Yesterday after tried moving many times all three lenses in the bench, it seems the first 100mm lens has very large effect on the beam shape. So we decided to change this lens into a large focal length one and re-design the telescope.
This is the new telescope parameters:(origin still at the first line of hole out of the cavity)
L1 f1=200mm z1=8cm
L2 f2=200mm z2=56.5cm
L3 f3=175cm z3=99cm
With this new telescope we had a better beam at the window of PR chamber and also before the beam goes inside the arm, the previous one produced a beam a little bit longer in horizontal direction. So we decided to go on with this configuration.
Then today after we realign everything on the bench, took some measurement to check the difference between the theoretical value and practical one is inside the tolerance or not. All the points we measured are equal or smaller than the result of the calculation. We also added the two telescope mirror inside the chamber to check if the beam in the tube has the size we want or not. At the two inch mirror we had a beam radius about 1mm which is good, and the largest size inside the arm is about 1.3cm.
The other thing we did is we measured several points on the bench after the third lens. With these data, we calculated the beam waist and simulate the propagation of beam in 300mm, but without the two telescope mirror. From the picture we can see that the beam shape on the bench can not produce astigmatism at 300m.