Abstract

Obtaining good agreement between multimode fiber loss sets has proven difficult. The standard deviation between locations (i. e., test sets) is limited to about 0.2 dB/km by the problem of reproducing test set launch conditions.¹ This particular problem should not exist with single mode fibers, since only one mode can propagate. Consequently, one would expect the interlocation standard deviation to be less than 0.2 dB/km when measuring single mode fiber attenuation. Indeed, because the loss of single mode fibers is frequently only a few tenths of a dB/km, a loss measurement accuracy of a few hundredths of a dB/km will be necessary. We have recently completed a loss measurement round robin involving seven Bell System test sets. The results indicate that this level of accuracy is indeed attainable with single mode fibers.