Fiber Optic Testing: Loss Budget Calculation, OLTS and OTDR Measurement

Compare the fiber line’s total loss with the acceptance limit. If the loss is high, use the OTDR trace to investigate where on the line the problem occurs: at a connection, a splice or along the cable.
Technical summary
Technology
ANSI/TIA-568.3-D, ISO/IEC 14763-3, IEC 61300-3-35, OLTS, OTDR, OM4, OS2
Topic
Fiber line acceptance testing and the loss budget
Level
Intermediate–advanced
Prerequisites
  • Knowledge of fiber types (multimode / single-mode)
  • Connector and splice concepts
In this article8 sections

The fiber line has been pulled, the modules have been inserted, the link is up. That isn’t enough for acceptance. A backbone carrying 10G today may not come up when moving to 40G/100G because the loss budget isn’t enough; if you have no measurements that day, you can’t tell whether the problem is in the cable, the splices or a dirty connector. Below we explain the two levels of fiber acceptance testing (Tier 1 OLTS and Tier 2 OTDR), a worked loss budget calculation, reference methods, reading the OTDR trace and connector contamination, the leading cause of failed tests in the field. If the campus backbone will also carry remote security cameras, calculate recording traffic and endpoints separately for the fiber optic camera network infrastructure.

Fiber test levels: Tier 1 and Tier 2

ANSI/TIA-568.3-D and ISO/IEC 14763-3 define how fiber cabling is tested in the field at two levels:

Level Instrument What does it measure? Result
Tier 1 (basic) OLTS: light source + power meter End-to-end total loss, length, polarity The primary value for acceptance
Tier 2 (extended) OTDR The position and loss of each event along the line Proof of installation quality, fault location

The decisive measurement for loss acceptance is OLTS. The total loss calculated by OTDR is based on a backscatter estimate and doesn’t replace the OLTS value. Tier 2 isn’t an alternative to Tier 1 but a complement: OLTS answers the question “is the line within budget?”, OTDR the question “where is the loss occurring?” On long campus backbones in an enterprise network infrastructure, on lines with splices and in projects that require a warranty certificate, ask for both together. If one is missing, the report is incomplete too.

Measurement wavelengths: 850 nm and 1300 nm on multimode (OM3/OM4/OM5) fiber; 1310 nm and 1550 nm on single-mode (OS2) fiber. Measuring at two wavelengths reveals bend-related losses, especially in single-mode. A report closed with a single wavelength can hide a pinched patch cord.

Loss budget components and a worked calculation

The loss budget is the highest loss a link may show according to standard component limits. It is the sum of three components:

  • Cable attenuation: Length (km) × attenuation coefficient (dB/km). The TIA-568.3-D upper limits are, for multimode fiber, 3.0 dB/km at 850 nm (3.5 dB/km in ISO/IEC 11801 and older TIA revisions) and 1.5 dB/km at 1300 nm; for single-mode fiber, 1.0 dB/km for indoor cable and 0.5 dB/km for outdoor cable.
  • Connector pair loss: At most 0.75 dB for each mated connector pair.
  • Splice loss: At most 0.3 dB for each fusion or mechanical splice.

Worked example: 150 m of OM4 backbone between two floor distribution rooms, a patch panel at each end (2 connector pairs), fusion splices with pigtails on the panel side (2 splices), at 850 nm:

Cable     : 0.150 km × 3.0 dB/km = 0.45 dB
Connector : 2 × 0.75 dB         = 1.50 dB
Splice    : 2 × 0.3 dB          = 0.60 dB
-------------------------------------------
Budget (850 nm)                 = 2.55 dB

If the measured loss is below 2.55 dB, the link passes against the general TIA limit. Don’t miss two points. First, the standard limit is an upper limit; in a properly done installation, this link’s measurement usually comes out noticeably below the budget. A measurement approaching the budget doesn’t mean “passed” but “there’s a problem somewhere”: a dirty connector or a bad splice. Second, the application’s own channel loss limit may be lower than the general limit. Channel budgets for 40G/100G parallel optic applications are tight; the value should be taken from the relevant clause of the IEEE 802.3 standard and the datasheet of the optical module to be used. Also check that the limit set selected on the test instrument (TIA or ISO) is the same as the limit in the specification. The right thing is to write both the general limit and the target application’s budget into the specification as test criteria.

Reducing the number of connectors and splices directly eases the budget. How much loss each connector family brings is covered in the choosing LC, SC and MPO connectors article.

Measuring with OLTS and the reference method

OLTS works on a “take a reference, then measure” logic. The source power is zeroed as a reference with the test cords, the link is inserted in between, and the difference is read as loss. The reference method determines which connections are included in the result:

Method Test cords used in referencing End connections included in the result
1 jumper One Both end connections included
2 jumper Two One end connection included
3 jumper Three End connections excluded

For permanent link testing, the method preferred by TIA-568.3-D and ISO/IEC 14763-3 is the 1-jumper reference, because it also measures the connector loss at both ends of the link. The 2- or 3-jumper method is used when the connector type of the instrument ports differs from the connector in the link; in that case the same number of connectors must also be deducted from the budget calculation. The method must be written in the report. If it isn’t, the same link appears with two different values in two teams’ reports and nobody can tell who is right.

Points to watch in the field:

  • After the reference is taken, the test cord isn’t unplugged from the source port; if it is, the reference is taken again.
  • In multimode measurement, the source’s modal distribution affects the result. An Encircled Flux (EF) compliant source or an EF-compliant test cord gives repeatable results across different instruments.
  • Test cords (TRCs) must be high quality and checked; a worn test cord produces a systematic error in the same direction on all links.
  • If negative loss (gain) is read, the reference is wrong. Stop the test and renew the reference.

