We're done with the initial alignment using the total station and the red laser. The final fine alignment should be done in chamber with ISCTEX in place.
Angle:
TMS yaw angle was adjusted by pushing the TMS SUS cube, and pitch by adjusting the sliding mass on the ISC table, such that the red laser beam from the total station is centered on the aperture on the ISC table. We used a small target placed in the aperture, but putting a target on the ISC table throws the PIT balance off. In the end we placed the second (dummy) target that is identical to the real target such that two targets are symmetrically placed around the center of the table.
Centering relative to ETM:
TMS centering relative to ETM was measured by the total station after putting the table/tele on the alignment stabilizing tooling such that the set screws barely touch the ISC table, then mounting the large target on the large input aperture of the telescope secondary plate.
TMS was 5 to 6mm lower than the ETM and off to the outside of L shape of the IFO by 1.5mm or 2 according to Jason. So the centering error is
sqrt(6^2+2^2) = 6.3 mm
in the worst case.
This is very acceptable both for the green and IR beam.
Green:
Outside of the arm on the AR side of the ETM, the green arm mode has the waist radius of about 6mm, the Rayleigh range of about 200m, the divergence angle of about 6mm/200m=30urad, and the waist is 1500m away from the ETM into the arm (see e.g. T1200200).
If we inject a beam from the center of TMS (6.3mm offset from the arm mode at the ETM) such that there's no lateral displacement at the waist, the misalignment angle would be 6.3mm/1500m = 4.2urad.
The power mismatch would be (4.2urad/30urad)^2 = 2%. This is nothing.
IR:
The requirement for the centering tolerance for the IR beam is much larger than the green, as the only important thing for IR is that both of the QPDs will see the light. Even if the IR beam is half radius off of the center of the QPDs that would be OK. Having said that, just for the completeness I'll show some numbers.
Outside of the arm on the AR side of the ETM, the IR arm mode has the waist radius of 8mm, the Rayleigh range of 200m (same as green), divergence angle of about 40urad, and the waist is 1500m away from the ETM into the arm.
If the TMS is aligned as described above (no lateral displacement at the waist but 6.3mm offset at the ETM) the power mismatch parameter would be (4.2/40)^2=1%.