alexan.staley@LIGO.ORG - posted 16:03, Wednesday 11 December 2013 - last comment - 09:36, Thursday 12 December 2013(8912)
Fiber power from MSR to EX
(Alexa, Sheila)
1. Input to MPC @ MSR: 280uW
2. Output of MPC @ MSR: 240uW
3. To EX @ MSR: 170uW
4. Output @ EX: 45uW
5. Outpt of patch panel on ISCTEX: 42uW
6. On table @ fiber output: 25uW
This is not good. 
. The loss between 2&3, 3&4, 5&6 are much higher than expected.
If you are measuring the total power, then you might try mandrel wrapping the fiber at the input; it may be that you are measuring power from higher modes in the fiber that will dissipate rapidly in the first part of the fiber chain. A mandrel wrap will dissipate these modes early, so you will be left with a more accurate reading of the power that will propegate the full length of the fiber. This will be particularly true for 1064nm light; singlemode fiber is designed for 1310nm - for 1064nm, it will behave more like multimode fiber.
Even so, something looks particularly bad with the insertion loss for the jumper between the polarizer box and the patch to EX (2&3 in your list). That's probably a bad or dirty fiber.