OSP Fiber Optic Testing
After fiber optic cables are installed, spliced and terminated, they must be tested. For every fiber optic cable plant, you need to test for continuity and polarity, end-to-end insertion loss and then troubleshoot any problems. If it's a long outside plant cable with intermediate splices, you will probably want to verify the individual splices with an OTDR test also, since that's the only way to make sure that each splice is good. If you are the network user, you may also be interested in testing transmitter and receiver power, as power is the measurement that tells you whether the system is operating properly.
Testing is the subject of the majority of industry standards, as there is a need to verify component and system specifications in a consistent manner. A list of TIA fiber optic standards is on the FOA website in Tech Topics. Perhaps the most important test is insertion loss of an installed fiber optic cable plant performed with a light source and power meter (LSPM) or optical loss test set (OLTS) which is required by all international standards to ensure the cable plant is within the loss budget before acceptance of the installation.
Testing fiber optic components and cable plants requires making several tests and measurements with the most common tests listed below. Some tests involve installer inspection and judgment, such as visual inspection or tracing while some use sophisticated instruments that provide direct measurements. Optical power, required for measuring source power, receiver power and, when used with a test source, loss or attenuation, is the most important parameter and is required for almost every fiber optic test. Backscatter measurements made by an OTDR are the next most important measurements, especially for testing outside plant installations and troubleshooting. Measurement of geometrical parameters of fiber and bandwidth or dispersion are essential for fiber manufacturers but not relevant to field testing. Troubleshooting installed cables and networks is required in every installation.
Testing fiber optics requires special tools and instruments which must be chosen to be appropriate for the components or cable plants being tested. See Jargon and Test Instruments to see a description of these instruments.
Continuity checking with a visual fiber tracer makes certain the fibers are not broken and to trace a path of a fiber from one end to another through many connections, verifying duplex connector polarity for example. It looks like a flashlight or a pen-like instrument with a light bulb or LED source that mates to a fiber optic connector. Attach the fiber to test to the visual tracer and look at the other end of the fiber to see the light transmitted through the core of the fiber. If there is no light at the end, go back to intermediate connections to find the bad section of the cable.
A good example of how it can save time and money is testing fiber on a reel before you pull it to make sure it hasn't been damaged during shipment. Look for visible signs of damage (like cracked or broken reels, kinks in the cable, etc.) . For testing, visual tracers help also identify the next fiber to be tested for loss with the test kit. When connecting cables at patch panels, use the visual tracer to make sure each connection is the right two fibers! And to make certain the proper fibers are connected to the transmitter and receiver, use the visual tracer in place of the transmitter and your eye instead of the receiver (remember that fiber optic links work in the infrared so you can't see anything anyway.)
Visual Fault Location
A higher power version of the fiber tracer called a visual fault locator (VFL) uses a visible laser that can also find faults. The red laser light is powerful enough for continuity checking or to trace fibers for several kilometers, identify splices in splice trays and show breaks in fibers or high loss connectors. You can actually see the loss of light at a fiber break by the bright red light from the VFL through the jacket of many yellow or orange simplex cables (excepting black or gray jackets, of course.) It's most important use is finding faults in short cables or near the connector where OTDRs cannot find them.
You can also use this gadget to visually verify and optimize mechanical splices or prepolished-splice type fiber optic connectors. By visually minimizing the light lost you can get the lowest loss splice. In fact- don't even think of doing one of those prepolished-splice type connectors without one. No other method will assure you of high yield with those connectors.
A note on VFL eye safety. VFLs use visible light. You will find it uncomfortable to look at the output of a fiber illuminated by a VFL. That's good, because the power level is high and you should not be looking at it. When tracing fibers, look from the side of the fiber to see if light is present.
Visual Connector Inspection by Microscope
Fiber optic inspection microscopes are used to inspect connectors to confirm proper polishing and find faults like scratches, polishing defects and dirt. They can be used both to check the quality of the termination procedure and diagnose problems. A well made connector will have a smooth , polished, scratch free finish and the fiber will not show any signs of cracks, chips or areas where the fiber is either protruding from the end of the ferrule or pulling back into it.
The magnification for viewing connectors can be 30 to 400 power but it is best to use a medium magnification. If the magnification is too low, critical details may not be visible. Inspecting with a very high magnification may cause the viewer to be too critical, rejecting good connectors. Multimode connectors should use magnifications in the range of 100-200X and singlemode fiber can use higher magnification, up to 400X. A better solution is to use medium magnification, but inspect the connector three ways: viewing directly at the end of the polished surface with coaxial or oblique lighting, viewing directly with light transmitted through the core, and viewing at an angle with lighting from the opposite angle or with quite oblique lighting.
Viewing directly allows seeing the fiber and the ferrule hole, determining if the ferrule hole is of the proper size, the fiber is centered in the hole and a proper amount of adhesive has been applied. Only the largest scratches may be visible this way, however. Adding light transmitted through the core will make cracks in the end of the fiber, caused by pressure or heat during the polish process, visible.
Viewing the end of the connector at an angle, while lighting it from the opposite side at approximately the same angle or using low-angle lighting and viewing directly will allow the best inspection for the quality of polish and possible scratches. The shadowing effect of angular viewing or lighting enhances the contrast of scratches against the mirror smooth polished surface of the glass.
One needs to be careful in inspecting connectors, however. The tendency is to sometimes be overly critical, especially at high magnification. Only defects over the fiber core are generally considered a problem. Chipping of the glass around the outside of the cladding is not unusual and will have no effect on the ability of the connector to couple light in the core on multimode fibers. Likewise, scratches only on the cladding should not cause any loss problems.
The best microscopes allow you to inspect the connector from several angles, either by tilting the connector or having angle illumination to get the best picture of what's going on. Check to make sure the microscope has an easy-to-use adapter to attach the connectors of interest to the microscope.
Video readout microscopes are now available that allow easier viewing of the end face of the connector and some even have software that analyzes the finish. While they are much more expensive than normal optical microscopes, they will make inspection easier and greatly increase productivity.
