When manufacturers are asked what is the biggest problem with installing of fiber optic components and systems, they invariably say “testing.” Practically everything they do depends on testing. During development, engineering and manufacturing in their labs, they test product performance. Testing reveals the most effective installation processes that they document and teach to their customers.


But once the products are purchased and installed, testing—usually performed by the installer—verifies the installation quality. Whenever questions about product performance arise after installation, the initial judgment is based on that testing. If there is a problem, more often than not, the dispute focuses on how the components were tested even more than how they were installed.


Why the controversy? 


Why is testing so controversial? Don’t today’s instruments give us an easily decipherable digital number we can read, store and print for our customer? Yes, they do. But what does that number mean? How close is it to the actual value we are trying to measure? What can change the reading and cause errors? Would anyone else making the same measurement—e.g., the manufacturer’s lab personnel—get the same reading?


The answers to those questions can be found in the study called “metrology,” which is the science of measurements. Don’t stop reading for fear I’m getting too technical, because I promise I’m not. Metrology is quite simple and logical. Electrical contractors (ECs) should learn about it because it is relevant to many of the things they do beyond fiber optics installation. It can be found in building construction, electrical installation or any part of business that involves taking measurements.


For measurements, accuracy is the topic of most concern. Accuracy refers to how well the measured value agrees with the actual value of what is being measured. For example, when measuring the length of a fiber with an optical time-domain reflectometer (OTDR) and 10.0 kilometers (km) is measured but the actual length is 10.1 km, the accuracy is 99 percent, or as we generally state it, your error is 1 percent.


The variations we get in our measurement indicate the uncertainty of the measurement or its accuracy.

“Precision” is another term widely used in measurements. A measurement is precise when repeated measurements produce nearly the same result. In the OTDR example, a person might look at the display, move the markers slightly and measure 10.0 km, try again and get 9.9 km, again and get 10.1 km and again to get 10.2 km, caused by the uncertainty in placing the cursors. Or, perhaps the trace is noisy and several measurements are taken with the same variations, or the OTDR is allowed to automatically record the distance as an event in several tries as 10.0, 10.1 and 9.9. All of these measurements have a variation of about 1 percent, which we call the precision of the measurement.


The variations we get in our measurement indicate the uncertainty of the measurement or its accuracy. In the 1980s, when I worked with the scientists at the U.S. National Bureau of Standards (now the National Institute of Standards and Technology) to develop a standard for optical power to calibrate fiber optic power meters, they always corrected me when I said “accuracy.” They preferred to say “measurement uncertainty” because we generally refer to measurement errors rather than how close we are, which sounds strange. So I continue with “uncertainty.”


There are two major contributions to measurement uncertainty: random errors and systematic errors. Using the example of measuring length with an OTDR, random errors would be variations caused by the operator variations in setting the marker positions or noise in the OTDR trace, affecting automatic measurements. Random errors would cause random variations in the length.


But the fiber length being measured is a function of the speed of light in the fiber, because the OTDR measures time and converts it to distance. The OTDR calibration uses the index of fiber refraction that determines the speed of light in the fiber. What happens if we use the wrong index of refraction? Every measurement will be short if the index of refraction is too high and long if it’s too low. This mistake produces a systematic error where all measurements are erroneous in the same way.


Next month, I’ll cover more about fiber optic measurement uncertainty.