Reading the OTDR trace: reflective, splice and break events

An OTDR sends pulses into the fiber and measures backscattered and reflected light over time; the result is a trace showing power level against distance. The sloping straight line in the trace is the fiber itself; sudden changes are events.

  • Reflective event: A peak followed by a step. It means a connector or a mechanical splice. The peak height gives information about reflectance, and the step height about loss.
  • Non-reflective event: A step without a peak. A fusion splice or a macrobend. If the loss at the same point is noticeably higher at 1550 nm than at 1310 nm, the event is most likely a bend; look for a fiber pinched in a cassette or an overtightened cable tie.
  • Fiber end or break: A large reflection peak followed by a drop to the noise floor. If it appears before the expected length, the line is broken or crushed at that point.
  • Splice showing a gain (gainer): At a point where two fibers with different backscatter characteristics are spliced, the splice can look like a gain in one direction. The correct splice loss is found by averaging the two-way measurement.
  • Ghost: A repeat of strong reflections at a distance in the trace that isn’t real. It usually appears after the fiber end and contains no loss step. Don’t chase it.

Pulse width is a trade-off between resolution and range: a short pulse distinguishes nearby events but gives a noisy trace on long lines. For lines at building and campus scale, a short pulse and sufficient averaging time are usually the right setting.

Why are launch and receive fibers needed?

The strong reflection at the OTDR’s own port briefly saturates the detector; this distance is called the dead zone. The event dead zone is the shortest distance at which two reflective events can be distinguished; the attenuation dead zone is the shortest distance at which an event’s loss can be measured. If the instrument is connected directly to the link, the first connector falls into this zone. It can’t be measured.

A launch fiber (launch reel), inserted between the instrument and the link, moves the first connector out of the dead zone. A receive fiber (tail reel), added at the end of the link, makes the loss of the last connector measurable; without it, the last connector blends into the fiber-end reflection. The reel length is chosen per the manufacturer’s recommendation based on the instrument’s dead zone value and the fiber type; the fiber type and connector must be the same as the link under test. An OTDR trace taken without reels hasn’t measured the two points where most faults occur: the end connectors.

Connector contamination: inspection and cleaning

One of the most common causes of failed fiber tests and fiber faults in the field is dust, oil and scratches on the connector end face. A dirty connector doesn’t just increase loss; it also carries dirt to the mating connector, and with high-power optics it can leave permanent damage on the end face. The rule is one sentence: inspect first, then connect.

  1. Check the connector and the mating connector inside the adapter with a fiber microscope or an inspection probe with automatic analysis. The acceptance criteria are defined by zone in IEC 61300-3-35.
  2. If it’s dirty, apply dry cleaning first (a cassette-type cleaner, or a pen-type cleaner for inside the adapter).
  3. If dry cleaning isn’t enough, move to the wet–dry method: wipe in one direction from an area moistened with a suitable solvent to a dry area.
  4. Inspect again. If it isn’t clean, don’t connect it.
  5. Remove dust caps only at the moment of connection. A used dust cap isn’t clean.

Fiber acceptance test checklist

  • The test level (Tier 1 / Tier 1 + Tier 2) is defined in the specification.
  • A budget calculation has been done for each link; both the general limit and the target application limit are known.
  • The OLTS measurement is at two wavelengths, and the reference method is written in the report.
  • An EF-compliant source or test cord was used for multimode measurement.
  • Launch and receive fibers were used for OTDR traces; splice losses were averaged in both directions.
  • Polarity (Tx–Rx matching) has been verified on every duplex link.
  • All connectors were inspected before connection; end face images are on record.
  • Link IDs match the labels and the as-built documents.

If the fiber type (OM3, OM4, OS2) hasn’t been chosen yet, start with the choosing Cat6, Cat6A and fiber article; for fiber count and route decisions on inter-building lines, see the campus backbone fiber design article. Test parameters for copper links are in the cable certification testing article. For teams that want fiber backbone installation, termination and acceptance measurements handled by a single provider, see our structured cabling and low-voltage infrastructure service page.

Frequently asked questions

Why is an OLTS test needed when there is an OTDR measurement?

OTDR estimates total loss from backscatter and can’t always measure the end connectors accurately. The end-to-end loss value used for acceptance is measured with OLTS. OTDR is the complementary test that shows where on the line the loss occurs.

What value should be used for connector loss in the loss budget calculation?

For the general acceptance limit, the upper value of 0.75 dB per connector pair in TIA-568.3-D is used. In design, a more realistic budget is drawn up with the value from the datasheet of the component to be used, and compared with the target application’s channel limit.

What does it mean to see negative loss in a test result?

In OLTS, negative loss shows that the reference was taken incorrectly; e.g. the test cord was dirty during referencing, or the source-side connection was unplugged after referencing. Clean the connections and take the reference again. In OTDR, a one-way gain is a measurement effect seen at points where different fibers are spliced, and it is corrected by two-way averaging.

Are alcohol and a tissue enough for cleaning connectors?

No. General-purpose tissues leave lint and residue, and the wrong solvent can form a film on the end face. Dry cleaners made for fiber, pen-type adapter cleaners and the wet–dry method with a suitable solvent are used; the result is always verified with an inspection probe.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Network

We build switching, routing, cabling and wireless networks with VLAN segmentation and redundant uplinks, and provide regular maintenance and fault support.

Knowledge Center

Other articles