Remember to check that no power is present in the cable before you look at it in a microscope to protect your eyes! The microscope will concentrate any power in the fiber and focus it into your eye with potentially hazardous results. Some microscopes have filters to stop the infrared radiation from transmitters to minimize this problem.
More on Visual Inspection.
Optical Power - Power or Loss? ("Absolute" vs. "Relative" Measurements)
Practically every measurement in fiber optics refers to optical power. The output of a transmitter or the input to receiver are "absolute" optical power measurements, that is, you measure the actual value of the power. Loss is a "relative" power measurement, the difference between the power coupled into a component like a cable, splice or a connector and the power that is transmitted through it. This difference in power level before and after the component is what we call optical loss and defines the performance of a cable, connector, splice, or other component.
Whenever tests are performed on fiber optic networks, the results are displayed on an instrument readout. Power measurements are expressed in "dB," the measurement unit of power and loss in optical fiber measurements. Optical loss is measured in “dB” while optical power is measured in “dBm.” Loss is a negative number (like -3.2 dB) as are many power measurements. Measurements in dB can sometimes be confusing.
In the early days of fiber optics, source output power was usually measured in milliwatts, a linear scale, and loss was measured in dB or decibels, a logarithmic scale. Over the years, all measurements migrated to dB for convenience causing much confusion. Loss measurements were generally measured in dB since dB is a ratio of two power levels, one of which is considered the reference value. The dB is a logarithmic scale where each 10 dB represents a ratio of 10 times. The actual equation used to calculate dB is
dB = 10 log (measured power / reference power).
Thus 10 dB is a ratio of 10 times (either 10 times as much or one-tenth as much), 20 dB is a ratio of 100, 30 dB is a ratio of 1000, etc. When the two optical powers compared are equal, dB = 0, a convenient value that is easily remembered. If the measured power is higher than the reference power, dB will be a positive number, but if it is lower than the reference power, it will be negative. Thus measurements of loss are typically expressed as a negative number.
Measurements of optical power such as the output of a transmitter or input to a receiver are expressed in units of dBm. The “m” in dBm refers to a reference power of 1 milliwatt. Thus a source with a power level of 0 dBm has a power of 1 milliwatt. Likewise, -10 dBm is 0.1 milliwatt and +10 dBm is 10 milliwatts.
To measure loss in a fiber optic system, we make two measurements of power, a reference measurement before the component we are testing and a loss measurement after the light passes through the component. Since we are measuring loss, the measured power will be less than the reference power, so the ratio of measured power to reference power is less than 1 and the log is negative, making dB a negative number. When we set the reference value, the meter reads “0 dB” because the reference value we set and the value the meter is measuring is the same. Then when we measure loss, the power measured is less, so the meter will read “- 3.0 dB” for example, if the tested power is half the reference value. Although meters measure a negative number for loss, convention is the loss is expressed as a positive number, so we say the loss is 3.0 dB when the meter reads - 3.0 dB.
Instruments that measure in dB can be either optical power meters or optical loss test sets (OLTS). The optical power meter usually reads in dBm for power measurements or dB with respect to a user-set reference value for loss. While most power meters have ranges of +3 to -50 dBm, most sources are in the range of +10 to -10 dBm for lasers and -10 to -20 dBm for LEDs. Only lasers used in CATV or long-haul telephone systems have powers high enough to be really dangerous, up to +20 dBm; that’s 100 milliwatts or a tenth of a watt.
It is important to remember that dB is for measuring loss, dBm is for measuring power and the more negative a number is, the higher the loss. Set your zero reference before measuring loss and check it occasionally while making measurements.
Read more about dB.
Calibration of Fiber Optic Power Measurements
Calibrating fiber optic power measurement equipment requires setting up a reference standard traceable to a national standards lab like the National Institute of Standards and Technology in the US for comparison purposes while calibrating every power meter or other instrument. The NIST standard for all power measurements is an ECPR, or electrically calibrated pyroelectric radiometer, which measures optical power by comparing the heating power of the light to the well-known heating power of a resistor. Calibration is done at 850, 1300 and 1550 nm. Sometimes, the wavelength of lasers at 1310 nm is used by manufacturers as the calibrated wavelength on a power meter, but the standard for power meter calibration is 1300 nm. To conveniently transfer their laboratory standard to fiber optic power meter manufacturers calibration laboratories, NIST currently uses a laboratory optical power meter which is sent around to labs as a transfer standard.
Meters calibrated in this manner have an uncertainty of calibration of about +/- 5%, compared to the NIST primary standards. Limitations in the uncertainty are the inherent inconsistencies in optical coupling, about 1% at every transfer, and slight variations in wavelength calibration. NIST is working continuously with instrument manufacturers and private calibration labs to try to reduce the uncertainty of these calibrations.
Recalibration of instruments should be done annually, however experience has shown that the accuracy of meters rarely changes significantly during that period, as long as the electronics of the meter do not fail. The calibration of fiber optic power meters requires considerable investment in capital equipment so meters must be returned to the original manufacturer or private calibration labs for calibration.
More on optical power calibration.
Understanding FO Power Meter Measurement Uncertainty
Much attention has been paid to developing transfer standards for fiber optic power measurements. The US NIST in Boulder, Colorado and standards organizations of most other countries have worked to provide good standards to work from. We can now assure traceability for our calibrations, but even so the errors involved in making measurements are not ignorable. Even when fiber optic power meters are calibrated within specifications, the uncertainty of a measurement may be as great as +/- 5% (about 0.2 dB) compared to standards. Understanding power meter errors and their probable causes will insure a realistic viewpoint on fiber optic power measurements.
The first source of error is optical coupling. Light from the fiber is expanding in a cone. It is important that the detector to fiber geometry be such that all the light from the fiber hits the detector, otherwise the measurement will be lower than the actual value. But every time light passes through a glass to air interface, such as the window on the detector, a small amount of the light is reflected and lost. Finally, the cleanliness of the optical surfaces involved can cause absorption and scattering. The sum total of these potential errors will be dependent on the connector type, wavelength, fiber size and NA.
