How to Test Fibre Insertion Loss with a Light Source and Power Meter (OLTS Guide)
An OTDR trace shows where the events are, but the number most specifications ask for when a fibre link is handed over is simpler: the insertion loss of each fibre, measured end to end with a light source and an optical power meter. This test, carried out with an optical loss test set (OLTS) or a separate source and meter, is the closest match to what the transmission equipment will actually experience. This guide explains how the test works, the 1-, 2- and 3-cord reference methods (and why they give different answers), a step-by-step procedure, how to decide pass or fail, and what to do when a reading looks wrong.
What insertion loss testing measures
Insertion loss is the drop in optical power, in dB, when a link is inserted between a light source and a power meter. A stable source sends light through a launch reference cord into one end of the link. At the far end, a receive reference cord carries the light to the power meter. The reading includes:
- the connection at each end of the link, where the reference cords mate with the patch panel or connectors;
- the attenuation of the fibre itself;
- every splice, intermediate connection and, on FTTH networks, any PLC splitter in the path.
This is the method described in the Fiber Optic Association's FOA-1 standard, TIA OFSTP-14 (multimode) and OFSTP-7 (single-mode), and IEC 61280-4-1 and 61280-4-2. In TIA-568 cabling terms it is often called Tier 1 testing; adding OTDR traces is Tier 2.
Equipment you need
- A light source and a power meter at the right wavelengths: 1310 and 1550 nm for single-mode (1625 nm is sometimes added), and 850 nm (often with 1300 nm) for multimode. The EXFO FOT-300 optical loss test set combines a source and meter in one unit, and a pair of them lets you test from both ends. A separate meter such as the EXFO FPM-600 or the Bluetooth-connected EXFO Optical Power Expert PX1 can be paired with a source.
- Reference cords 1 to 2 m long, of the same fibre type and connector polish as the link. For single-mode UPC links an SC/UPC OS2 patch cord is typical, for APC links an SC/APC OS2 cord, and a hybrid such as an LC/UPC to SC/UPC cord adapts an SC tester port to an LC patch panel. See the full range of fibre patch cords.
- Good mating adapters with ceramic or metal alignment sleeves, not plastic.
- Inspection and cleaning tools. A probe such as the EXFO FIP-400B Wireless lets you check every end face before it is mated.
The three ways to set the 0 dB reference
Before testing, the power meter must be “zeroed” to the output of the launch cord, so that it shows only the loss added by the link. All the standards allow three ways of doing this, using one, two or three reference cords. The choice depends mainly on whether the tester's ports can take the same connectors as the link.
| Reference method | How the 0 dB reference is set | When to use it | Effect on the result |
|---|---|---|---|
| 1-cord (formerly Method B) | Launch cord from the source plugged straight into the power meter | Tester ports accept the same connector type as the link | Measures both end connections plus everything in between. Lowest uncertainty: the preferred method in most cabling standards |
| 2-cord (formerly Method A) | Launch cord and receive cord mated through one adapter | Tester connectors do not match the link, but a single mated pair can be made | Reads lower than the 1-cord method by roughly the loss of the one reference connection |
| 3-cord (formerly Method C) | Launch cord, a short third cord and receive cord, through two adapters | Plug-and-jack connectors such as MPO, or when the tester cannot accept the link connector | Reads lower again, by roughly the loss of two reference connections. Highest uncertainty |
A common misunderstanding is that the 2-cord method “leaves out one connector” and the 3-cord method “leaves out both”. In fact every method measures all the connections in the link. What changes is that the loss of the connections made while setting the reference is subtracted from the result, and those connections are never exactly the same as the ones in the actual test. That is why each extra cord adds uncertainty. The FOA tested one cable plant all three ways under controlled conditions and reported:
| Reference method | Measured loss | Standard deviation |
|---|---|---|
| 1-cord | 2.96 dB | ±0.02 dB |
| 2-cord | 2.66 dB | ±0.20 dB |
| 3-cord | 2.48 dB | ±0.24 dB |
The lesson is simple: use the 1-cord reference whenever your equipment allows it, and always write down the method used with the results. A loss value is only comparable to another test, or to a specification, when the reference method is known. Some network specifications name a particular method, so check before you start.
