Insertion Loss - Testing the Installed Fiber Optic Cable Plant
Typical fiber optic cable plants are composed of a backbone cable connecting patch panels and several short jumper cables which connect the equipment onto the cable plant. Premises cabling systems look like the photo to the right, where the backbone fiber is terminated in wiring closets and short jumpers connect wall outlets or directly to the equipment. These installations often have no splices at all, since distances are short. Premises cabling, which is short and often lower speed, is often multimode fiber, unless the network speed is >10Gb/s as in data centers, uses a GPON passive optical LAN or is shared with a cellular distributed antenna system.
Telco cable plants look similar, but the the fiber is all singlemode and cable runs may be long, requiring splices every 2-4 km. In addition, the fibers are not terminated directly, but high quality factory made pigtails are spliced onto the backbone cable. The process of testing any fiber optic cable plant during and after installation includes all the procedures covered so far.
To thoroughly test the cable plant, one needs to test it three times, a continuity test of the fiber optic cable on the reel before installation, insertion loss of each installed segment and complete end to end loss. One should test the cable on the reel for continuity before installing it, to insure no damage was done in shipment from the manufacturer to the job site. Since the cost of installation usually is high, often higher than the cost of materials, it only makes sense to insure that one does not install bad cable. It is generally sufficient to just test continuity, since most fiber is installed without connectors and then terminated in place, and connectors are the most likely problem to be uncovered by testing for loss. However, if any damage is visible on cable reels, OTDR testing may be needed to verify the cable is still good.
After installation and termination, each segment of the cable plant should be tested individually as it is installed, to insure each connector and cable is good. Finally each end to end run (from equipment placed on the cable plant to equipment) should be tested as a final check. Measured loss should be compared to the calculated loss budget for the cable plant to determine if the measured loss is acceptable.
is Insertion Loss?
Optical loss in a fiber optic link. The meter shows the decrease in optical power in the pulse as it traverses the fiber.
The goal of an insertion loss test is to simulate link operating conditions by using a test source to launch power into the fiber or cable under test and a power meter to measure the loss at the other end. That requires creating launch conditions from the test source that are similar to transmitter sources and using the power meter to measure the power from the source before and after the component or cable is inserted in the test setup. How this is done is the subject of many standards in fiber optics that ensure the test results from laboratories, manufacturers, installers or users are comparable.
Why is this test called an "insertion loss" test?
The name comes from the fact that one performs the test by inserting the components under test between a test source and a power meter. The most obvious version of this test is the method used by manufacturers to evaluate connector or splice loss as shown in the diagram below. It can also be used to measure the attenuation of optical fiber.
Insertion loss test for a fiber optic connector.
Cutback test to measure the attenuation coefficient of optical fiber.
This is the insertion loss test used by manufacturers for evaluating the performance of connectors or splices. This test connects the test source to a power meter over an optical fiber. Often this test is done with bare optical fiber with the connection to the power meter using a bare fiber adapter. The meter and source are turned on and a "0 dB" reference is set.
To test the component the fiber is cut and a pair of connectors or a splice is inserted in the fiber and the change in power measured. The change in power indicates the loss generated by the insertion of the component. When evaluating a new component, manufacturers may do hundreds or thousands of these tests to determine the average performance of the component that they specify in their product datasheets.
For tests involving multimode fiber, the test results are highly dependent on the test setup, particularly the modal distribution launched into the fiber used in the test. Test sources can have significant variations in modal distribution so it is important to include some form of modal conditioning in the test fiber. Modal conditioning and other issues that affect all multimode tests will be covered in a section below on test conditions.
Note: It's common in fiber optics to talk about the loss of a connector, but a connector loss is measured when mated to another connector. In fact, a single connector has no loss because a connector is defined as a component that allows making connections between two fibers or connects one fiber to an active device like a transmitter or receiver. The correct term is "connection loss" because it is the loss of a mated pair of connectors. In practice, we assign a loss to a connector by testing it against a reference connector, as you will see when we discuss single-ended insertion loss testing below.
Testing Cables And Cable Plants With Connectors
The most common tests involve testing fiber optic cable plants, patchcords and other components that have connectors on each end. These are tested with a test source and power meter with reference test cables to connect to the component under test. There are two versions of this test, a double ended test and a single ended test.