Beyond the coupling errors, one has errors associated with the wavelength calibration. Semiconductor detectors used in fiber optic instruments (and systems too) have a sensitivity that is wavelength dependent. Since the actual source wavelength is rarely known, there is an error associated with the spectral sensitivity of the detector. By industry convention, the three cardinal wavelengths (850, 1300 and 1550 nm) are used for all power measurements, not the exact source wavelength.
Another source of error exists for high and low level measurements. At high levels, the optical power may overload and saturate the detector, causing the measurement to be in error. At low levels, the inherent detector noise adds to the signal and becomes an error. If the signal is 10 dB above the noise floor (10 time the noise), the offset error is 10% or 0.4 dB.
Instrument Resolution vs. Measurement Uncertainty
Considering the uncertainty of most fiber optic measurements, instrument manufacturers have provided power and loss meters with a measurement resolution that is usually much greater than needed. The uncertainty of optical power measurements is about 0.2 dB (5%), loss measurements are more likely to have uncertainties of 0.2-0.5 dB or more, and optical return loss measurements have a 1 dB uncertainty.
Instruments which have readouts with a resolution of 0.01 dB are generally only appropriate for laboratory measurements of very low component losses or changes caused by environmental changes. Within the laboratory, a resolution of 0.01 dB can be extremely useful, since one often measures the loss of connectors or splices that are under 0.10 dB or changes in loss under environmental stress that are under 0.1 dB. Stability of sources and physical stress on cables limits measurement uncertainty to about 0.02 to 0.05 dB per day, but 0.01 dB resolution can be helpful in determining small changes in component performance.
Field measurements have higher uncertainty because more components are measured at once and losses are higher. Practically, measurements are better when the instrument resolution is limited to 0.1dB. Readings will be more likely to be stable when being read and more indicative of the measurement uncertainty.
Take a minute and read more about dB the measurement unit of power and loss in optical fiber measurements.
Measuring Power With A Fiber Optic Power Meter
Measuring power requires only a power meter (most come with a screw-on adapter that matches the connector being tested), a known good fiber optic cable (of the right fiber size, as coupled power is a function of the size of the core of the fiber) and a little help from the network electronics to turn on the transmitter. Remember when you measure power, the meter must be set to the proper range (usually dBm, sometimes microwatts, but never "dB" - that's a relative power range used only for testing loss! read more about dB and the proper wavelength , matching the source being used in the system (850, 1300, 1550 nm for glass fiber, 650 or 850 nm for POF). Refer to the instructions that come with the test equipment for setup and measurement instructions (and don't wait until you get to the job site to try the equipment, try it in the office first!)
To measure power, attach the meter to the cable attached to the source that has the output you want to measure (see diagram to the right). That can be at the receiver to measure receiver power, or using a reference test cable (tested and known to be good) that is attached to the transmitter to measure output power. Turn on the transmitter/source and give it a few minutes to stabilize. Set the power meter for the matching wavelength and note the power the meter measures. Compare it to the specified power for the system and make sure it's enough power but not too much.
More on measuring Power.
Testing Optical Loss or Insertion Loss
An insertion loss measurement is made by mating the cable being tested to known good reference cables with a calibrated launch power that becomes the "0 dB" loss reference. Why do you need reference cables to measure loss? Why can't you just plug the cable to test into a source and power meter and measure the power? There are several reasons:
1)You need a cable to measure the output power of the source for calibration of "0 dB" loss.
2) In order to measure the loss of the connectors you must mate them to a similar, known good, connector.
This is an important point often not fully explained. When we say connector loss, we really mean "connection" loss - the loss of a mated pair of connectors. Thus, testing connectors requires mating them to reference connector which must be high quality connectors themselves to not adversely affect the measured loss when mated to an unknown connector.
3)Testing with reference cables on each end simulates a cable plant with patchcords connecting to transmission equipment.
Fiber Optic Test Sources
In addition to a power meter, you need a test source. The test source should match the type fiber ( generally LED for MM or laser for SM) and wavelength (850, 1300, 1550 nm) that will be used on the fiber optic cable you are testing. If you are testing to some standards, you may need to add some mode conditioning, like a mandrel wrap, to meet the standard launch conditions.
A fiber optic test source must be chosen for compatibility with the type of fiber in use (singlemode or multimode with the proper core diameter) and the wavelength desired for performing the test. Most sources are either LED's or lasers of the types commonly used as transmitters in actual fiber optic systems, making them representative of actual applications and enhancing the usefulness of the testing. Some laboratory tests, such as measuring the attenuation of fiber over a range of wavelengths requires a variable wavelength source, which is usually a tungsten lamp with a monochromator to vary the light source wavelength.
Typical wavelengths of sources are 650 or 665 nm (plastic fiber), 820, 850 and 870 nm (short wavelength multimode fiber ) and 1300 (long wavelength multimode fiber) or 1310 nm and 1550 nm (long wavelength singlemode fiber). LED's are typically used for testing multimode fiber and lasers are used for singlemode fiber, although there is some crossover. High speed LANs which use multimode fiber may be tested with VCSELs like the system sources and short singlemode jumper cables may be tested with LEDs.
The source wavelength can be a critical issue in making accurate loss measurements on long links, since the attenuation coefficient of the fiber is wavelength sensitive. Thus all test sources should be calibrated for wavelength in case corrections for wavelength variations are required.
Test sources almost always have fixed connectors. Hybrid test jumpers with connectors compatible with the source on one end and the connector being tested on the other must be used as reference cables. This may affect the type of reference setting mode used for loss testing.
