Key takeaways
- The launch fibre offsets the dead zone to measure the first connector of the link.
- A compliant fusion splice generates an insertion loss without a reflection peak on the trace.
- Comparing wavelengths instantly reveals macrobends concealed in the splice trays.
- The measured budget must remain below the theoretical sum of the losses of the installed components.
Locating connectors, splices, and bypassing the dead zone
Optical time-domain reflectometry sends a series of calibrated light pulses along the strand. The instrument analyses Rayleigh backscatter as well as Fresnel reflections generated by the various glass interfaces. The measurement of each strand is systematically performed across the two operating spectral windows to qualify the integrity of the link. The choice of pulse width determines the resolution: a short pulse separates very close events at the patch panel, while a long pulse allows long sections to be spanned. On the graph, the horizontal axis indicates the linear distance in metres and the vertical axis measures the residual optical power in decibels. An optical fibre in perfect condition results in a steady downward straight line whose constant slope reflects the natural attenuation coefficient of the cable. Any sudden change in this trajectory reveals a component, a break, or a physical stress.
- Establishing the reference lineThe launch lead absorbs the saturation of the transmitter and allows the backscatter curve to stabilise before reaching the input of the measured link.
- Qualifying the first connectionPassing through the patch panel generates a clearly visible Fresnel reflection followed by an attenuation step that reveals the alignment quality of the ferrules.
- Checking inline splicesEach correctly executed fusion splice appears as a downward step without reflection, the height of which measures the insertion loss of the joint.
- Delimiting the end of the linkAdding a tail fibre at the end of the link enables the last connector to be evaluated with the same metrological rigour as the initial connector.
| Event on the strand | Graphic appearance | Typical attenuation | Reflectance level |
|---|---|---|---|
| UPC straight polished connector | Pronounced reflective peak and downward step | Under 0.50 dB | Below -50 dB |
| APC angle polished connector | Reduced reflection peak and downward step | Under 0.50 dB | Below -60 dB |
| Thermal fusion splice | Pure downward step without any peak | Under 0.15 dB | Non-reflective |
| Clean break or open connector | High reflective peak followed by drop to noise floor | Total | Below -14 dB |
Tracking coiling stresses and checking the optical budget
Signal degradation in a patch cabinet commonly stems from an unintentional macrobend caused during the coiling of fibre slack. A strand pinched under a tray hinge or coiled with too small a radius allows light to escape from the core. To isolate this mechanical problem from a splice defect, the technician exploits the spectral sensitivity of silica. In single-mode, measurements are taken successively at 1310 nm and 1550 nm. A defective splice maintains constant attenuation across both wavelengths. Conversely, a bending stress generates a significantly more severe loss at 1550 nm than at 1310 nm, thus pinpointing the crushed bend precisely.
- Optical macrobend
- Excessive mechanical bending imposed on an optical fibre, causing abnormal refraction of light out of the waveguide core without physical breakage of the glass.
The final validation of a link is based on comparing the measured optical budget with the expected theoretical budget. The technician calculates the projected loss by adding the cable's attenuation coefficient, the permissible attenuation of each connector pair, and the cumulative losses of the splices. This theoretical calculation incorporates an operational margin ensuring the proper functioning of active transceivers throughout the service life of the installation. Any overall drift exceeding this envelope requires a detailed point-by-point analysis. According to art. 367 CO, after delivery, the client inspects the state of the work and notifies the contractor of defects. The measurement documentation provides the indispensable technical proof to record compliance or justify the necessary rework.
Measuring each strand and managing dead zones
The acceptance of an optical link requires the measurement of each strand without exception. In IT networks and premises cabling, sample testing is prohibited: each fibre presents its own mechanical stresses. The network electrician performs the acquisition at two distinct wavelengths, typically 1310 nm and 1550 nm in single-mode, or 850 nm and 1300 nm in multimode. This dual measurement unambiguously distinguishes a geometric bending anomaly from a point splice defect.
To obtain a clear trace, the instrument settings are adjusted according to the link. A short pulse width of 5 to 10 nanoseconds provides high spatial resolution to distinguish two closely spaced connectors, whilst a wider pulse of 30 to 100 nanoseconds spans long links without degrading the signal-to-noise ratio. The average acquisition time is kept at 30 seconds per wavelength to stabilise Rayleigh backscatter and smooth the noise floor.
Analysing the start of the link is hampered by the dead zone at the beginning of the trace. During laser emission, the saturation of the receiver caused by the interface reflection masks defects over several tens of metres. To certify the first patch connector, the installer inserts a launch fibre of at least 150 metres in single-mode. An identical tail fibre is connected at the opposite end. This setup pushes the event and attenuation dead zones outside the measured link, ensuring the rigorous inspection of the end connections.
| Type of event | Curve shape | Spectral behaviour | On-site origin |
|---|---|---|---|
| Optical connector | Reflective peak followed by a downward step | Identical attenuation at 1310 and 1550 nm | Contaminated or scratched optical endface, or mechanical misalignment |
| Fusion splice | Pure downward step without reflective component | Stable attenuation at both wavelengths | Misaligned splice, contaminated core, or poorly calibrated arc |
| Bending stress | Downward attenuation step without reflection | Significantly higher attenuation at 1550 nm | Bend radius below 30 mm, tight coiling |
| Spurious ghost peak | Isolated reflective peak without downward step | Identical presence at both wavelengths | Multiple reflection between two highly reflective connectors |
Checking the optical budget and validating the acceptance file
Acceptance of an optical section is based on comparing the expected optical budget with the measured budget. The theoretical budget is modelled in the technical office by adding the cable attenuation coefficient (0.35 dB/km at 1310 nm and 0.22 dB/km at 1550 nm for a standard single-mode fibre), the planned losses for each connector (0.30 dB to 0.50 dB per plug), and the splice losses. Any discrepancy exceeding the permissible margin indicates an installation drift that must be located directly on the trace.
