How I can launch light into PM fiber A polarization maintaining fiber preserves the linear polarization of the light at the output, the light at the input must be linear polarized to see signal at the photodetector... However I can't see anything.
What ist the standard procedure?
 A: Polarization maintaining fibers are (as far as I know always) single-mode fibers. It is inherently difficult to efficiently couple light into single-mode fibers because, by definition, they will only support one specific optical mode which you have to carefully match with your input beam. So this mode matching will most likely be your main challenge. Getting the polarization right can be a separate step and should be easier.
Lets assume you start with a nice monochromatic single-mode gaussian beam, e.g. from a solid-state laser. To get (most of) this light into your single-mode fiber you have to make the beam mode and the fiber mode overlap which requires the tuning of six parameters:


*

*the lateral position of the beam axis in horizontal and vertical direction

*the angle of the beam axis in horizontal and vertical direction

*the longitudinal position and the size of the beam waist


Getting the waist size and position right will require some planning and knowledge of the fiber size and beam parameters of the input beam. Usually you need one or two lenses to collimate the beam onto the fiber end. Sometimes your fiber end already comes with collimating optics. Software can help you with the planning (I have used JamMt before, but there might be better ones).
The alignment degrees of freedom can be controlled by sending the input beam over a pair of steering mirrors or by mounting the fiber end on a 4-axis table (or a combination of both).
Once you have everything set up roughly (e.g. by eye or by sending some light backwards through the assembly) you should make sure you have a way to monitor how much light gets into the fiber. Then start fine-tuning the alignment and mode-matching parameters to gradually increase the light power in the fiber. After the initial optimization of the individual parameters, you will need to do "beam walking": Offset one parameter (this will make things worse) and compensate with the others as best as possible. If this made things better, continue in the same direction, otherwise go back and try the other direction.
This can be a fiddly process if you aim for coupling efficiencies beyond say 50%. Luckily, for many applications imperfect coupling is enough, as long as you get some light into the fiber.
Now on to the polarization: The PM fiber only works properly if the polarization axis of the linearly polarized input beam is either parallel or orthogonal to the fiber's polarization axis. Injecting linearly polarized light at a any other angle will generally produce elliptical polarization after transmission through the fiber. If the correct injection angle is unknown, it can be found by analyzing the light after the fiber with a polarizer and turning the injection angle (with a half-wave plate) until the resulting light is linearly polarized.
I hope this helps as a general overview. More specific advice might be possible if we know your particular setup and application.