Source-related factors affecting measurement accuracy are the stability of the output power and the modal distribution launched into multimode fiber. Source stability is mainly a factor of the electronic circuitry in the source. Industry standards have requirements on the modal output of test sources for multimode fiber that are important to the manufacturers of the test sources. Various standards have called for mode scramblers, filters and strippers to adjust the modal distribution in the fiber to approximate actual operating conditions. Today, most standards call for sources to meet output requirements and for a mode conditioner to be used in testing. The effects of mode power distribution on multimode measurements are covered in the chapter on optical fiber.
Reference cables and mating adapter
Loss testing requires one or two reference cables, depending on the test performed and the appropriate mating adapters for the connectors. Reference cables are typically 1-2 meters long, with fiber and connectors matching the cables to be tested. The accuracy of the measurement will depend on the quality of the reference cables, since they will be mated to the cable under test. The quality and cleanliness of the connectors on the launch and receive cables is one of the most important factors in the accuracy of loss measurements. Always test reference cables by the patchcord or single ended method shown below to make sure they are in good condition before you start testing other cables.
Standards groups have not been able to successfully specify the quality of reference cables in terms of tightly toleranced components like the fiber and connectors but only by the insertion loss of the connectors on the cables when tested with other high quality test cables. The best recommendation for qualifying reference cables is to choose cables with low loss, tested "single-ended" per FOTP-171 below.
In order to mate the reference cables to the cables you want to test, you need mating adapters. Mating adapters are as important to low connection loss as the quality of the connectors since the mating adapter is responsible for aligning the two connector ferrules correctly. Mating adapters must be kept clean, like connectors and discarded after some number of uses as they wear out from repeated matings. Mating adapters may have alignment sleeves made from plastic, metal or ceramic. Plastic alignment sleeves used on the cheapest mating adapters should not be used for testing as they wear out in only a few insertions, leaving dusty residue on the connectors. Metal adapters are good for many more insertions and provide a better alignment, so they are acceptable. Ceramic alignment sleeves are the best, providing the best alignment and practically never wearing out.
More on reference cables.
Single-ended (patchcord) test
Double-ended testing for installed cable plant
There are two methods that are used to measure loss, a "patchcord test" which we call "single-ended loss" (TIA FOTP-171) and an "installed cable plant test" we call "double-ended loss" (TIA OFSTP-14 (MM) and OFSTP-7 (SM).) Single-ended loss uses only the launch cable, while double-ended loss uses a receive cable attached to the meter also when making the loss measurement.
Single-ended loss is measured by mating the cable you want to test to the reference launch cable and measuring the power out the far end with the meter. When you do this you measure the loss of the connector mated to the launch cable and the loss of any fiber, splices or other connectors in the cable you are testing. Since you are aiming the connector on the far end of the cable mated to the power meter at a large area detector instead of mating it to another connector, it effectively has no loss so it is not included in the measurement. This method is described in FOTP-171 and is shown in the drawing. An advantage to this test is you can troubleshoot cables to find a bad connector since you can reverse the cable to test the connectors on the each end individually.
In a double-ended loss test, you attach the cable to test between two reference cables, one attached to the source and one to the meter. This way, you measure two connectors' loses, one on each end, plus the loss of all the cable or cables, including connectors and splices, in between. This is the method specified in OFSTP-14 (multimode, the singlemode test is OFSTP-7), the standard test for loss in an installed cable plant.
Three Ways to Set "0" dB Loss Reference
One Cable Reference
Most fiber optic connectors are constructed so that the fiber is held in a protruding ferrule, called a “plug” style connector. Two plug connectors are mated using a mating adapter that holds the ferrules in alignment and allows them to meet in the center. If connectors like these are being tested and the test equipment has interfaces that fit those connectors, the single cable reference (OFSTP-14 Method B) can be used. This method is the simplest method and generally considered the method of choice as no connections are included when setting the 0 dB reference.
After setting a reference, the launch cable is detached from the meter, but not the source. The launch reference cable should never be removed from the source after setting the reference to ensure the launch power remains constant. The receive cable is attached to the meter and then both reference cables are attached to the cable to test. The loss reading will include both connections to the cable under test and the loss of the fiber and any other components in the cable itself.
Two Cable Reference
If the test equipment has an interface for a different style of connector, so the connectors on the cables being tested cannot be attached to the instruments, a two cable reference method (OFSTP-14 Method A) can be used. Reference cables must be hybrid cables with connectors on one end to match the interfaces of the instruments and the other end to mate to the connectors on the cable to be tested. The 0 dB reference is set by attaching both reference cables to the instruments and connecting the other ends with a mating adapter. After setting the reference, the two cables are disconnected at the middle and the cable to be tested inserted in between them.
The loss reading will include both connections to the cable under test and the loss of the fiber and any other components in the cable itself less the loss of the connection between the two reference cables when setting the reference. Thus loss measured using the two cable reference will be lower than the one cable reference by the connection included when setting the reference. The uncertainty of this connection loss included in the reference also adds to the uncertainty of the loss measurement of any cables tested in this manner.
Three Cable Reference
Some fiber optic connectors are “plug” and “jack” style connectors where one has a protruding ferrule while the other has a jack or receptacle. Some have alignment pins that are only on one side, like the MTP connector where pins are used on the jack side. They are generally used with plugs on both ends of patchcords and jacks or receptacles on the permanently installed cables terminated in racks or outlets.
Either of these two styles of connectors can only be mated to an appropriate style of connector, making it hard to do a one or two cable reference. The solution is a three cable reference (OFSTP-14 Method C), where the hybrid cables attached to the instruments for reference cables are terminated in plugs and a third cable terminated in jacks is inserted between them to create a 3 cable reference. After setting the reference, the two reference cables are disconnected from the third cable at the middle and the cable to be tested inserted in between them in place of the reference cable.
As before, the loss reading will include both connections to the cable under test and the loss of the fiber and any other components in the cable itself less the loss of the two connections between the third reference cable and the two reference cables when setting the reference. Since the third cable is usually only a short length of fiber with connections on each end, the loss of the fiber is ignorable. The loss measured using the three cable reference will be lower than the one cable reference by the two connections included when setting the reference. The uncertainty of these two connection losses included in the reference also adds to the uncertainty of the loss measurement of any cables tested in this manner.
