fault that requires connector cleaning or replacement. Our model automatically calculates the expected ORL at each connector from the connector type (PC/APC) and the measured splice loss, flagging connectors whose measured Fresnel amplitude exceeds the expected value by more than 3 dB—the threshold at which back-reflection can destabilize the DFB laser in 400ZR coherent transceivers.