The double-ended test is the standard test for installed cable plants that allows testing the entire cable plant including the connectors on each end. It is a simulation of the loss an actual transmission system will see when connected to the cable with patchcords. Here is the block diagram of the test:
Standard cable plant test, double-ended with launch and receive cables
The source has a launch reference cable because it launches power into the component under test. The power meter has a receive reference cable because it receives the power after the component has been inserted. Each of these mate to one end of the cable under test to measure the connection loss on both ends, hence the name "double-ended." The cable under test is sometimes called the "permanent link," a reference to the name of the installed cable in UTP copper cable testing.
The "0 dB" reference measurement is made with the test source and power meter with their reference cables using one of three different but acceptable methods, depending on the types of connectors on the test equipment and the cable plant. The reason for the existence of three methods is the compatibility of test equipment to the cable plant; whether the test equipment has connector interfaces that allow direct connection to the cable under test.
Three ways to set a "0dB" reference for insertion loss testing.
Which method is used depends on the connectors on the cable plant you are testing and the connector interfaces on your test equipment. (And some history about how different companies defined testing.)
The options for use of these three methods are:
The insertion loss test then measures the loss generated by adding the cable under test. The value measured is called the loss of that particular fiber optic cable. The loss can be compared to the calculated loss budget to see if the cable plant meets specification.
Testing the complete cable plant is done per standard test procedures, e.g. FOA Standard FOA-1, TIA OFSTP-14 for multimode fiber or OFSTP-7 for singlemode, which use the same procedures or ISO/IEC 61280, ISO/IEC 14763, etc. These standards offer 3 different ways to set a "0 dB" reference and cover the peculiarities of multimode fiber in detail. Information for multimode cables covers the problems of controlling mode power distribution, but the same procedures apply for singlemode fiber, less the concerns expressed for mode power distribution errors.
multimode fibers, testing is now usually done at 850 and
sometimes also at 1300 nm, using LED sources with some
control over mode power distribution. This will prove the
performance of the cable for every datacom system,
including FDDI and ESCON, and meet the requirements of all
OFSTP-7 was written for singlemode fiber cables. Testing is usually done at 1300 nm, but 1550 or 1625 nm is sometimes required also. The1550 nm testing will show that the cable can support wavelength division multiplexing (WDM) at 1300 and 1550 nm for future service expansion. In addition, 1550 and 1625 nm testing can show microbending losses that will not be obvious at 1300 nm, since the fibers are much more sensitive to bending losses at the longer wavelengths.
If cable plant end to end loss exceeds total allowable loss, the best solution is to retest each segment of the cable plant separately, checking suspect cables each way, since the most likely problem is a single bad connector or splice. If the cable plant is long enough, an OTDR may be used to find the problem. Bad connectors must then be repolished or replaced to get the loss within acceptable ranges.
One can also do a variation of the insertion loss test with only a launch cable, called a "single ended" test, that only tests the connector on the source end of the cable. This test is often used to test patchcords since it allows the measurement of the connectors on each end separately, a more rigorous method of testing these short cables.
Fiber optic patchcord test, also called "single-ended" test
The single ended test looks like the double ended test except there is no receive reference cable. The cable under test is mated to the connector on the launch cable and the power meter. The "0 dB" reference measurement is made with the test source, a launch reference cable and the power meter, exactly like the one cable reference for double ended insertion loss tests.
The output of the launch cable is the 0 dB loss reference, so for short cables like patchcords, the measurement is basically the loss of the connection between the launch cable and the test cable. The test cable can be reversed and the connector on the other end tested separately. This test is used on patchcords because it allows testing each connector individually, ensuring that both connectors meet specifications. If the cable were tested with the double ended test, one would measure a total loss of both connectors and could not identify if one connector were out of specification.
Choosing Appropriate Test Equipment
In order to perform an insertion loss test, it is necessary to have appropriate equipment for the cables or cable plant under test. Whether one uses a test source and power meter or optical loss test set (OLTS), the equipment must be compatible with the test requirements.
The fiber optic power meter used for insertion loss testing should be calibrated at the wavelength of the test source being used. The meter should have a connector adapter compatible with the connectors on the cable plant being tested. Having a special "dB" range that will allow setting a "0 dB" loss reference power level will simplify testing.
An OLTS will offer both source and power meter for loss testing at the appropriate wavelengths. The requirements for the source for modal conditioning (see below) and compatibility of the instrument to the connectors on the cable plant being tested may have to be accommodated by the reference test cables.
Test Source Wavelengths
For multimode fiber, the test source should be a LED at 850 nm that is the wavelength used for virtually all multimode communications systems. There is an option for testing at 1300 nm with a LED also, but there are few systems today that operate at that wavelength so testing at that wavelength is generally unnecessary. You will find some references to using 1300 nm testing to find stress on the cable, since multimode fiber is much more sensitive to bending losses at 1300 nm. But even that reason is no longer relevant in most cases because most multimode fiber is of the bend-insensitive (BI) type. Finding stress areas in cables can be tricky anyway, but if the cable is long enough, an OTDR that offers both 850 and 1300 nm testing is a better instrument for finding the location of the loss.