While this three cable method has the highest uncertainty, it is the only method that works for any connectors and any test equipment. Therefore, it has become the preferred method in several international standards.
The most preferred method is the one cable method.
Choosing a Reference Method
Some reference books and manuals show setting the reference power for loss using only a launch reference cable, both a launch and receive cable mated with a mating adapter or even three reference cables. Industry standards, in fact, include all three methods of setting a "0dB loss" reference. The two or three cable reference methods are acceptable for some tests and are the only way you can test some connectors, but it will reduce the loss you measure by the amount of loss between your reference cables when you set your "0dB loss" reference. Also, if any of the reference cables are bad, setting the reference with the cables does not reveal that problem. Then you could begin testing with bad launch cables making all your loss measurements wrong. This means that it is very important to inspect and test reference cables to ensure they are in good condition.
More on the differences in insertion loss test reference methods.
Clean your connectors for every test!
Making The Insertion Loss Measurement
Turn on the source, attach a launch reference cable and select the wavelength you want for the loss test. It's always a good idea to let the source warm up to a stable power output. Turn on the meter, select the "dBm" or "dB" range and select the wavelength you want for the loss test. Attach the source and launch cable to the meter, along with other reference cables if the two or three cable reference methods are being used. Measure the power at the meter. This is your reference power level for all loss measurements. If your meter has a "zero" function, set this as your "0" reference.
Single-ended test used for patchcords or testing cable one connector at a time.
Double-ended test for installed cable plants
Connect the cable to test as shown above and measure the loss. Record the wavelength of the source used for the test, the 0 dB reference method and the loss.
It is very important to always record the method of setting a "0dB loss" reference when testing insertion loss. The two or three cable reference methods are acceptable for some tests and are the only way you can test if the connectors on the cable plant are not the same as your test equipment, but it will reduce the loss you measure from the loss by a single cable reference by the amount of loss between your reference cable connections when you set your "0dB loss" reference. Also, if either the launch or receive cable is bad, setting the reference with both cables hides the fact. Then you could begin testing with bad reference cables making all your loss measurements wrong. Always test and clean reference cables before beginning to test cables.
Below is an animation of what happens when you test the same cable with each of these options.
Testing the same cable with different methods of setting the "0dB" reference gives different test results.
Videos on loss testing on the FOA Channel on
Modal Effect on Multimode Loss Measurements
If you measure the attenuation of a long graded-index multimode fiber in EMD (or with EMD simulated launch conditions) and compare it to a normal fiber with "overfill launch conditions " (that is the source fills all the modes equally), you will find the difference is about 1 dB/km, and this figure is called the "transient loss". Thus, the EMD fiber measurement gives an attenuation that is 1 dB per Km less than the overfill conditions. Fiber manufacturers use the EMD type of measurement for fiber because it is more reproducible and is representative of the losses to be expected in long lengths of fiber. Some standards call for using a higher attenuation coefficient when estimating cable plant loss than the tested attenuation coefficient of most fibers would justify, because cables covered in this standard are much shorter than EMD lengths.
Likewise, when testing cables with connectors, the loss measured depends on the mode power distribution in the fiber. An EMD measurement can give optimistic results, since it effectively represents a situation where one launches from a smaller diameter fiber of lower NA than the receive fiber, yielding lower connector loss. The difference in connector loss caused by modal launch conditions can be dramatic. Using the same pair of connectors, it is possible to measure several tenths of a dB more with a fully filled launch than with a EMD simulated launch.
Most testing standards for cables with multimode fibers call for some method of controlling mode power distribution. Manufacturers use sophisticated methods that analyze the output power of the test source coupled into a reference cable. More practical field test methods call for a specification on source modal output followed by a mandrel wrap on the launch cable to remove higher order modes.
You should refer to relevant standards to determine the proper type of mode filter that should be used. If you are using patchcords or cables that are made with bend-insensitive fiber, the dimensions of a modal wrap may be different and you should use the recommendations of the manufacturer of the fiber.
More on modal effects on multimode cable tests.
What Loss Should You Get When Testing Cables?
Before testing, preferable during the design phase, you should calculate a loss budget for the cable plant to be tested to understand the expected measurement results. Besides proviiding reference loss values to test against, it will confirm that the network transmission equipment will work properly on this cable. While it is difficult to generalize, here are some guidelines:
(0.5 dB X # connectors) + (0.2 dB x # splices) + fiber loss on the total length of cable
If you have high loss in a cable, make sure to reverse it and test in the opposite direction using the single-ended method. Since the single ended test only tests the connector on one end, you can isolate a bad connector - it's the one at the launch cable end (mated to the launch cable) on the test 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.
If you see no light through the cable (very high loss - only darkness when tested with your visual tracer), it's probably one of the connectors, and you have few options. The best one is to isolate the problem cable, cut the connector of one end (flip a coin to choose) and hope it was the bad one (well, you have a 50-50 chance!)
FOA Tech Bulletin on troubleshooting.
OTDRs are an essential test tool for most OSP installation and restoration. OTDRs are used during installation to check cable before installation for shipping damage and after installation for damage caused by the installation process. Each fiber splice is checked by an OTDR before the splice is placed in a splice tray and the closure is sealed. After a cable run is concatenated, the OTDR will inspect the complete cable and every fiber in it and keep a record of the cable for future reference in case of changes in the network or comparison for restoration. During restoration, OTDRs can be used to locate problems like a cable cut or compare the current data to the installed data for comparison and determination of problems.
OTDRs are fiber optic instruments that can take a snapshot of a fiber, showing the location of splices, connectors, faults, etc. OTDRs are powerful test instruments for fiber optic cable plants, if one understands how to properly set the instrument up for the test and interpret the results. When used by a skillful operator, OTDRs can locate faults, measure cable length and verify splice loss. Within limits, they can also measure the loss of a cable plant. About the only fiber optic parameters they don't measure is optical power at the transmitter or receiver. There is a lot of information in the OTDR trace, as shown in the actual trace below.
