Use the cable manufacturer’s wavelength-specific attenuation, the specified insertion loss of each mated connection, and the specified loss of each splice when preparing a fiber loss budget. If those specifications are not yet available, use clearly labeled reference values—not an unexplained mixture of typical performance and maximum allowances. Corning’s test guidelines identify cable data sheets as the source for attenuation and provide standards-based allowances as a fallback.
The tables below distinguish planning references, manufacturer specifications, and standards-based allowances. The essential distinction is basis: a typical value estimates expected performance; a maximum specification or agreed acceptance allowance serves a different purpose.
Choose Values That Match the Budget’s Purpose
| Parameter basis | What it represents | Appropriate use |
|---|---|---|
| Manufacturer typical value | Representative performance for the identified product and conditions; not a guaranteed maximum | An expected-loss planning scenario |
| Manufacturer maximum specification | A stated upper limit for the identified product, configuration, and specified conditions | A specified-component design scenario |
| Standards-based allowance | A component or cable allowance associated with a particular standard edition and defined application | An acceptance-budget calculation when adopted by the project |
| Project measurement | A result for a particular installed path, wavelength, and test setup | Comparison with the agreed budget or an established baseline |
| Illustrative assumption | A value selected to demonstrate a method | Preliminary scenario analysis, explicitly labeled as assumed |
FOA distinguishes the cable plant loss budget—the estimated attenuation of the passive path—from the equipment power budget, which describes the loss the communications link can tolerate. A component allowance does not, by itself, establish that an application will operate over the resulting channel.
Keep the expected-loss scenario and the acceptance-budget scenario separate. If the project adopts tighter manufacturer or owner requirements, do not replace them with broader generic allowances merely because those allowances produce a passing result. Corning recommends agreeing on the link-loss budget between the installer and owner before testing begins.
Fiber Attenuation Reference Values
Fiber attenuation is expressed in dB/km at a stated wavelength. Multiply it by the length of the fiber-bearing cable included in the budget; do not enter it as a fixed loss per cable or per connection.
General Planning References
These values come from FOA’s loss-budget discussion. They are preliminary references, not guaranteed performance for every cable carrying the same fiber classification.
| Fiber type | Wavelength | Attenuation reference | Basis and limitation |
|---|---|---|---|
| Multimode | 850 nm | 3.0 dB/km | FOA planning value; replace with the selected cable specification |
| Multimode | 1300 nm | 1.0 dB/km | FOA planning value; not an application reach limit |
| Single-mode | 1310 nm | 0.4 dB/km | FOA planning value; not a universal OS2 maximum |
| Single-mode | 1550 nm | 0.3 dB/km | FOA planning value; product-specific specifications may be lower |
Source: FOA, “Calculating Fiber Optic Loss Budgets,” passive-component calculation.
Select the wavelength before selecting the attenuation coefficient. A 1550 nm value is not a substitute for a 1310 nm value. For a path carrying multiple operating wavelengths, prepare a budget for each relevant wavelength using supported component data. FOA identifies wavelength, fiber type, length, connections, splices, and other passive devices as necessary loss-budget inputs.
Standards-Based Cable Allowances
The following values are reproduced from Table 1 of Corning’s Recommended Fiber Optic Test Guidelines, LAN-1561-AEN, February 2020. Corning identifies them as maximum fiber attenuation values associated with ANSI/TIA-568.3-D. They are presented here as a version-specific manufacturer reference, not as a direct quotation from the standard or a statement of the requirements for every project.
| Cable or fiber category in Corning’s table | Wavelength | Attenuation allowance |
|---|---|---|
| OM1 or OM2 multimode | 850 nm | 3.5 dB/km |
| OM1 or OM2 multimode | 1300 nm | 1.5 dB/km |
| OM3, OM4, or OM5 multimode | 850 nm | 3.0 dB/km |
| OM3, OM4, or OM5 multimode | 1300 nm | 1.5 dB/km |
| OM5 multimode | 953 nm | 2.3 dB/km |
| Single-mode indoor/outdoor cable | 1310 or 1550 nm | 0.5 dB/km |
| Single-mode inside-plant cable | 1310 or 1550 nm | 1.0 dB/km |
| Single-mode outside-plant cable | 1310 or 1550 nm | 0.4 dB/km |
Source: Corning LAN-1561-AEN, Table 1, page 5. The inside-plant, indoor/outdoor, and outside-plant distinctions belong to that table and should not be discarded when copying its values.
Confirm the standard edition and test requirements adopted by the contract before using these allowances for acceptance. Do not label an allowance simply “TIA maximum” without identifying its source, edition, fiber category, and installation context.
