Many people do diy, so they have a leg up on it, but getting into the case on some would be interesting. I have Radions, and there are screws, but mine talks to the manufacturer every time I run changes: it has a USB you plug in, linked to your computer (I don't have the interface, just use a laptop) and I think probably it is capable of detecting a malfunction and reporting same. Not sure at all. You might ask Ecotech whether they do service on a dead bulb. As these units do age, there's still a lot of 'room' in the program---I run mine at 50% of what they can output, and the corals are happy. So just ramping up the percentage might hold you a while, I have absolutely no idea---Again, you might ask a specific manufacturer. If they last more than 5 years, you'll have spent only a certain amount for the unit per year---and I know I was spending about 100.00 a year on regular MH and actinics. So I figure what with the less power draw, no ballast to buy separate, and all, the per annum cost even at a 5 year life wouldn't be killer, compared to the annual bulb bill. I have no day/night temperature fluctuation, temp is rock steady, and colors are variable during the day and very nice. If they should run 50,000 hours I'd be ecstatic.
SOme other issues to consider: ease of setup, 'spread' and 'depth'. An LED is much more straight down than sideways, but some seem to have more spread than others: I was able to get away with a single unit on my wedge bow tank, because my rockwork is essentially a pyramid. And how deep they will reach while still delivering coral-capable light is also variable by brand. If you have a very deep or very shallow tank this also is something to take into consideration.