OTDRs are almost always used on outside plant cables to verify the loss of each splice as it is made and find stress points or damage caused by installation. They are also widely used as OSP troubleshooting tools since they can pinpoint problem areas such as loss caused by stress placed on a cable during installation. Most ODTRs lack the distance resolution for use on the shorter cables typical of premises networks.
OTDRs are available in versions for standardized fiber optic systems, singlemode or multimode, at the appropriate wavelengths. In order to use an OTDR properly, it's necessary to understand how it works, how to set the instrument up properly and how to analyze traces. OTDRs offer an "auto testing" option, but using that option without understanding the OTDR and manually checking its work often leads to problems.
(Read more about the OTDR)
How OTDRs Work
Unlike sources and power meters which measure the loss of the fiber optic cable plant directly, the OTDR works indirectly. The source and meter duplicate the transmitter and receiver of the fiber optic transmission link, so the measurement correlates well with actual system loss.
The OTDR, however, uses backscattered light of the fiber to imply loss. The OTDR works like RADAR, sending a high power laser light pulse down the fiber and looking for return signals from backscattered light in the fiber itself or reflected light from connector or splice interfaces.
At any point in time, the light the OTDR sees is the light scattered from the pulse passing through a region of the fiber. Only a small amount of light is scattered back toward the OTDR, but with wider test pulses, sensitive receivers and signal averaging, it is possible to make measurements over relatively long distances. Since it is possible to calibrate the speed of the pulse as it passes down the fiber, the OTDR can measure time, calculate the pulse position in the fiber and correlate what it sees in backscattered light with an actual location in the fiber. Thus it can create a display of the amount of backscattered light at any point in the fiber.
Since the pulse is attenuated in the fiber as it passes along the fiber and suffers loss in connectors and splices, the amount of power in the test pulse decreases as it passes along the fiber in the cable plant under test. Thus the portion of the light being backscattered will be reduced accordingly, producing a picture of the actual loss occurring in the fiber. Some calculations are necessary to convert this information into a display, since the process occurs twice, once going out from the OTDR and once on the return path from the scattering at the test pulse.
Actual OTDR Trace
Diagram of OTDR trace with events shown
There is a lot of information in an OTDR display. The slope of the fiber trace shows the attenuation coefficient of the fiber and is calibrated in dB/km by the OTDR. In order to measure fiber attenuation, you need a fairly long length of fiber with no distortions on either end from the OTDR resolution or overloading due to large reflections. If the fiber looks nonlinear at either end, especially near a reflective event like a connector, avoid that section when measuring loss.
Note the large initial pulse on the OTDR trace. That is caused by the high-powered test pulse reflecting off the OTDR connector and overloading the OTDR receiver. The recovery of the receiver causes the "dead zone" near the OTDR. To avoid problems caused by the dead zone, always use a launch cable of sufficient length when testing cables.
Connectors and splices are called "events" in OTDR jargon. Both should show a loss, but connectors and mechanical splices will also show a reflective peak so you can distinguish them from fusion splices. Also, the height of that peak will indicate the amount of reflection at the event, unless it is so large that it saturates the OTDR receiver. Then peak will have a flat top and tail on the far end, indicating the receiver was overloaded. The width of the peak shows the distance resolution of the OTDR, or how close it can detect events.
Here is an interactive lesson in how to read an OTDR trace from FOA's Fiber U.
OTDRs can also detect problems in the cable caused during installation. If a fiber is broken, it will show up as the end of the fiber much shorter than the cable or a high loss splice at the wrong place. If excessive stress is placed on the cable due to kinking or too tight a bend radius, it will look like a splice at the wrong location.
Making Measurements With The OTDR
All OTDRs display the trace on a screen and provide two or more markers to place at points on the screen to measure loss and distance. This can be used for measuring loss of a length of fiber, where the OTDR will calculate the attenuation coefficient of the fiber, or the loss of a connector or splice.
Fiber Attenuation Coefficient and Length
In addition to measuring fiber attenuation, this method is used to measure length.
To measure the length and attenuation of the fiber, we place the markers on either end of the section of fiber we wish to measure. The OTDR will calculate the distance difference between the two markers and give the distance. It will also read the difference between the power levels of the two points where the markers cross the trace and calculate the loss, or difference in the two power levels in dB. Finally, it will calculate the attenuation coefficient of the fiber by dividing loss by distance and present the result in dB/km, the normal units for attenuation. If the fiber segment is noisy or does not look straight, the OTDR can average the measurement with a method called least squares analysis (LSA).
Splice or Connector Loss
The OTDR measures distance to the event and loss at an event - a connector or splice - between the two markers. To measure splice loss, move the two markers close to the splice to be measured, having each about the same distance from the center of the splice. The OTDR will calculate the dB loss between the two markers, giving you a loss reading in dB. Measurements of connector loss or splices with some reflectance will look very similar, except you will see a peak at the connector, caused by the back reflection of the connector.
The OTDR can also use a least squares method to reduce noise effects and remove the error caused by the loss of the fiber between the two markers. The OTDR fits a line to the fibers on each side of the splice or connector and looks at the dB offset to measure loss.
To measure reflectance, the OTDR measures the amount of light that's returned from both backscatter in the fiber and reflected from a connector or splice. Calculating reflectance is a complicated process involving the baseline noise of the OTDR, backscatter level and power in the reflected peak. Like all backscatter measurements, it has a fairly high measurement uncertainty, but an OTDR has the advantage of showing where reflective events are located so they can be corrected if necessary.
Comparing two traces in the same window is useful for confirming data collection and contrasting different test methods on the same fiber. Comparisons are also used to compare fiber traces during troubleshooting or restoration with traces take just after installation to see what has changed. All OTDRs offer this feature, where you can copy one trace and paste it on another to compare them.