While most 850 nm multimode systems operating over OM3 or OM4 fiber today use VCSEL sources (vertical cavity surface-emitting lasers), these sources are not recommended for use as test sources due to the unpredictable mode power distribution of individual devices. Instead, standards now call for sources with mode power distribution or conditioning the output of the source and its launch cable to approximate an ideal VCSEL mode condition. This will be covered below in reference cables and again in the section on mode conditioning for multimode fiber testing.
Singlemode fiber is tested with laser sources, similar to the devices that will be used in the communications systems which operate over the fiber. Singlemode fiber will be tested with 1310 nm and/or 1550 nm lasers depending on the cable plant to be tested. Short links, from hundreds of meters in a data center or building up to 20 or 30 km in metro networks, are always tested at 1310 nm and often at 1550 nm if wavelength division multiplexing (WDM) or passive optical networks (PONs) like FTTH are planned for use on the link. Long links are tested at 1550 nm to match the wavelengths of networks using them.
There may also be a reason to test short singlemode links at 1550 nm to find stress loss in a cable plant caused by installation. But as mentioned above, this is not relevant if bend-insensitive fiber is being used in the link, common today in high fiber density cables. Also testing at 1550 nm for finding stress is more relevant to ODTR testing where the cause of the stress loss can be determined.
Reference Test Cables
Just as important as the choice of test equipment is the choice of reference test cables for the launch cable and receive cable. The basic requirements for test cables is that they be about 1 to 2 meters long, match the size of fiber in the cable plant under test and have connectors compatible to the connectors on the cable plant. Multimode graded index fiber in test cables should be 62.5/125 for OM1 cable plants or 50/125 for OM2, OM3, OM4 or OM5 fiber cable plants. There are no significant differences in types of 50/125 fiber so any type of this size fiber can be used to test any other type. Today most of the multimode graded index fiber is bend insensitive (BI) fiber. Many standards recommend not using BI fiber for reference test cables even if testing BI fiber cables, but this may not be possible. We'll discuss BI fiber in the section on modal conditioning for multimode fiber. When testing step-index multimode cable plants using plastic optical fiber (POF) or plastic coated silica fiber (PCS), one must likewise choose a matching fiber for reference cables.
The connectors on the test cables should be PC polished (physical contact) and must be of very high quality, determined by having low loss when tested against each other in the single ended test mode. One should use low loss patchcords, typically under 0.3 dB, so the test results will be consistent. With use, the connectors will wear, even when cleaned frequently. If the patchcords exceed 0.5 dB over time, they should be replaced. Careful repolishing with diamond polishing film by an experienced tech may bring the loss back down. Or the cables should be replaced.
The new TIA OFSTP-14 and ISO/IEC 61280-4-1 standards for multimode call for 2 meter long "reference quality" test cables with connector loss of 0.1 dB. The standard explains that this may require special cables with selected fibers and selected connectors as well as higher quality mating adapters such as those used for singlemode fiber with ceramic mating adapters. For singlemode fiber the loss specified is 0.2 dB.
There are very few sources of "reference quality" test cables. Multimode cables with losses of 0.1 dB will degrade with use, showing the effects of the many mating cycles required when testing in the field. Using high quality cables with relatively loss is the practical solution. Cables with loss of 0.2 up to 0.5 dB maximum are generally adequate for testing multimode fiber.
The launch reference cable combines with the test source to create the modal conditions used for testing. If the source modal conditions are not specified, and they rarely are, the modal conditions can be modified by the user using special patchcords or modal conditioning on the cable itself. This is discussed in the section on modal conditioning below. Do not ignore modal conditioning in multimode testing. When coupled into multimode fiber, LEDs typically have higher modes than are specified by test standards. Mode conditioning will remove those higher modes and tests will be more consistent and will consistently show lower loss.
Singlemode reference cables should also be high quality cables. The connectors must also be compatible to the type of connector on the cable plant being tested, e.g. PC (physical contact, color coded blue) or APC (angled physical contact, color coded green.) For singlemode fiber "reference quality" test cables, the loss specified is 0.2 dB. High quality singlemode patchcords should be 0.3 dB or less so this limit is not unreasonable.