Step-by-step: how to test insertion loss
Step 1: Prepare the equipment
Switch the source on and let it warm up until its output is stable to within about 0.1 dB. Select the same wavelength on the source and the meter. On multimode, follow the launch condition your standard asks for (a mandrel wrap or an encircled flux compliant launch), because the result depends strongly on how the light is launched.
Step 2: Inspect and clean
Inspect and clean every connector before mating it: reference cords, tester ports and both ends of the link. A single dirty end face can add tenths of a dB or more, and can transfer dirt to the patch panel. Our guide on how to inspect and clean fibre optic connectors explains the routine.
Step 3: Check the reference cords
Mate the launch and receive cords together and measure the connection. Reference-quality cords should be about 0.2 dB or better on single-mode and 0.1 dB on multimode. Ordinary good patch cords are usually under 0.3 dB. Replace any cord that reads more than about 0.5 dB.
Step 4: Set the 0 dB reference
For the 1-cord method, plug the launch cord into the power meter and store the reading as the reference. From now on, do not disconnect the launch cord from the source. Removing it and plugging it back changes the launch power and invalidates the reference.
Step 5: Connect the link
Connect the launch cord to one end of the link and the receive cord, attached to the power meter, to the other end. On long routes the source and meter are at different sites, so agree the fibre numbering by phone before starting, following the TIA-598 colour code.
Step 6: Measure and record
Read the loss at each wavelength and record it with the fibre number, direction, wavelength, reference method and date. Many specifications ask for results in both directions. Typical measurement uncertainty is about ±10% of the reading, so small differences between fibres are normal.
Step 7: Re-check the reference
At the end of the job, or every few hours, connect the launch cord back to the meter and confirm the reference has not drifted. If it has, set it again and retest the fibres measured since the last check.
Safety: only test dark fibres. Never connect a light source to a fibre carrying live traffic, and never look into a fibre or connector. To measure the signal on a live FTTH line, use a PON power meter instead; see our guide to testing FTTH signal levels with a PON power meter.
Pass or fail: compare with the loss budget
A measured loss means little on its own. Compare it with the expected loss calculated from the link design, as explained in our guide on how to calculate a fibre optic loss budget. TIA-568 sets maximum values of 0.75 dB per mated connection and 0.3 dB per splice, and up to 0.5 dB/km for outside plant single-mode cable at 1310 and 1550 nm. The example below uses the design values from that guide:
| Item (example 10 km single-mode link) | Expected loss using design values | Maximum allowed by TIA-568 |
|---|---|---|
| Fibre at 1310 nm (10 km) | 10 × 0.4 = 4.0 dB | 10 × 0.5 = 5.0 dB |
| 2 end connections (patch panels) | 2 × 0.3 = 0.6 dB | 2 × 0.75 = 1.5 dB |
| 4 fusion splices | 4 × 0.1 = 0.4 dB | 4 × 0.3 = 1.2 dB |
| Total at 1310 nm | 5.0 dB | 7.7 dB |
If this link measures 4.6 dB at 1310 nm, it is close to the expected 5.0 dB and well under the 7.7 dB maximum: a pass. A reading of 7.0 dB would still be inside the maximum, but it is 2 dB more than expected, so something along the route deserves a closer look. Many installers treat a result well above the design value as a fault to investigate, even when the standard's maximum is not exceeded. Litech's own G.652.D duct cable, for example, has a cabled attenuation of ≤ 0.35 dB/km at 1310 nm and ≤ 0.21 dB/km at 1550 nm on its datasheet, comfortably inside the design figures above.
