Skip to content
Send us the packageRequest engineers
Request engineersSend us the package
Company profile (PDF)

Pro tool · for testers, QA and commissioning

Loss budget & OTDR

Work out the maximum loss for a fibre link and whether the optic will run on it. Paste splice results and see what fails on the bidirectional average. Look up what an OTDR event means, how to confirm it and what to fix first.

Fibre
Wavelength
Link length (m)
LC / SC mated pairs
MPO mated pairs
Splices

With a 1-cord reference, count the mated pairs at both test-cord interfaces. A cassette link usually has one LC pair at each panel front and one MPO pair at each cassette rear.

Component limits

0.75 dB per mated pair, 0.75 dB per MPO pair, 0.3 dB per splice. The standard's worst case. Use it when the specification names the standard and nothing stricter.

OM4 · 850 nm · 100 mTIA-568.3 / ISO/IEC 11801-1 maximum

Maximum allowed loss

1.85 dB

Fibre0.35
LC / SC pairs1.50
MPO pairs0.00
Splices0.00

3.5 dB/km: TIA-568.3 / ISO/IEC 11801-1 multimode at 850 nm.

Measured loss (dB)
Will the optic run on it?
ApplicationReachChannel ILHeadroom
10GBASE-SR
IEEE 802.3 cl. 52		400 m2.90 dB+1.05 dB
25GBASE-SR
IEEE 802.3by cl. 112		100 m1.90 dB+0.05 dB
40GBASE-SR4
IEEE 802.3ba cl. 86		150 m1.50 dB-0.35 dB
100GBASE-SR4
IEEE 802.3bm cl. 95		100 m1.90 dB+0.05 dB
400GBASE-SR8
IEEE 802.3cm cl. 138		100 m1.90 dB+0.05 dB
400GBASE-SR4.2 (BiDi)
IEEE 802.3cm cl. 150		100 m1.90 dB+0.05 dB

Headroom = the application’s channel insertion loss minus your designed budget. Negative means the worst-case parts do not fit: specify low-loss components or shorten the link.

Test it right: 850 and 1300 nm, IEC 61280-4-1 with an Encircled Flux launch, 1-cord reference with reference-grade test cords (ISO/IEC 14763-3).

How it works

Rules and tables you can check.

  • Limits you can name

    Connector, splice and fibre limits come from TIA-568.3 and ISO/IEC 11801-1, or from your specification. The page prints which set it used.

  • Will the optic run?

    The designed budget is checked against each application's channel insertion loss from IEEE 802.3, so you see the headroom before you buy parts.

  • Splices judged properly

    Acceptance is on the average of both directions. Gainers and fibres far off their cable-mates are flagged, because that is where a bend hides.

  • Stays on your device

    Pasted results are read in your browser and never sent anywhere.

Straight answers

What engineers ask.

How do you calculate a fibre loss budget?

Add the fibre attenuation (length in km times dB/km), the mated connector pairs times the connector loss, and the splices times the splice loss. TIA-568.3 allows 0.75 dB per mated pair and 0.3 dB per splice; multimode fibre is 3.5 dB/km at 850 nm. For 100 m of OM4 with two mated pairs that is 0.35 + 1.5 = 1.85 dB.

What is an acceptable fusion splice loss?

TIA-568.3 allows up to 0.3 dB per splice. Most project specifications ask for 0.1 to 0.15 dB, measured as the average of OTDR readings in both directions. A splicer's own estimate is not a measurement.

What is a gainer on an OTDR trace?

An event where the trace steps up instead of down. It happens when two fibres with different backscatter are spliced. It is not a gain: test from the other end and take the average of the two readings.

How do I tell a bend from a bad splice?

Test at two wavelengths. A splice loses about the same at 1310 and 1550 nm; a bend loses clearly more at 1550 nm (and more again at 1625 nm).

Free, no sign-up

Need the engineers too?

GO2DataWorks puts cabling and fibre engineers on data-centre sites, or takes the whole package: installed, terminated, tested and handed over.

Request engineers Send us the package

More Pro tools

All Pro tools
Engineers, buyers, site managersCable LensPhoto of the jacket print to a cable cardOpen Splicers, testers, supervisorsFibre Colour Code FinderAny fibre number to tube and colour, any standardOpen Data-centre engineers, designersMPO Polarity PlannerBuild the channel, see every lane landOpen