In the Oclaro diode thread I posted about my experience with beam manipulation with the P73 diodes. It got me to thinking about low order aberrations ie the ones we can afford to do something about. I isolated a single P73 diode ran it at 500 mA and measured the near field dimensions at 6mm x 3.2mm and @ 16.0 M the spot measured 14mm x 4mm. You can calculate the divergence if you want. I also noticed that no mater how carefully I adjusted the cylinder pair and the two 75mm PCX lenses there was always a significant amount of "junk" around the far field spot. This is why I like spatial filtering so much. But, I digress. I then started reversing the orientation of the lenses ie which face was closest to the diode and...
EUREKA!
All this should have changed is the spherical aberration that the lenses add, but because there is a cylinder pair and a radially symmetric pair then you can add or subtract negative or positive spherical aberration in any rotational orientation you require. In my case, I even changed the cylinder pair FL's. I didn't change the magnification, but both lenses were increased in power allowing a stronger spherical aberration compensation. The source of the spherical aberration is unclear, but likely due to the complex nature of the diode emission, the housing windows (in my case) and the thick glass cube when a PBS is used in the uni-axially expanding beams. In any case, I suppose you want the numbers? The near field remained at 6mm x 3.2mm and the baseline as well @ 16.0M, but the far field decreased to 10mm x 4 mm and was much crisper with significantly less "junk". It looked a lot better! I have ordered some more lenses including meniscus and bi-convex lenses to see if I can carry this any further.
The lenses I am using are the 4mm Optima aspherical collimators, a 10mm PBS cube, a -25mm cylinder @ 6mm from the cube face (curve away from the diodes), a 100mm positive cylinder (curve toward the diodes) @ 50mm from the negative cylinder. The first 75mm PCX is located 50mm later and both 75mm PCX lenses have their flat surfaces closest to the mid point focus. The cylinder lens pair orientation goes against convention, but the spherical aberration introduced seems to counteract what is introduced earlier.
Calculations of divergence can be misleading because as the beam is expanded in the near field, the far field shrinks and the differential rapidly disappears, but these results are pretty impressive. 1/4 mRad by zero as is and 0.43mRad by zero if the beam was shrunk to fit a 5mm width.