Fiber-Optic Link Loss and Dispersion Budget Project

This project builds a commissioning package for a single-mode fiber-optic link. The engineering task is not only to prove that light reaches the far end. The task is to show that the installed link has enough optical power margin, enough dispersion margin, acceptable reflection and event behavior, traceable measurement evidence, and clear operating limits for the team that will maintain it.

The final deliverable is a fiber link acceptance report. It should be detailed enough for a design reviewer, field engineer, or operations engineer to reproduce the decision and diagnose the link later without guessing which assumptions were used.

Project Objective

Commission a fiber-optic link that carries a 10 Gbit/s Ethernet service between an aggregation site and a remote technical building. The project must answer:

1. Does the installed loss stay below the design loss limit at the operating wavelength?
2. Is the received optical power above receiver sensitivity after reserving design margin?
3. Is the maximum received power safely below receiver overload?
4. Is accumulated chromatic dispersion inside the transceiver tolerance?
5. Do OLTS and OTDR records agree closely enough to support acceptance?
6. Which residual risks, monitoring thresholds, and maintenance controls should be handed to operations?

The project produces a calculation worksheet, an optical test plan, an acceptance table, an exception log, and an as-built handover record.

Baseline Scenario

Use the following scenario as the design basis. Replace the numbers with site data when applying the workflow to a real project.

Parameter	Value
Service	10 Gbit/s Ethernet backhaul
Fiber type	single-mode OS2
Operating wavelength	1550\ \text{nm}
Route length	38\ \text{km}
Fiber attenuation coefficient	0.22\ \text{dB/km} at 1550\ \text{nm}
Mated connector pairs	6
Design loss per mated connector pair	0.35\ \text{dB}
Fusion splices	18
Design loss per splice	0.08\ \text{dB}
WDM/filter allowance	0.80\ \text{dB}
Bend and route allowance	0.50\ \text{dB}
Required design margin	3.0\ \text{dB}
Transmitter minimum optical power	+1.0\ \text{dBm}
Transmitter maximum optical power	+4.0\ \text{dBm}
Receiver sensitivity	-18.0\ \text{dBm}
Receiver overload threshold	-3.0\ \text{dBm}
Chromatic dispersion coefficient	17\ \text{ps/(nm km)}
Transceiver dispersion tolerance	800\ \text{ps/nm}
PMD coefficient	0.10\ \text{ps}/\sqrt{\text{km}}
PMD tolerance	10\ \text{ps} differential group delay

The route is long enough that attenuation, connector practice, repair margin, and dispersion all matter. It is also short enough that receiver overload must still be checked if a high-output optic or a short bypass patch is later installed.

Step 1: Define the Acceptance Boundary

The acceptance boundary must be explicit. For this project, accept the path from optical handoff at Site A to optical handoff at Site B, including:

• transceiver type and wavelength;
• patch cords and connector end faces at both sites;
• outside plant fiber, splice trays, and intermediate closures;
• WDM/filter or passive optical components in the path;
• optical distribution frames and labeling;
• monitoring values reported by the transceivers;
• OLTS, OTDR, and traffic-test evidence.

Do not accept only a switch port that happens to show link up. A link can come up with poor margin, dirty connectors, high reflection, hidden bend loss, or dispersion close to its operating limit. Those conditions often become outages during temperature changes, maintenance, repair splicing, or transceiver replacement.

Step 2: Build the Optical Loss Budget

The end-to-end physical loss before design margin is:

Fiber attenuation is:

Substituting the baseline values:

Connector loss:

Splice loss:

Total path loss before reserved design margin:

The received optical power using minimum transmitter output is:

Available margin after the required design allowance is:

The link closes, but the post-allowance margin is not large. That result affects maintenance policy: future patching, additional filters, repair splices, or connector degradation cannot be added casually. Any change that consumes more than about 2.8\ \text{dB} after the design allowance needs engineering review.

Step 3: Check Receiver Overload

Sensitivity checks only weak-signal operation. A commissioning package must also check strong-signal operation, especially when optics can be substituted during maintenance.

Use a conservative low-loss condition to estimate maximum received power. Suppose the same route, after cleaning and optimized splicing, could have:

Loss item	Low-loss estimate
Fiber attenuation	8.36\ \text{dB}
Six connector pairs at 0.20\ \text{dB}	1.20\ \text{dB}
Eighteen splices at 0.04\ \text{dB}	0.72\ \text{dB}
WDM/filter and bend allowance	1.00\ \text{dB}
Total low-loss path	11.28\ \text{dB}

Maximum received optical power is:

Overload margin is:

The receiver is not expected to overload in the baseline configuration. The engineering comment is still important: if a future short reroute, higher-power optic, or optical amplifier is installed, overload must be recalculated. A link can fail because it has too much optical power, not only because it has too little.

