Procuring an unlit Dark Fibre DWDM pair between carrier points of presence (PoPs) or enterprise data centres gives infrastructure teams absolute transmission sovereignty. However, taking delivery of dark fibre without full Tier-2 OTDR characterization traces is among the costliest mistakes in telecommunications engineering.
An Optical Loss Test Set (OLTS) only reports aggregate end-to-end decibel loss. It cannot indicate whether 80% of your link budget is being consumed by a pinched fiber in an underground Openreach chamber, a dirty patch panel, or a substandard fusion splice 14km out.
Physics of the Trace: Rayleigh Backscatter vs. Fresnel Reflection
An OTDR operates as optical radar: it injects high-speed laser pulses down the glass core and measures the minute reflections that bounce back over time. The resulting trace represents two distinct physical phenomena:
- Rayleigh Backscattering (Downward Slope): Caused by natural, microscopic density fluctuations in silica glass. As light travels down the cable, a predictable fraction scatters backward toward the detector. The steady, downward angle of the trace represents fiber attenuation, measured in dB/km.
- Fresnel Reflections (Sharp Upward Spikes): Occur when light encounters an abrupt change in the refractive index—most commonly an air gap at a mechanical connector, a patch lead, or a clean fiber end. While these spikes pinpoint connector locations, excessive reflectance degrades high-speed coherent optics.
The Dead Zone Trap: Event vs. Attenuation Dead Zones
When an OTDR pulse strikes a reflective connector, the photodetector saturates. During this recovery period, the OTDR is effectively blind to subsequent fiber anomalies:
- Event Dead Zone (EDZ): The minimum distance required for the reflectometer to distinguish between two consecutive reflective events. A short EDZ (e.g. 0.8m) allows an engineer to differentiate two patch leads on adjacent rack frames.
- Attenuation Dead Zone (ADZ): The longer distance required for the backscatter trace to settle within ±0.5 dB of its linear trajectory. Splice losses or macro-bends located inside the ADZ cannot be measured accurately.
Never plug an OTDR directly into an enterprise patch panel. Always insert a 500m to 1,000m launch fiber lead at the transmitter and a matching tail lead at the remote terminus. This shifts the initial and terminal connector saturation outside the cable under test, allowing 100% of both patch panel connectors to be characterized inside the linear trace window.
OTDR Event Evaluation Standards Matrix
When reviewing carrier handover packets or commissioning private enterprise dark fibre, compare every trace event against ITU-T G.652 single-mode standards:
Optical Time-Domain Reflectometry Acceptance Standards
| Cable / Termination Event | Trace Signature | Typical Measured Loss | Maximum Permissible Loss | Tier-2 Acceptance |
|---|---|---|---|---|
| Fusion Splice | Non-Reflective Event | 0.02 dB to 0.05 dB | 0.10 dB (Openreach / ITU-T) | Strong |
| Angled Physical Contact (LC/APC) | Reflective Event | 0.20 dB to 0.35 dB | 0.50 dB (ORL > 60 dB) | Exceptional |
| Ultra Physical Contact (LC/UPC) | Reflective Event | 0.25 dB to 0.40 dB | 0.75 dB (ORL > 50 dB) | Suitable |
| Macro-bend / Pinch | Non-Reflective Wavelength-Dependent | High at 1625/1550nm, Low at 1310nm | 0.15 dB delta between wavelengths | Limited |
| Mechanical Splice / Dirty End-face | High Reflective & Attenuation Spike | 0.50 dB to 2.50 dB+ | Must be cleaned or re-fused | Not Recommended |
Diagnosing Macro-Bends: The Dual-Wavelength Test
A common installation defect during underground cable pulling through congested Openreach ducts or footway boxes is a macro-bend—where a tight bend radius causes light to leak from the core into the cladding.
An OLTS testing at 1310nm will often show normal attenuation. However, longer wavelengths are significantly more vulnerable to bend-induced leakage. By testing the fiber simultaneously at both 1310nm and 1550nm (or 1625nm):
- Fusion Splice: Exhibits identical loss (~0.04 dB) at both 1310nm and 1550nm.
- Macro-Bend: Shows negligible loss at 1310nm (<0.05 dB), but spikes to 0.5 dB to 2.0 dB loss at 1550nm.
Model Your Optical Link Budget & Attenuation Limits
Calculate cumulative fiber attenuation, splice counts, connector margins, and receiver saturation across 10G, 100G, and 400G optics:
Bidirectional Averaging: Correcting "Gainer" Splices
In spliced cables comprising different manufacturing reels, differences in Mode Field Diameter (MFD) and backscatter coefficients frequently produce an apparent "gainer"—a trace event where signal power appears to increase after a splice.
This violates thermodynamic laws: the splice does not amplify light; rather, the downstream fiber backscatters more light to the receiver. To establish the true physical insertion loss, field engineers must conduct bidirectional OTDR testing (testing from Point A to Point B, and from Point B to Point A) and compute the arithmetic average:
True Splice Loss (dB) = [ Loss(A → B) + Loss(B → A) ] / 2