The reflectometric trace accurately pinpoints physical stresses invisible during an optical power meter test. Slack improperly coiled in a junction box, a loose tube pinched in an optical patch panel, or a cable crushed under a rigid cable tie cause wave leakage. This macrobend is recognised by its spectral dependence: the loss is negligible at 1310 nm but often exceeds 1 dB at 1550 nm, as the longer wavelength is less confined within the silica. The metric reading of the suspect event directs the field team to the precise point of stress to release the cable without replacing the link.
The rigour of the measurement deliverables guarantees the conformity of the work during acceptance. According to art. 367 CO, after delivery, the client inspects the state of the work and notifies the contractor of defects. By recording bidirectional OTDR traces, measurement tables, and the comparison with the theoretical optical budget in the as-built documentation, the contractor provides irrefutable proof of installation quality and protects its technical liability.
Rework procedure on non-compliant links and clearing reservations
When the instrument reports a failure, the trace provides the exact linear position of the defect to intervene without guesswork on the construction site. An anomaly generally stems from three distinct origins: excessive reflectance on a connector, an abnormal point attenuation on a splice, or a drop in signal caused by mechanical stress. Directly comparing the expected and measured optical budget enables the defect to be qualified, determines whether the link remains usable, and allows targeted rework of the faulty component to be ordered.
Careful examination of the slope prevents destructive and unnecessary dismantling. A telematics technician does not replace a cable termination without distinguishing the attenuation coefficient from the point loss. The preliminary calculation of the theoretical optical budget, based on the actual length and number of splices, constitutes the reference template against which the measured trace must justify every drop in power.
- Inspection of interfaces and elimination of artefactsDisconnect launch leads, inspect ferrules with a video microscope, and clean optical endfaces. A dirty or scratched connector generates an abnormal reflective peak and widens the dead zone at the start of the trace, masking the first metres of the cable and distorting the analysis.
- Differentiating between a macrobend and a spliceOverlay the traces obtained at the two test wavelengths, for example 1310 nm and 1550 nm in single-mode. If the insertion loss remains identical, the event corresponds to an imperfect splice. If the loss increases sharply at 1550 nm, the defect stems from crushing or an excessive bend radius.
- Physical rework of slack and trayOpen the enclosure identified by the OTDR metric coordinate. Unpack the poorly coiled slack to restore a bend radius greater than 30 mm, reposition the strands without mechanical stress in the routing comb, and redo the fusion if the splice protector compresses the core.
- Verification trace and validationRerun the reflectometry measurement on the corrected strand from both ends of the link in order to calculate the bidirectional average. The final values must fall within the contractual tolerances before exporting the digital traces to the as-built documentation intended for site management.
According to art. 364 CO, the contractor is liable for the proper and faithful execution of the work entrusted to their care. No acceptance file may be submitted with concealed questionable events or artificially widened tolerance thresholds. The measurement of each strand must remain transparent and reproducible.
Frequently asked questions
Why is an optical gain sometimes observed on a splice during an OTDR trace?
A false gain appears when the signal passes from a fibre with a high backscatter coefficient to a fibre with a lower coefficient, often due to differing core diameter tolerances between two manufacturers. The trace displays an upward step. This phenomenon does not create energy. To obtain the true attenuation of the splice event, it is mandatory to perform a bidirectional measurement at both ends and then calculate the arithmetic mean of the two values.
What is the purpose of the launch fibre and what minimum length must it have?
The launch fibre allows the dead zone at the start of the trace to be offset outside the link to be certified. Without this launch lead, receiver saturation caused by the OTDR output connector masks the first connection of the installation. In premises or local area networks, a lead of 150 to 500 metres is used in single-mode (at least 100 metres in multimode). It enables the precise measurement of the insertion loss and reflectance of the first connector.
How can a macrobend in a tray be distinguished from a fusion defect?
The method relies on comparing two wavelengths. In single-mode, the fibre is tested at 1310 nm and 1550 nm. A defective splice shows identical attenuation at both wavelengths. Conversely, a macrobend or a fibre pinched in too tight a radius suffers much more severe light leakage at 1550 nm than at 1310 nm. A significant attenuation difference points directly to a mechanical coiling problem.
What should be done if the measured optical budget exceeds the theoretical calculated budget?
If the measured overall attenuation exceeds the expected optical budget, the origin of the deficit must be isolated metre by metre along the trace. First check the cleanliness of each connector at the cable terminations. If the optical transitions are clean, track down point losses exceeding the permissible thresholds on each strand. Any component or connection displaying an abnormal loss must be cleaned, realigned, or respliced until the safety margin becomes positive again.
Texts and standards cited
- CO art. 364contractor's obligations: liability subject to the same rules as that of an employee; the contractor performs the work in person or under personal supervision; unless customary or agreed otherwise, the contractor procures at their own expense the necessary means, machinery, and tools.