Manufacturer Typical and Maximum Values Are Different Inputs
A product data sheet may provide both typical and maximum attenuation for the same cable option.
For example, Corning’s ALTOS all-dielectric cable family specification, revision October 15, 2017, lists the following for the SMF-28 Ultra option with fiber code Z and performance option 22:
| Wavelength | Typical attenuation | Maximum attenuation |
|---|---|---|
| 1310 nm | 0.32 dB/km | 0.34 dB/km |
| 1550 nm | 0.18 dB/km | 0.22 dB/km |
These values illustrate the distinction within a specific manufacturer document. They are not specifications for every ALTOS cable, every SMF-28 Ultra cable, or every OS2 product; confirm the ordered configuration and applicable data-sheet revision.
Also distinguish an optical fiber specification from a finished cable specification. Corning’s July 2025 SMF-28 Ultra optical fiber sheet lists maximum attenuation of 0.32 dB/km at 1310 nm and 0.18 dB/km at 1550 nm. Those are specifications for the fiber product described in that sheet; do not silently substitute them for the attenuation specified for a finished cable assembly.
Connector Loss Reference Values
Budget connection loss per mated pair, per optical path—not per loose connector housing. Two connectors joined through an adapter form one connection. FOA explicitly distinguishes a connector from the connection whose insertion loss is measured.
| Connection or termination type | Reference insertion loss | Nature of the value |
|---|---|---|
| Conventional multimode connection | 0.2–0.5 dB per mated pair | FOA typical range |
| Factory-made single-mode connector termination using a fusion-spliced pigtail | 0.1–0.2 dB per mated pair | FOA typical connection-loss range; account separately for the pigtail splice |
| Generic connector-pair allowance in Corning’s TIA-568.3-D worksheet | 0.75 dB per mated pair | Version-specific allowance, not typical performance |
| MPO/MTP or a preterminated modular assembly | Use the exact product or assembly specification | Do not substitute a generic single-fiber connector value |
FOA provides the typical connection-loss ranges. Corning’s LAN-1561-AEN worksheet uses 0.75 dB per connector pair and directs users of preterminated modular systems to product-specific budgeting guidance.
An LC or SC designation alone is not enough to select insertion loss. Obtain the manufacturer’s specification for the actual assembly and mating conditions. Likewise, use insertion loss—not return loss or reflectance—as the connection-loss input in the attenuation sum. FOA’s budget method uses the optical power lost through each connection.
Count Connections Along One Fiber Path
Trace one optical path from the selected starting reference plane to the ending reference plane.
- Count each mated interface encountered by that path once.
- Do not double the count because each connection contains two connector ends.
- Do not multiply loss by the number of fibers in a duplex or multifiber cable; calculate each relevant optical path.
- Include equipment patch cords and their connections only when they lie within the boundary being evaluated.
FOA shows that a cable plant budget changes when equipment patch cords are included or excluded. It also explains that the test reference method affects which end connections appear in the measured insertion loss.
For an installed link tested with a one-jumper reference, Corning states that the measurement includes the installed system and both end connection pairs. Other reference arrangements have different boundaries. Match the budget to the documented test method rather than assuming that every test includes the same interfaces.
Splice Loss Reference Values
Splice loss is expressed in dB per splice traversed by the optical path. The number of splice closures is not the splice count: count the actual joints on the selected fiber. Corning’s worksheet requires the number of splices in the installed system, including separate pigtail splices.
| Splice type or budgeting basis | Reference loss | Interpretation |
|---|---|---|
| Single-mode fusion splice | Typically less than 0.05 dB | FOA typical-performance reference; not a guaranteed field result |
| Multimode splice | 0.1–0.5 dB | FOA reference range; not an acceptance band |
| Generic splice allowance in Corning’s TIA-568.3-D worksheet | 0.3 dB per splice | Version-specific allowance used for mechanical or fusion splices |
Sources: FOA’s splice-loss discussion and Corning LAN-1561-AEN, page 6. FOA’s multimode range does not establish one universal value for every mechanical or fusion-splice product.
For a selected mechanical-splice product, use its specified insertion loss and conditions. For fusion splices, use the project’s agreed design and acceptance criteria rather than treating a favorable typical value as a promise.
Avoid Counting an Integrated Termination Twice
A separate pigtail fusion splice and its connector mating interface are separate loss contributions. Include both when both are within the budget boundary.
An internal splice may already be included in a connector assembly’s insertion-loss specification. Corning specifically instructs users not to add a separate mechanical-splice allowance for the internal splice in its UniCam or OptiSnap connectors; those terminations are counted as connectors. Apply the same boundary check to other products using their own documentation.
