I finally got the lens stable enough and far enough away from the fiber (as to not melt it) but getting that to where I can test the growth of something is a bit different.
The setup works as a flash light (and death-ray) but does little given the fibers short length and it being daylight. Really the sun tracking and IR are show stoppers. Unless you want a collector the same size as the area you're lighting you need it to track the sun and without limiting the IR you are only using about 30% of the collected light -- even then I think these fibers would continue to melt in a couple hours if I could keep tracking the sun.
Oh and I did contact those guys about a custom IR lens.... note to self: when people laugh at your questions... it's not a good sign. I got the impression the conversation was likened to a poor kid talking about beefing up his station wagon to a race car driver. Which is ok. God knows I've been there before with web clients. But dag-nab-it, I want that lens!... and don't have a $200k budget for research, designing, testing, designing, bla-de-bla-bla.
[edit] Actually I did think about going small on it. If I put normal plastic caps (lens) on the end of fibers (which they sell.. cheaply) and spread them out on a grid (on roof again probably) it would probably be low enough heat and high % of light captured. So, that's where I've been. Deciding if there is a whole different approach that could be taken -- for uber cheap of course.