At first take, that low of a figure hit me as too low for that diameter expander lens, with a 6 mm diameter collimation lens alone, that diode produces about 2.6 mRad which in my opinion isn’t good at all, as the beam turns into a wide rake 300 meters out and more at that amount of divergence. However, the 6x beam correction must be why you can achieve that low of a divergence.

However, thinking about this more you very well may have achieved that low of a divergence. Here is my thought process: If your beam has a divergence of 2.6 using a 6 mm collimation lens with roughly a 4 to 5 mm beam width diameter and you expand that diameter to 35 mm, as an example, that’s roughly a factor of 7 to 9x expansion. Take that 2.6 mRad figure and divide it by 9 and you get a divergence of roughly less than .3 mRad which really isn’t that far off from what you found, but since you expanded one axis of the beam prior to the Sanwu beam expander which for one axis, reduces the divergence below 2.6 before being expanded, I can now see you very well may be getting that low of a divergence figure.

Although there should be two different divergence figures, one for the fast axis and one for the slow axis, unless the beam correction C lenses perfectly correct the two to exactly match together, one side will have some amount of difference. I’ve yet to see anyone call out both figures for these direct to wavelength diodes, so for most if not all of our hobby laser pointers, it’s unlikely both sides are perfectly matched. Regardless of all of this, or if one side is slightly different, the divergence figure either way is awesome, good work. You have my address