As an Amazon Associate we earn from qualifying purchases. Details

How to Use a Fiber Optic Light Source

A fiber optic light source sends a known level of light into one end of a fiber link. On its own it does not tell you anything about the link. It becomes useful when a power meter sits at the other end, because the difference between the two readings is the loss of the fiber, its splices, and its connectors. This guide walks through the setup steps that most loss-test procedures share, and points out the places where people usually get the result wrong.

Quick answer

To use a fiber optic light source for a loss test, match the wavelength and fiber type, clean every connector, set the reference with a short test cord, then remove the reference and measure the link under test. Keysight describes the method as a two-step process: connect the source to one end of the link and the meter to the other, measure, then connect the source directly to the meter and measure again. The difference is the net loss of the link.

The rest of this guide explains each step, what the readings mean, and how to check that the result is trustworthy.

What a light source does

A light source is the transmitter in a loss test. It launches light at a set wavelength into the fiber, and a power meter measures what comes out at the far end. Keysight’s article on optical power meters describes the pair in exactly these terms: the source is connected to one end of the link and the meter to the other, and the meter is set to the same wavelength as the source. Without the source, the meter has nothing to measure in a controlled way.

Many meters and sources are sold together as a kit. Fluke’s CertiFiber Pro, for example, has a power meter and light source function in one handheld unit. Fluke’s guidance describes how the user taps Tools, then Power Meter/Light Source, and then selects the fiber type, MM or SM. The same logic applies to any kit: the source sends, the meter reads, and the reference step removes the test cords from the result.

Before you start

• Confirm the fiber type. Select multimode or single-mode on the source and meter so they match the link. Fluke’s CertiFiber guidance asks for this choice before the test.
• Match the wavelength. Keysight notes that 850 nm is used for short-distance applications such as local area networks, and 1,550 nm for long-distance applications such as undersea cable. Fluke’s examples include 850 nm and 1300 nm for power readings on a CertiFiber. Set the source and meter to the same wavelength for the link you are testing.
• Clean and inspect connectors. Fluke’s CertiFiber procedure begins with cleaning and inspecting all connectors. A dirty ferrule adds loss that looks like a fault in the fiber, so clean both ends of every cord and of the link.
• Use the right reference cords. The reference cords connect the source and the meter. Keep them short and in good condition. A damaged reference cord will add loss to the reference and skew the result.

Step 1: Power the units and let them settle

Turn on the source and the meter and let them stabilise before you set anything. Fluke’s bidirectional procedure starts from a reference-grade launch cord, and a stable output is needed for a repeatable reference. If the reading drifts, give the source more time before you continue.

Step 2: Set the reference (0 dB)

Connect the source output to the meter input with a short reference cord. Select the same wavelength on both units, then use the REF function to store the current power as the baseline. On a CertiFiber Pro, Fluke’s knowledge base says to tap REF first to zero out the reference cable. Once the reference is stored, the meter shows the loss relative to that baseline, not the absolute power.

Fluke’s bidirectional guidance describes the same idea using a one-cord reference method. The launch cord from the source is kept in place for the whole test, which keeps the reference stable. If the source output is disturbed or disconnected, the reference becomes invalid. Fluke’s CertiFiber article says to reset it before continuing.

Step 3: Measure the link under test

Remove the reference cord and insert the link under test between the source and the meter. Clean its connectors the same way as before. The relative reading on the meter is the loss of the link. Record the value, the wavelength, the fiber type, and the direction of the test.

Keysight’s two-step method is the same logic in a different form. First measure the link with the source at one end and the meter at the other. Then connect the source directly to the meter and measure again. “The difference between these readings is the net power loss of that fiber link.” This gives the same answer as a reference step, and it is a good way to check your understanding of what the reference is doing.

Step 4: Test the other direction on complex links

Fluke’s guidance says bidirectional measurements are needed when a link is more complex, or when components may cause loss to differ by direction. In the procedure, the first direction (X-Y) is measured without disconnecting the launch cord. Then the receive cord and meter move to the other end and the reverse direction (Y-X) is measured.

Fluke’s threshold is a useful check. If the two directions differ by more than 0.2 dB, the reference may be wrong, the test cord may be faulty, or the fiber may be bent. The worse of the two results is the one to report. This check takes only a few minutes and catches a large share of setup errors.