Step 4: Screen Chromatic Dispersion

For a first-pass single-mode dispersion check, accumulated chromatic dispersion is:

where D is the fiber dispersion coefficient and L is route length.

Compare this with the transceiver tolerance:

The link passes the stated transceiver tolerance with 154\ \text{ps/nm} of dispersion margin. This is a real constraint, not a paperwork number. If a replacement optic has only 500\ \text{ps/nm} tolerance, the same fiber route would fail the dispersion screen:

That negative margin would require a different optic, dispersion compensation, lower rate, shorter route, different wavelength plan, or a more detailed system test. The acceptance report must therefore record transceiver model and dispersion tolerance, not only the fiber loss.

Step 5: Screen Polarization Mode Dispersion

A simple PMD estimate uses:

where D_{PMD} is the PMD coefficient.

Against a 10\ \text{ps} tolerance, the PMD screen has large margin:

This does not prove every installed fiber is perfect. It shows that PMD is unlikely to be the governing impairment for this route if the cable type and installation condition are as stated. For older cable, long-haul coherent systems, high bit rates, or unusual stress history, PMD evidence should be stronger than a catalogue estimate.

Step 6: Set OLTS Acceptance Limits

An optical loss test set measures end-to-end insertion loss. The OLTS acceptance limit should be tied to the engineering budget, not invented in the field.

For this project, set:

where L_{test} is a small allowance for measurement uncertainty and installation variation. Use 1.0\ \text{dB}:

The measured results are:

Direction	Measured loss at 1550\ \text{nm}	Limit	Result
Site A to Site B	12.7\ \text{dB}	14.2\ \text{dB}	pass
Site B to Site A	12.9\ \text{dB}	14.2\ \text{dB}	pass

Use the worse measured direction for receiver margin:

Measured available margin after the design allowance:

The link passes. The measured margin is slightly better than the conservative design estimate because actual connector and splice losses are lower than their design allowances. The comment to include in the report is that the 3.0\ \text{dB} design margin remains reserved for aging, cleaning uncertainty, repairs, temperature, and future patching. It is not spare capacity for undocumented additions.

Step 7: Use OTDR to Locate Events

OLTS gives the best end-to-end insertion-loss acceptance value, but it does not locate faults. OTDR evidence helps locate splices, connectors, bends, breaks, and reflective events. It should not be used as the only pass/fail measurement for end-to-end service acceptance.

For this project, define these OTDR criteria:

Criterion	Acceptance limit
Individual fusion splice loss	not above 0.15\ \text{dB} without engineering disposition
Mated connector event loss	not above 0.50\ \text{dB} after inspection and cleaning
Unexpected macrobend signature	none accepted without reroute or bend-radius correction
Reflective event at connector	within connector type expectation and not trending upward
End-to-end trace length	consistent with as-built route within documented tolerance

Example OTDR event review:

Event	Distance from Site A	Measured event	Interpretation
Connector at Site A ODF	0.00\ \text{km}	reflective, 0.31\ \text{dB}	pass after inspection and cleaning
Splice closure 1	7.4\ \text{km}	0.06\ \text{dB}	pass
Splice closure 2	15.8\ \text{km}	0.12\ \text{dB}	pass, monitor if future repair occurs nearby
Mid-route bend signature	22.6\ \text{km}	0.18\ \text{dB} wavelength-sensitive loss	hold for field inspection
Splice closure 3	29.1\ \text{km}	0.08\ \text{dB}	pass
Connector at Site B ODF	38.0\ \text{km}	reflective, 0.34\ \text{dB}	pass after inspection and cleaning

The mid-route bend signature is the only exception. It does not make the OLTS result fail, but it is not accepted silently. The project disposition should require a field inspection or a documented engineering waiver before service handover. A small bend loss can grow after temperature cycling, cable movement, or repeated maintenance access.

