Hehe... Back then was a totally different era. If you had more than a watt of RGB, you were king shit. These days that's a typical lumia projector! Crazy...
No. 635 nm red has always been a direct-injection diode. There is a 671 nm red that is DPSS, but it's rather uncommon. (It's also not very good for laser show use, because 671 is such a deep red that you can't see it very well.)Does that mean that the 635nm from that time was dpss?
Yes, the coatings were really good. That's why the dichro was so expensive. To put this in perspective, back then a high-quality dichro from Edmund's Optics cost around $35 in quantities of 1. These special Semrock dichros were over $150 each, and we had to buy 10 of them to get that price. I imagine they're quite a bit more expensive today. Also, those Edmund Dichros had a dead-band of 30 to 40 nm between full-reflect and full-pass. This dichro had a dead-band of less than 12 nm. So, yeah, it was a very precise optic, and it had a price to match.About the efficiency he states: "Efficiency will be around 98% for the reflected beam, and 96% for the transmitted beam." I don't know much about lab lasers or their optics but I do know that is insanely efficient given that the difference in wavelength is only 15-25nm.
As for the issue of buy vs build - you *can* get some very nice stuff on the retail market. But if you go super-cheap, you'll likely be disappointed. Conversely, building your own projector is a great way to learn, and you can usually save money as well, but it isn't the best solution for everyone.
And regarding A1XD's projector, I'm reasonably certain they used a PBS cube to mix the two reds. A custom dichro would be way too expensive, and so long as you don't need really high power, it's just as easy to use a PBS cube.
The reason we needed this dichros back then was because many people were using the long open-can 660 nm red diodes. They had fantastic beam specs, but they only made about 200 mw tops. So lots of people used a PBS cube to combine a pair of those diodes to get ~ 400 mw of red. That sounds like a lot of red, but remember that 660 nm is very deep on the spectrum. Our eyes aren't very sensitive to it, so it appears to be less bright. So typically everyone was always looking for ways to get more red.
However, if you were already using a pair of PBS'ed 660 nm diodes and you still wanted more red (and you didn't want to get into knife-edging multiple beams together), then you needed a second red wavelength in your projector and a dichro to mix it in with the 660 nm red. You couldn't use another PBS cube to mix in the new red wavelength, because the beam from the two 660 nm reds was already randomly polarized after exiting the PBS cube used to combine those two. Also note that in 2008, the 637 / 638 nm red diodes hadn't been released yet, so the best candidate was either 642 nm (which was right at the cut-off point for the dichro), or 635 nm, which was cheaper and easier to find.
Adam