In a standard single-mode patch cord, the two wavelengths used for insertion loss testing are 1310nm and 1550nm. All single-mode fibers work very similarly at any wavelength - that is, you don't need to buy fiber by wavelength, one fiber will fit all wavelengths. So, if your fiber optic assemblies are properly constructed using quality materials and good technology, the insertion loss values for any given fiber optic connector should be very similar when tested on a 1310 or 1550.
This has led some manufacturers to use only one wavelength to test their products, and only test two wavelengths when a customer specifically requires it. Testing at both wavelengths requires additional equipment, which in some cases appears to be nothing more than a "necessary evil". However, there are benefits to making it standard practice for testing all fiber optic cable assemblies at 1310 and 1550: the variation in insertion loss between the 1310nm and 1550nm test wavelengths can be very helpful in identifying serious problems with the product and/or process.
A useful tip to troubleshoot any single-mode insertion loss testing of your product is to keep the following in mind:
(1) 1310nm is more sensitive to alignment problems
(2) 1550nm is more sensitive to fiber bending problems
(3) 1310 and 1550 are similar
Fiber optic assemblies will test about the same on a 1310 or 1550 if made properly. Insertion loss results for the 1550 are typically a few percent better, due in part to its lower fiber attenuation. Typically, the insertion loss value of a connector is ~0.01-0.05 dB better at 1550 than at 1310.
(4) 1310 is higher than 1550
The insertion loss at 1310 is significantly higher than at 1550 for the A connector or the entire product design, indicating a possible problem with core-to-core alignment between two mated ferrules. The difference may be small and practically acceptable. The greater the misalignment, the greater the insertion loss at the 1310 compared to the 1550. Misalignment can be caused by a number of factors, the most common being contamination of the product and test assembly, or poor fiber core-to-metal ferrule concentricity.
It is hoped that contamination can be eliminated, and previous manufacturing processes can be improved to eliminate contamination prior to testing. However, poor concentricity is often caused by using an oversized ferrule, so insertion loss cannot be improved without replacing the connector. "Relatively oversized" is relative: the diameter of the ferrule hole is larger than the outer diameter of the fiber, the fiber will at most be able to sit on the center side of the ferrule, so the expected insertion loss (1310) is greater.
(5) 1550 is higher than 1310
A connector or entire product design has significantly higher insertion loss at 1550 than at 1310, indicating that there may be a stress point on the fiber somewhere in the fiber assembly - most likely the fiber is bent beyond the working bend radius, or within the product The fiber "pinches" or microbends somewhere. The higher the stress (more bending), the higher the insertion loss at 1550 compared to 1310. However, while the aforementioned core shift issues are usually a normal consequence of raw material selection, any excessive stress directly applied to the fiber represents a serious problem. Product reliability is at risk, so an IL value @ 1550 is especially important for monitoring and troubleshooting.
It is good practice to test all products at 1550, even if "reduced bend radius" fibers are present. Doing so could uncover serious product defects, especially in products where fiber is routed in a wiring closet (such as a cassette or fiber closet) or functional area fan-out transitions). If your product has a significantly higher insertion loss of 1550 than 1310, your product is most likely stressed by the fiber, so you need to understand why.