Step 5: Read and save the result

Fluke’s CertiFiber guidance says the user taps POWER to switch between dBm power readings and LOSS readings in dB, and taps SAVE to store the result. Keep the saved file or note with the link name, the wavelength, and the direction. A loss number with no record of its setup is hard to use later.

The result has no pass or fail limit by default. Fluke’s guidance says that if the project needs pass/fail results, the user should create a custom fiber test limit. Set the limit from the cable or project specification, not from a guess.

Reading the results

A single high loss reading tells you the link is not performing as expected. It does not tell you where the problem sits. Fluke’s guidance gives a rough pattern to look for. Loss on every fiber in a cable suggests the cable is damaged or kinked. Loss on a single fiber more often points to a bad splice or connector.

If the reading is unexpected, check the setup before you blame the link. Confirm the reference was set with the same wavelength and fiber type. Look at the connectors, and clean them again. Then run the reverse direction. A large difference between directions points to a setup error or a bent fiber, not to a clean link with a fault.

Common mistakes

The most frequent mistake is setting the reference at one wavelength and measuring at another. The second is skipping cleaning, which adds loss that looks like a real problem. The third is moving the source after the reference is set. The fourth is assuming a single direction is enough on a link with more than one component. Each of these produces a number that looks valid but is not.

When a light source and power meter cannot find the fault, the next tool is an OTDR. An OTDR can locate events along the fiber, while a light source and meter only measure the total loss between two ends. Keysight’s power meter article notes that an OTDR can help detect fiber breakages, loose connections, and faulty or dirty connectors or couplers.

What we could not verify

• The exact menu names, button paths, and reference steps on a light source from a brand other than Fluke. Check the manual for your model, because the steps vary.
• Maximum output power, source type (LED or laser), and wavelength options for any particular source. These are product details and were not in the sources used.
• Whether a given source meets a specific project’s acceptance limit. Use the cable or project standard for that.

Related guides

For a wider look at fiber fault tools, read best cable fault locator. For copper network testing, see best network cable tester. A loss test with a light source and meter is usually the first step, with a fault locator as the next step if the loss is too high.

Summary

To use a fiber optic light source, match the fiber type and wavelength, clean every connector, set the reference with a short cord, and measure the link under test. For complex links, measure both directions and check that they agree within 0.2 dB, as Fluke recommends. Save the result with its setup, and use a custom limit when the project needs a pass or fail call. If the loss is high and you need to find where it sits, move on to an OTDR.

How we chose these sources

We used the Keysight blog for the source-and-meter pairing and the two-step loss method, and two Fluke Networks pages for the handheld workflow and the bidirectional procedure. Each page was opened and read on the date shown above. We did not use retailer pages or forum posts for any technical step.

Sources

1. Keysight Technologies, “Optical Power Meters: Understand Their Uses and Internals,” keysight.com/blogs/en/tech/educ/2023/optical-power-meters. Opened 2026-10-28.
2. Fluke Networks, “Bi-directional Testing with Light Source and Power Meter,” flukenetworks.com/blog/cabling-chronicles/bi-directional-testing-light-source-and-power-meter. Opened 2026-10-28.
3. Fluke Networks, “Measuring Optical Power – CertiFiber Pro,” flukenetworks.com/knowledge-base/certifiber-pro/measuring-optical-power-certifiber-pro. Opened 2026-10-28.

FAQ

Do I need to set a reference every time?

Set it again whenever the source output is disturbed, the wavelength changes, or you change the reference cords. Fluke’s guidance says an invalid reference must be reset before testing continues.

What wavelength should I use?

Match the wavelength to the network. Keysight gives 850 nm for short-distance local area networks and 1,550 nm for long-distance links, and Fluke’s CertiFiber examples include 850 nm and 1300 nm.

Why does the reading change when I reconnect a cord?

Each connection adds some loss, and a dirty connector adds more. Clean and inspect the connectors before each reading, and keep the reference cords in good condition.

Is a light source enough to find a break?

No. A light source and power meter measure total loss between two ends. To find where a break or bad splice sits, use an OTDR.

What does a 0.2 dB difference between directions mean?

Fluke’s guidance says a difference above 0.2 dB suggests the reference may be wrong, the test cord may be faulty, or the fiber may be bent. Recheck the setup before accepting the result.