Singlemode does not have mode conditioning standards like multimode fiber, but short launch patchcords connected to a typical laser source may support 2 modes for short distances and that can affect measurement results. To ensure the launch is singlemode, a mode filter made with a small loop of fiber 40 to 60 mm diameter (1.6 to 2.4 inches) in the launch cable will ensure singlemode launching at the end of the cable.
Test Conditions For Accurate Insertion Loss Testing
As with any testing, to reduce measurement uncertainty, it is important to consider and as, far as possible, control test conditions. It is important to understand all the contributions to measurement uncertainty and how their effects can be minimized by the test operator. Many sources of measurement errors are not controllable by the user. They depend on the manufacturer of the test equipment, optical fiber, connectors, etc. and their quality control.
The first and most important issue is to decide which reference method to use (1, 2 or 3 cables) and record that with the test data. Since the test results change considerably with the choice of reference method, this is very important information to record. After setting the "0 dB" reference, it is important to not disconnect the reference launch cable from the source. The connection between the source and cable can change if it is disconnected and reconnected invalidating the "0 dB" reference set.
Secondly, monitor the type and condition of all the test reference cables. This is probably the biggest cause of random errors in testing. Of course the fiber in the reference cables must match the type in the cables being tested and the connectors must be compatible. The condition of the connectors on the reference cables should be inspected for cleanliness and damage, cleaned and inspected again to ensure proper cleaning. Then the reference cable connectors should be tested against each other to ensure they are still low loss. As the loss increases from more testing, replace the cables or refurbish the connectors by repolishing using diamond film.
The connections to test cables have another important component, the mating adapter used with connectors. All the single fiber connectors use mating adapters to align the ferrules. These adapters are available in several types, depending on the alignment sleeve for the ferrules. Some inexpensive mating adapters use plastic alignment sleeves that should never be used for testing. These sleeves wear out quickly and leave dust and residue on the connectors. The acceptable mating adapters use metal or ceramic alignment sleeves. Metal sleeves work for hundreds of tests but ceramic sleeves will last many times longer and leave no residue.
Test equipment used for insertion loss testing should be checked and power meters calibrated regularly, according to manufacturers specifications. When making tests, the biggest problems are with sources. First of all, the source needs to be stable; if its output power drifts, the 0 dB reference will be lost and all tests will be wrong. You can test your own source by connecting it to a meter with a patchcord and turning it on. Note how long it takes the output to be stable and use that time as a warmup time when setting up to make tests. If the source is not stable over time, varying more than 0.1 dB after warmup, it should be checked or replaced with a more stable source.
The second issue with sources is modal conditioning. This is mainly a multimode fiber and LED test source problem, but even single mode sources with lasers can have modal problems. The user should ensure that multimode launch cables have proper mode power distribution since that can affect the loss measured significantly. Using a simple mandrel wrap and checking the modal distribution with a HOML (higher order mode loss) test as described below will greatly reduce measurement errors. For singlemode fiber, a simple loop mode filter is all that is needed. With all reference cables, be careful to not stress them during the tests as that can induce loss that will change the 0 dB reference and or create changes in the modal distribution.
The technician performing the tests should be experienced in the process, familiar with the procedures and conduct every test in the same manner. Even small issues like stress on reference cables can make big differences in measurements.
If reasonable precautions are taken, what is the likely accuracy of loss measurements? Experience has shown that typical measurements have an uncertainty of approximately +/-10% of the measured value in dB. Thus a 2dB loss has an estimated uncertainty of +/-0.2 dB.
Most problems with high cable loss are caused by bad or dirty connectors, high loss splices or stress loss in the cable caused during installation.
The first step is connectors should be inspected with a microscope for dirt, scratches cracks, or other damage and thoroughly cleaned. Visual fault locators can check for continuity, proper connections and, if the cable jacket permits, high loss bends or breaks.
If you have high loss in a single cable with connectors on each end, you can reverse it and test in the opposite direction using the single-ended test method. Since the single ended test only tests the connector on the end connected to the launch cable, you can isolate a bad connector this way. The bad connector is the one at the end mated to the launch cable when you measure high loss.
High loss in the double-ended test should be isolated by retesting single-ended and reversing the direction of test to see if the end connector is bad. If the loss is the same, you need to either test each segment separately to isolate the bad segment or, if it is long enough, use an OTDR.
What About OTDR Testing?
The OTDR is better at troubleshooting problems than making quantitative measurements because it uses a different test method. However it can be invaluable at testing in certain situations and troubleshooting. Read more about OTDRs here.
This page was taken from the FOA Guide To Fiber Optic Testing, Chapter 8.