In order to place the waist of the pump beam on the right position, I characterized the pump beam using the beam profiler and moving it with the translation stage. 

Laser settings: I = 1A, P = 210mW , hwp_angle=0, not chopped.

I aligned the pump beam to be parallel to the z-axis translation stage within 0.2mm on the whole translation range 250mm (±0.8mrad).

The beam profiler position is set 0 at the position of the KAGRA-size sample first surface (see picture).

Without the focusing lens, the waist is 377um at -0.88m

Using the focusing lens f=125mm placed at -60mm, the waist becomes w0=53um  at 0.081m

According to this calculation the deepest measurable point is 0.08*1.81=0.149m for sapphire (and less for silica).  Since KAGRA-size thickness is 15cm, the cross point (which will be at the pump waist) cannot go beyond the second surface.  There is not enough space to place the lens closer to the sample, therefore I need to change the focal lens.

Making a quick simulation with JamMt, a focal lens of 150mm will move the waist to 0.111m (w0=62.8um) without sample, With a 15cm-thick sapphire substrate the waist will move to 0.178m, and with a 15cm-thick silica substrate, the waist moves to 0.16m, which is enough to measure the entire thickness of the sample.