Insertion Loss Testing
This domain covers the acceptance test for a installed cable plant: end-to-end insertion loss measured with a light source and power meter (an OLTS). The exam tests reference methods, multimode launch conditions, link versus channel, and the reasons readings go wrong.
Setting the reference
Insertion loss is a relative measurement, so you first zero the setup at each test wavelength. The three standard reference methods are the two-jumper (Method A), one-jumper (Method B) and three-jumper (Method C) configurations described in TIA-526-14 and IEC 61280-4-1/4-2. They differ in how many test-jumper connections sit inside the reference, so the same link can read differently under different methods. The one-jumper method is the most repeatable and therefore preferred; the three-jumper method is used when the link's connector type differs from the test jumpers.
The rule that protects your data: choose one method, record it on the report, and use it for every link in the project. Mixing methods makes results incomparable and can push a passing link over a limit.
Multimode launch condition
IEC 61280-4-1 requires a controlled launch for multimode loss testing - an encircled-flux-conditioned source, or a mandrel wrap (commonly five turns around a 25 mm mandrel for 50/125 fiber). An uncontrolled overfilled launch excites high-order modes that die quickly, so the measured loss comes out optimistically low.
Permanent link versus channel
A permanent link test covers the installed cable and connectors but not the equipment cords; a channel test includes them. They have different limits, and a channel result can never be compared with a permanent-link limit. Count connectors and splices before you compare anything.
Direction and repeatability
Field insertion loss is normally measured in one direction. Connector repeatability of ±0.2 to 0.5 dB is normal, so a reading sitting on the limit is not a reliable pass - re-clean, re-seat, and re-measure before deciding.
Worked example
A 1310 nm single-mode link has a design budget of 2.0 dB. After referencing to 0.0 dB, the first reading is 2.6 dB. The technician cleans and inspects both end connectors, re-seats the test jumpers, and re-reads 1.7 dB. Reasoning: two contaminated mating points contributed roughly 0.9 dB; with clean connectors the link passes with margin. The original 2.6 dB would have triggered an unnecessary re-termination.
Traps that cost marks
Forgetting to set the reference, or setting it at the wrong wavelength; a negative loss reading, which always means a bad reference or a disturbed jumper; APC jumpers on a UPC link; a jumper bent tighter than its minimum radius during the test; reporting a single reading as final when it sits within connector repeatability of the limit.
Key points
- Insertion loss is relative: reference to 0.0 dB at each wavelength first.
- One-jumper reference is preferred; never mix reference methods in one project.
- Multimode testing requires encircled-flux or mandrel-controlled launch.
- Permanent link and channel have different limits; count connectors before comparing.
- A negative result means the reference or the test setup is wrong.