Optical Time Domain Reflectormeter
When installing a fiber optic system or cable plant the system must be tested to verify performance. This is carried out to ensure that light will pass through the system properly. Three types of optical testing are carried out routinely in the field: continuity, power and Optical Time Domain Reflectometer (OTDR) testing. A distinct advantage of a fiber optic system is its electrical immunity. Because fiber cable is nonmetallic, it cannot emit or pick up electromagnetic interference (EMI) or radio frequency interference (RFI). Both EMI and RFI can cause significant problems with metallic conductors.
An Optical Time Domain Reflectometer is essentially an optical radar: it sends out a flash of bright light, and measures the intensity of echo or reflections. This weak signal is averaged to reduce detection noise, and computation is used to display a trace and make a number of mathematical deductions.
This instrument is really good for measuring points loss on installed systems, where it is used to find faults and measure point losses such as caused by splicing. However to do this accurately is more complicated and time consuming than is commonly supposed, since a measurement should be taken from both ends of the system, and then averaged. If this is not done, spurious excess losses and "gainers" may be recorded where different fibers are joined, resulting in wasted splicing effort while non-existent faults are "repaired". This is a particular issue when measuring the fusion splice joints, where the loss is small, and the adjacent sections may have fibers with different intrinsic backscatter characteristics.
OTDRs can be used for return loss measurements, although quoted accuracy is not very high.
This is most commonly used during installation acceptance and maintenance of outside plant cables. In this role, it is likely to be used to identify point losses, the length of various cables, and to measure return loss.
1. Interpreting the trace requires too much skill for most field technicians . These people must rely on the built in automation program to compile data tables, and may have little idea what to do with the trace. Since only highly skilled users can set up the parameters for this automation, in some circumstances most users can get into major difficulty.
2. Acceptance verification is relatively easy, since standard procedures and automated measurement can generally be used. However using the same instrument for fault finding may require a totally different class of operator, who understands how control the measurement process in great detail, and also interpret the trace accurately.
3. Because of the skill requirements, the majority of organisations end up with a small number of identified "experienced" operators, who train others, and are called out to problem situations.
4. Limited ability to separate multiple point losses that are fairly close together. This problem happens quite regularly in practice, and is due to the "dead zone" effect. Although instruments may advertise an event dead zone of say 5 m, this is only under specific conditions. In practice the dead zone may go up to a km for long distance work. This makes these instruments of less use on short systems. Other tools, such as a visible laser, may be required to precisely identify the fault. This has become a big issue as the fibre count in cables has increased, which has caused an increase in the requirement to avoid disturbing already installed closures and racks.
5. The distance measurement accuracy is only about 1 - 2 % at best. For example a displayed result of 12.1567 Km is actually more realistically 11.91 - 12.39 Km, an uncertainty to field staff of nearly half a Km. The reasons for this are fundamental and are due to variations in cable manufacture and index of refraction. So a measurement of 1 Km, is typically not 1 Km of cable, and certainly not the exact route length. Use of a Cold Clamp can greatly improve distance accuracy.
6. Limited accuracy when determining the end to end loss of a system. It typically makes a poor job of measuring the loss of the end connectors, which are themselves a cause of problems.
7. Limited use on "passive optical network" systems that use couplers or splitters to connect one source to multiple locations. This is because measuring in this configuration only works in one direction, and so this method cannot be reliable.
8. Cannot be used in compliance with new multimode fiberoptics loss measuring standards, which mandate the use on an LED source with defined characteristics.
9. Accidental connection to a receiver can damage the receiver due to the high instantaneous power levels. There can be some optical safety issues associated with the high pulse powers in these instruments, which often exceed +20 dBm.
10. Factors to look for are now typically ease of use, quality of automation program, good local support, and compatibility with previously acquired measurement file types.