A normal spec for a splice is usually written in two ways, each event (splice) cannot exceed 0.10 when tested bi-directionally, and all splices when averaged cannot exceed 0.07 bi-directionally. This is a generec statement we see on many of our jobs. We splice with the Fitel as well because it has proven itself in variety of conditions, some extreme, and gives an average 0.02 bi-directional loss. Our remake percentage is less than 1% after testing and this machine I feel has proven itself better than Fuji, Sumi, and Siecor. All core alignment splicers do basically the same thing. All manufacturers state a average splice loss of 0.03, but that spec really doesn't matter since the machine has no clue as to the actual loss of the splice. The OTDR shows you reflectance, not actual loss, but is what we use to say a particular splice is good or bad. I'm not saying -0.04 is good or 0.04 is bad. If your spec says 0.07 bi-directionally, then that splice is good. If you are only testing one direction, then that situation opens you up for losses from core mismatch (Gainers/losses) that will have you chasing your tail to meet spec. A spec of 0.02 was mentioned, this spec is one that would raise my eyebrows. It is a very tight spec that most customers don't need, and most contractors don't want. We have worked on some of the most technologically advanced systems with some of the highest data rates, and this spec wasn't needed. There are the "LIDS" splicers that inject light into the fibers and align accordingly such as the M90 5000 and 6000 series. This machine is good, but slow, expensive, and for the most part not necessary in building a perfect network. I have the FMTAP (Tektronics) program if you want to send me an private message. I will give you my e-mail so that I can analyze your trace and call you directly.
This is CABL.com posting #142478. Tiny Link: cabl.co/mLec