Carry the Reference Values Into a Loss Budget
For a passive path containing multiple cable sections and component types:
=
\sum_i \alpha_i(\lambda)L_i
+
\sum_j C_j(\lambda)
+
\sum_k S_k(\lambda)
+
\sum_q D_q(\lambda)
where:
- (A_{\text{plant}}) is estimated passive-path insertion loss, in dB.
- (\alpha_i) is attenuation of cable section (i), in dB/km, at wavelength (\lambda).
- (L_i) is the included length of that section, in km.
- (C_j) is insertion loss of mated connection (j), in dB.
- (S_k) is insertion loss of splice (k), in dB.
- (D_q) is insertion loss of another included passive device, in dB.
This is the component-sum method described by FOA and used in Corning’s link-loss worksheet, with a separate term for additional passive devices. FOA specifically calls out splitters and couplers as components that must not be omitted.
Use the installed or planned cable length within the chosen boundary, including retained slack. A geographic route distance alone is not a complete length input when additional cable is installed. Corning identifies installed cable length, obtained from sheath markings or suitable test records, as a required budgeting input.
For PON splitters, filters, or other passive assemblies, obtain the insertion loss for the actual model, wavelength, and path through the device. Check whether its specification includes connector losses before adding separate connection terms. A fiber-plus-connections-plus-splices subtotal is incomplete when such devices lie in the path.
Example: Typical-Performance Assumptions Versus Component Allowances
Consider one illustrative single-mode path at 1310 nm with:
- 10.0 km of cable within the budget boundary.
- Two mated connection pairs.
- Eight fusion splices, including any separate pigtail splices.
- No splitters or other passive devices.
- No design reserve included in the passive-loss subtotal.
The first scenario assumes 0.4 dB/km, 0.2 dB per connection, and 0.05 dB per splice. These are illustrative inputs informed by FOA’s planning discussion; using 0.05 dB as an assumption does not make it a guaranteed splice maximum. The second scenario retains the same fiber coefficient but uses the 0.75 dB connection and 0.3 dB splice allowances in Corning’s TIA-568.3-D worksheet.
| Contribution | Illustrative planning scenario | Allowance-based scenario |
|---|---|---|
| Fiber | (10.0 \times 0.4 = 4.0) dB | (10.0 \times 0.4 = 4.0) dB |
| Connections | (2 \times 0.2 = 0.4) dB | (2 \times 0.75 = 1.5) dB |
| Splices | (8 \times 0.05 = 0.4) dB | (8 \times 0.3 = 2.4) dB |
| Calculated passive-path loss | 4.8 dB | 7.9 dB |
The 3.1 dB difference comes entirely from the connection and splice assumptions. The results are two budgeting scenarios, not a predicted measurement interval or proof that either value is an appropriate acceptance limit.
Keep Design Reserve Separate From Component Loss
A design reserve is not a measured connector, splice, or cable loss. If the project reserves (M) dB for defined future changes or degradation, keep it visible:
Here, (A_{\text{channel,max}}) is the allowable channel insertion loss for the selected application and matching reference planes. Use the equipment or application specification for that limit. FOA explains that transmitter-to-receiver power differences alone may not capture dispersion and other system penalties, and that some short-reach applications have very small loss allowances.
Do not add a generic reserve automatically or count one twice if the supplied system limit already incorporates it. A favorable attenuation comparison also does not establish compliance with bandwidth, dispersion, reflectance, or other application requirements.
Record the Source With Every Copied Parameter
A usable reference entry contains more than a number. Preserve the following information when transferring values into a worksheet or calculator:
| Field | What to record |
|---|---|
| Component identity | Cable part number and option, connector assembly, splice type, or passive-device model |
| Wavelength | The operating or test wavelength to which the value applies |
| Value and unit | dB/km for cable attenuation; dB per mated pair, splice, or device |
| Value basis | Typical, manufacturer maximum, adopted allowance, measurement, or assumption |
| Source | Document title, revision, and table or page where available |
| Boundary | Whether patch cords, end connections, pigtail splices, and integral device connections are included |
| Test context, if measured | Fiber ID, wavelength, direction, reference method, instrument, and date |
Corning’s documentation guidance includes cable and fiber specifications, route and splice records, reference-jumper details, test procedures, fiber IDs, and link-loss results. Those records make subsequent comparisons traceable.
When an installed result is close to its agreed limit, review the test setup and measurement uncertainty before assigning a pass or fail from a rounded display value. Use the specified insertion-loss test method for end-to-end attenuation; Corning notes that an OTDR provides complementary event information but does not replace OLTS or light-source-and-power-meter testing.