OTDR Measurement Uncertainty
Like all instruments, OTDRs cannot measure with perfect accuracy. Some of the measurement uncertainty with OTDRs is due to what is actually being measured. For example, when measuring the length of a fiber optic cable or the distance to a fault, the OTDR measures time and calculates distance based on the speed of light in the fiber or inversely the index of refraction of the fiber which determines the speed. Not all fibers have the same index of refraction, depending on the type of fiber and the manufacturer, and it is usually different for different wavelengths.
Group Index of Refraction (IOR) – Single Mode Fiber
The difference is only about 0.5% between the highest and lowest index of refraction, but that can mean an difference of 50 meters at 10 km, a significant error if one is trying to locate a fault.
The other problem with OTDR distance measurements is the difference between fiber length and cable length. Fiber is loosely fitted in buffer tubes and the buffer tubes are helically wound around the cable so the fiber will not be stressed when the cable is pulled.
But with this cable construction, the fiber length is about 1% longer than the cable length, a difference of 100 meters at 10 km. Since the OTDR measures fiber length, not cable length, this provides another error in measuring cable length, distance to an event like a splice or locating events in a long cable run.
If the actual cable length is known, and it can sometimes be read off the cable jacket or might have been recorded in the documentation when the cable was installed, one can use that to calibrate OTDR distances to actual cable distance. If OTDR traces are taken from both ends of a cable to an event or a fault, the relative distance to the event or fault can be estimated by using the ratio of the two measurements.
The biggest source of measurement uncertainty that occurs when testing with an OTDR is a function of the backscatter coefficient of the fibers being tested, the amount of light from the outgoing test pulse that is scattered back toward the OTDR. The backscattered light used for measurement is not a constant, but a function of the attenuation of the fiber and the diameter of the core of the fiber.
If you look at two different fibers spliced or connected together in an OTDR, any difference in backscatter from each fiber is a major source of error. If both fibers are identical, such as splicing a broken fiber back together, the backscattering will be the same on both sides of the joint, so the OTDR will measure the actual splice loss. However, if the fibers are different, unequal backscatter coefficients will cause a different percentage of light to be sent back to the OTDR.
If the fiber nearer the OTDR has more scattering (shown in the OTDR trace as attenuation) than the one after the connection, the percentage of light scattered back to the OTDR from the test pulse will go down, so the measured loss on the OTDR will include the actual loss plus a loss error caused by the lower backscatter level, making the displayed loss greater than it actually is. Looking the opposite way, from a low attenuation fiber to a high attenuation fiber, we find the backscatter goes up, making the measured loss less than it actually is. In fact, if the change in backscatter is greater than the splice loss, this shows a gain, a major confusion to new OTDR users.
While this error source is always present, it can be practically eliminated by taking readings both directions and averaging the measurements, and many OTDRs have this programmed in their measurement routines. This is the only way to test inline splices for loss and get accurate results.
If you are testing short cables with highly reflective connectors, you will likely encounter ghosts. These are caused by the reflected light from the far end connector reflecting back and forth in the fiber until it is attenuated to the noise level. Ghosts are very confusing, as they seem to be real reflective events like connectors, but will not show any loss. The best way to determine is a reflection is real or a ghost is to compare it to cable plant documentation. You can eliminate ghosts by reducing the reflections, for example using index matching fluid on the end of the launch cable.
The limited distance resolution of the OTDR makes it very hard to use in a LAN or building environment where cables are usually only a few hundred meters long. The OTDR has a great deal of difficulty resolving features in the short cables of a LAN and is likely to show "ghosts" from reflections at connectors, more often than not simply confusing the user.
Using The OTDR Correctly
When using an OTDR, there are a few cautions that will make testing easier and more understandable. Always use a long launch cable, which allows the OTDR to settle down after the initial pulse and provides a reference cable for testing the first connector on the cable. If testing the final connector on the cable is desired, a receive cable on the far end of the cable plant is required.
The OTDR operator must carefully set up the instrument for each cable. Again, good documentation will help setting up the test parameters. Always start with the OTDR set for the shortest pulse width for best resolution and a range at least 2 times the length of the cable you are testing. Make an initial trace and see how you need to change the test parameters to get better results. Some users are tempted to use the OTDR’s autotest function. More problems are caused by novices using autotest than any other issue in using OTDRs. Never use autotest until a knowledgeable technician has set up the OTDR properly and verified that autotest gives acceptable results.
More information on OTDRs and a free OTDR simulator to download.
Testing Long Haul, High Speed Fiber Optic Networks (Fiber Characterization):
Chromatic Dispersion, Polarization Mode Dispersion and Spectral Attenuation
One of the big advantages of fiber optics is its capability for long distance high speed communications. Attenuation at long wavelengths is lower. Singlemode fiber has extremely high bandwidth. Fibers can be fusion spliced with virtually no loss. High-powered lasers and fiber amplifier regenerators mean long distances are easily obtained.
However over very long distances, new factors in fiber performance become important. Chromatic dispersion, the dispersion caused by light of different wavelengths, and polarization mode dispersion, caused by the polarization in fibers, become factors limiting fiber links. Even the variation of fiber attenuation with wavelength can become an issue when networks using wavelength-division multiplexing are implemented. All 3 factors may need testing on long distance networks to ensure proper link performance. (See the page on fiber for descriptions of these fiber characteristics.)
Chromatic Dispersion (CD) in the Cable Plant
As with any other component, optical fiber performance parameters can vary from batch to batch, so a long concatenated cable plant with many different fibers will have a end-to-end chromatic dispersion which is an integration of the CD of all the individual fibers. Therefore fiber in long distance links will probably be tested for CD after installation or before upgrading a link to higher bit rate electronics.
Testing Chromatic Dispersion
There are several methods used for testing CD in fibers. All involve testing at a variety of wavelengths using several discrete sources of various wavelengths, a tunable laser or a broadband source with a monochromator in the receiver and measuring the relative speeds of the signals. The data taken at discrete wavelengths is then analyzed to calculate the dispersion in terms of ps/nm/km.
