Reproducibility Test of Machined Connector
S. Razin tested one of the 12 fiber machined connector pairs
for reproducibility of light transmission when the connection
is broken and remade. MSU loaded wls fibers into the surface mount
half of the connector. Four of the fibers were damaged by the time
we started the test. The other half of the connector was loaded with long
clear fiber which connected to a a barrel style pmt disk. Stepan connected
the clear bundle to one of our PMTs. After mating the connector he
inserted each of the 8 good green fibers successively into a test tile
with straight fiber groove. He illuminated the tile with a radioactive
source. Spectra were obtained and the mean number photo-electrons
were extracted. The connection was broken and remade 4 times for
a total of 5 measurements. The measurements are summarized in
the table and figures below. Typical statistical errors on the
measurements were about 1.5%. When averaged over the full connector
it looks like the first connection was 2% low and the last about
2% high compared to the average. When all the fibers and connections
are viewed as one body of data the sigma is 2% implying about 1.5%
variation over that expected from the statistical error.
Jim Sowinski