Single-mode G.652.D fibre normally shows less loss at 1550 nm than at 1310 nm. If a link reads higher at 1550 nm, suspect a bend.
Troubleshooting suspicious readings
| What you see | Likely cause | What to do |
|---|---|---|
| Loss far above the expected value on one fibre | Dirty or damaged connector, bad splice or a tight bend | Inspect and clean both ends, retest, then use an OTDR to find the event |
| Loss much higher at 1550 nm than at 1310 nm | Macrobend or crush point along the route | Check slack loops, closures, ties and duct entries |
| Negative loss (a “gain”) | Reference was set with a dirty or poorly mated connection, or the launch cord was moved after referencing | Clean everything, set a new reference and retest |
| Results drift while testing | Source not warmed up, or the launch cord was disturbed | Let the source stabilise, re-reference, keep the launch cord still |
| Very high loss on an APC link | UPC cord mated to an APC connector (or the reverse) | Use reference cords with the same polish as the link |
| Same high loss on every fibre | Faulty reference cord | Check the cords against each other and replace any that read over 0.5 dB |
A light source and power meter tell you how much loss there is, but not where. When a fibre fails, an OTDR such as the EXFO MaxTester 715B shows the distance to the problem; see our OTDR testing guide. For quick checks by field teams, the EXFO Optical Explorer OX1 measures loss, ORL and length in one instrument. For a single patch cord, a single-ended test from source to meter works: test it, reverse it and test again to find which end is bad.
Key takeaway: warm up the source, clean and inspect every connector, check your reference cords, set the 0 dB reference with the 1-cord method whenever you can, never disconnect the launch cord afterwards, and record the method with every result. Then compare each reading with the expected loss from your loss budget and test at both 1310 and 1550 nm on single-mode.
Frequently asked questions
What is an OLTS?
An optical loss test set is a light source and a power meter, often in one instrument, used to measure the insertion loss of a fibre link. It is the main tool for Tier 1 acceptance testing.
Which reference method should I use?
Use the 1-cord method when the tester accepts the link's connectors, because it has the lowest uncertainty. Use the 2- or 3-cord method only when the connectors do not match, and always record which method you used.
Why did my power meter show a negative loss?
A link cannot add power. A negative reading means the reference was wrong, usually because a dirty connector was cleaned after referencing, or the launch cord was moved or reconnected. Clean everything, reset the reference and retest.
Which wavelengths should I test?
For single-mode, test at 1310 and 1550 nm; a higher reading at 1550 nm points to bends. For multimode, test at 850 nm and, where specified, 1300 nm.
Do I still need an OTDR if I have an OLTS?
The OLTS gives the most direct measure of end-to-end loss. The OTDR shows the individual splices, connections and faults along the route and their distances. Long outside plant routes are usually documented with both.
Can I test through a PLC splitter?
Yes. The splitter loss is included in the reading, so add it to the expected loss for the link. Use the splitter datasheet value, which is around 17 dB for a 1:32 splitter.
How long should reference cords be?
About 1 to 2 m, so the loss of the fibre inside them can be ignored. They must use the same fibre type and connector polish as the link.
Need test equipment or reference cords?
Litech Pakistan (Private) Limited supplies EXFO test equipment, including loss test sets, power meters, inspection probes and OTDRs, along with fibre patch cords for reference and test use. We also manufacture single-mode optical fibre cables in Pakistan. Message us on WhatsApp +92 313 7776766 with your link type and test requirements for a quotation.
Sources: FOA-1, TIA OFSTP-14 and OFSTP-7 (insertion loss of installed fibre cable plant); IEC 61280-4-1 and IEC 61280-4-2 (attenuation measurement of installed multimode and single-mode cabling); ANSI/TIA-568.3-D (component and cable attenuation limits); Fiber Optic Association, “The Math of Insertion Loss” and insertion loss testing guidance; EXFO product information; Litech cable datasheets.