Test methods use phase delay or time of flight and generally require access to both ends of the fiber as well as a second fiber for synchronization of the two test instruments at either end. However, an OTDR test method is also used, since the OTDR measures time to convert to distance. Since the speed of light is different at various wavelengths, the OTDR would measure different distances on the same fiber at different wavelengths. Traces are taken at several discrete wavelengths and CD can be calculated from the difference in the distance data (or round trip time of the test pulse) obtained from the traces. This allows CD testing in the field from one end of the fiber.
All these methods have international standards for the test methods, instruments and data analysis.
Phase Shift Method
Pulse Delay Method
OTDR Test Method
Testing Polarization Mode Dispersion (PMD)
PMD is generally tested for fibers during manufacture or when being cabled. In the field, it is common to test PMD on newly installed fibers which are intended for operation at high speeds, generally above 2.5 Gb/s or when upgrading fibers installed some time in the past. Since PMD varies over time, a single test becomes an average and tests at a later time may be done for comparison.
There are a number of commonly used test methods for PMD, some of which are limited to the manufacturing environment, while others can be used in the field. Essentially, all the test instruments have a source which can vary the polarization of the test signal and a measurement unit that can analyze polarization changes.
Here are descriptions of the methods and relevant standards.
PMD testing is not an easy, reproducible, accurate test. The measurement uncertainty can be as high as 10-20%, as shown by testing done within international standards committees. These committees have concluded that all these measurement techniques are permissible, that there are factors in making these measurements that are not well understood, and the methods of data analysis are not without question.
All this uncertainty of PMD measurements has the effect of making comparisons between tests and test methods difficult. Variations are particularly high on tests of older fiber links. Presentations of field have even shown that variations in PMD can be correlated to wind speed for aerial cables.
PMD testing needs to be done on long links but the data must be analyzed intelligently to be of real use.
More on CD and PMD.
With the develoment of low water peak fibers, the possibility of transmission from 1260 to 1675 nm has been considered. This results from careful manufacturing of the fiber to reduce the water in the fiber (in the form of OH- ions) that causes higher spectral attenuation at around 1244 and 1383 nm.
Since one may want to use available fibers of unknown spectral attenuation for CWDM which uses lasers from 1260 to 1670 nm in 20 nm windows, it becomes necessary to test for spectral attenuation to verify the usability. At the water peaks, legacy fibers may have attenuation coefficients around 2 dB/km while low water peak fibers may be as low as 0.4 dB/km.
Testing spectral attenuation is done per TIA/EIA-455-61 or IEC 61300-3-7 with broadband sources like LEDs and a spectrum analyzer on the receiving end of the fiber. Calibration is done with a short fiber length, the instrument calculates the spectral attenuation on a long length being tested. The measurement of spectral attenuation uses instruments similar to those used for CD testing by the phase shift method, so some instruments do both measurements at one time.
Troubleshooting and Restoration
The time may come when you have to troubleshoot and fix the cable plant. If you have a critical application or lots of network cable, you should be ready to do it yourself. Smaller networks can rely on a contractor. If you plan to do it yourself, you need to have equipment ready (extra cables, mechanical splices, quick termination connectors, etc., plus test equipment.) and someone who knows how to use it.
We cannot emphasize more strongly the need to have good documentation on the cable plant. If you don't know where the cables go, how long they are or what they tested for loss, you will be spinning you wheels from the get-go. And you need tools to diagnose problems and fix them, and spares including a fusion splicer or some mechanical splices and spare cables. In fact, when you install cable, save the leftovers for restoration!
And the first thing you must decide is if the problem is with the cables or the equipment using it. A simple power meter can test sources for output and receivers for input and a visual tracer will check for fiber continuity. If the problem is in the cable plant, the OTDR is the next tool needed to locate the fault. More from FOA Tech Topics: Troubleshooting and Restoration
Getting Started in Fiber Optic Testing
Even if you're an experienced installer, make sure you remember these things.
1. Have the right tools and test equipment for the job.
2. Know how to use your test equipment
Before you start, get together all your tools and make sure they are all working properly and you and your installers know how to use them. It's hard to get the job done when you have to call the manufacturer from the job site on your cell phone to ask for help. Try all your equipment in the office before you take it into the field. Use it to test every one of your reference test jumper cables in both directions using the single-ended loss test to make sure they are all good. If your power meter has internal memory to record data be sure you know how to use this also. You can often customize these reports to your specific needs - figure all this out before you go it the field - it could save you time and on installations, time is money!
3. Know the network you're testing...
This is an important part of the documentation process we discussed earlier. Make sure you have cable layouts for every fiber you have to test and have calculated a loss budget so you know what test results to expect. Prepare a spreadsheet of all the cables and fibers before you go in the field and print a copy for recording your test data. You may record all your test data either by hand or if your meter has a memory feature, it will keep test results in on-board memory that can be printed or transferred to a computer when you return to the office.
A note on using a fiber optic source: eye safety. Fiber optic sources, including test equipment, are generally too low in power to cause any eye damage, but it's still advisable to check connectors with a power meter before looking into it. Besides, most fiber optic sources are at infrared wavelengths that are invisible to the eye, making them more dangerous. Connector inspection microscopes focus all the light into the eye and can increase the danger. Some telco DWDM and CATV systems have very high power and they could be harmful, so better safe than sorry. Read our page on Safety.
More on Testing & Troubleshooting Fiber Optic Systems
More good reading on fiber optic testing- The JDSU Reference Guide to Fiber Optic Testing: The JDSU Reference Guide to Fiber Optic Testing is written for fiber optic installers, project managers, telecom technicians, and engineers who need to understand, apply, and correctly measure and record the performance of fiber infrastructures. It is the best guide available for fiber optic testing and it's free. Download yourself a copy and read it!
Test Your Comprehension
More articles on Testing in The FOA Guide.
Table of Contents: The FOA Reference Guide To Fiber Optics