|
Testing for optical transmission of a multiple channel fiber optic block at a number of different geographic locations presents serious problems in correlating the results. In one of these situations, there was a requirement to test the light transmission of a single-input, 13-channel-output fiber optic block. Each output channel was to be tested twice: once within . 010 inches from the output surface and again at .375 inches from the exit surface. In addition, the test cycle was to be held to six minutes (10/hour). The testing was to take place at two vendor locations and at International Business Machines Corporation's Systems Manufacturing Division plant at Poughkeepsie, New York. The test results had to be accurate to 10% from tester to tester and repeatable within 5%. Specifications were met by using a single photovoltaic cell to measure the input light to the block and to take both of the channel output measurements. The photocell responses were recorded on a strip chart recorder. By adjusting the recorder to read the photocell response to the input light as full scale, and by using recorder paper graduated from 0 to 100, the optic block outputs were displayed directly in percent transmission on a hard copy output. In order to shorten the test cycle and to reduce the influence of the operator on the results, the equipment was automated. Thus the photocell was automatically indexed across the output face of the fiber optic block. To insure uniformity between the pieces of test equipment, the mechanical components of the test fixture and indexing mechanism were held to precise tolerances. The results were a maximum tester to tester correlation error of 9.4%, a repeatability error of less than 5% on each tester, and a two-minute test cycle per unit.
|
