I think that is where possibly some are saying that these are not usable. I am just after 500mW, corrected to fit on some 3mm mirrors. I don't care if I lose 75% in all the optics (collimaters, prisms, wave plates, PBS and pinhole). If I need 2W to manipulate it down to a usable beam and get 500mW, I would be fine with this. The price of these will still be cheaper. These have also been overdriven by a lot and hold up well. Of course, overdriving them will shorten the lifespan, but if you only get 2500 hours out of them that is still a lot of time (unless you do gigs every night). I use my projector in the garage so, not a lot of use. I just want something to go with my 300mW green and my corrected 445, which after optics, is about 600mW. My LW 640nm is not able to keep up without attenuating the blue a little and the green a lot. 500mW of 635 would look mighty sweet.
I do see, however, Steve's point here. When you start stacking beams and try to get multi-watts out of them, you are going to get a shit beam. Stacking too many diodes is pointless if you are going to pinhole. It may be possible to run a quad, knife edge the pairs tightly and combine them in the same orientation via wave plate and PBS, correct them via prism and then pinhole with a slightly larger diameter aperature (say 4mm), this is what I have in mind for my setup. Beyond that, kinfe edging a quad, or knife edging two groups of 3 or more and PBSing them... now the pinhole idea will start cutting off the additional beams and would be the same effect of trying to make a sailboat move by aiming a large fan at the sail from inside the boat. So yes, I can see where Steve is coming from. 2W of graphics capable red... I think we would seriously need Indiana Jones and his dad to help us find that grail.
If you're the smartest person in the room, then you're in the wrong room.