Step 8: Run Service-Level Validation

Optical tests prove physical-layer condition. The commissioned service still needs traffic and monitoring checks:

Test	Acceptance target
Transceiver digital diagnostics	received power consistent with measured budget and below overload
Bit error or frame error test	no errors during agreed test interval at line rate
Throughput test	meets committed 10 Gbit/s service configuration or documented shaped rate
Latency record	one-way or round-trip method stated, with packet size and load condition
Jitter or packet delay variation	measured under representative traffic load
Alarm verification	loss of light, high/low received power, and port-down alarms reach operations
Label and as-built review	port, fiber pair, route, closures, and patching match handover records

For the baseline route, approximate propagation delay is:

using n=1.468 for fiber and c=3.0\times10^8\ \text{m/s}:

This is only propagation delay through the fiber. The service latency record must also include transceiver processing, switches, forward error correction if used, packet buffering, routing, encryption, and measurement method. The calculation is still useful because it catches impossible field measurements. A reported one-way delay of 20\ \mu\text{s} for this physical route would be physically suspect.

Step 9: Build the Acceptance Decision

Summarize the engineering decision in one table.

Acceptance item	Evidence	Result	Comment
Loss budget	Design loss 13.20\ \text{dB} before margin	pass	Physical model is below optical budget.
Receiver sensitivity	Measured worst direction gives 3.1\ \text{dB} margin after design allowance	pass	Maintenance changes must preserve this margin.
Receiver overload	Maximum estimated received power -7.28\ \text{dBm}	pass	4.28\ \text{dB} below overload threshold.
Chromatic dispersion	646\ \text{ps/nm} against 800\ \text{ps/nm} tolerance	pass	Optic model must be controlled.
PMD	0.62\ \text{ps} against 10\ \text{ps} tolerance	pass	Not governing for baseline assumptions.
OLTS	12.7 and 12.9\ \text{dB} at 1550\ \text{nm}	pass	Worse direction used for margin.
OTDR	One wavelength-sensitive bend signature at 22.6\ \text{km}	conditional	Requires inspection or waiver before final handover.
Service test	Traffic, latency, alarms, and labels not yet attached to the package	hold until complete	Physical acceptance is not the whole service.

The correct decision is conditional acceptance, not unconditional release. The link has adequate optical and dispersion margin, but the OTDR bend signature needs closure and the service tests must be attached to the final package. This is a stronger engineering answer than “link is up.”

Deliverable Structure

The final project package should include:

1. Design basis: service, route, fiber type, wavelength, optics, rate, and acceptance boundary.
2. Loss budget worksheet: assumptions, formulas, units, margins, and measured comparison.
3. Dispersion worksheet: chromatic dispersion, PMD, transceiver tolerance, and replacement-optic constraints.
4. Test plan: OLTS method, OTDR settings, reference cords, wavelengths, connector-cleaning procedure, calibration records, and test direction.
5. Acceptance table: pass, fail, hold, or conditional disposition for every criterion.
6. Exception log: observed bend, high-loss event, dirty connector, missing label, or measurement discrepancy.
7. Operations handover: alarm thresholds, received-power baseline, patching restrictions, spare optic requirements, route records, and retest triggers.

Common Mistakes

Common mistakes include:

• accepting a fiber link because Ethernet comes up;
• using catalogue fiber attenuation but omitting connectors, splices, filters, bends, and repairs;
• checking sensitivity but forgetting receiver overload;
• recording received power without wavelength, reference method, connector condition, or test direction;
• treating OTDR loss as a substitute for OLTS insertion-loss acceptance;
• ignoring dispersion tolerance when replacing optics;
• consuming design margin with undocumented patching;
• failing to close OTDR exceptions because the traffic test happened to pass.

Limits of the Model

This project uses first-pass engineering calculations. It does not replace detailed vendor limits, coherent-system design, nonlinear fiber analysis, high-power optical safety review, multimode launch-condition testing, bidirectional wavelength-specific analysis, or long-duration bit error testing when those are required.

The model assumes the route length, fiber type, attenuation coefficient, component count, connector condition, and transceiver specifications are correct. If any of those assumptions are weak, the acceptance report should say so and require stronger evidence before release.

Engineering Closeout

A defensible closeout statement is:

The fiber link is acceptable for conditional service release after closure of the OTDR bend exception and completion of service-level traffic and alarm tests. The measured optical loss provides approximately 3.1\ \text{dB} margin after the required 3.0\ \text{dB} design allowance. Chromatic dispersion is inside the specified transceiver tolerance, PMD is not governing for the stated route, and receiver overload is unlikely in the baseline configuration. Future optic substitution, route repair, added passive components, or unexplained received-power drift require engineering review and retest.

This is the purpose of the project: convert calculations and field measurements into an acceptance decision that operations can trust.