@bluumax, thanks for pointing that out with the sensors, I actually have my machine in the study :-), so I'd planned on building a fully enclosed MDF shroud for it and then duct the outside dust extractor to it through the wall ( I already have a cabling entry so I can add another 35mm pipe easily enough). As the Proxxon is way quieter than most mills, it would be more to stop dust billowing than noise control, but that's a great point about dust getting into/on the optos. I'll have to think on that more but I think it would be easily solved by several miniature dust extraction tubes fitted above them to collect any particles, as I will be building an extraction shroud to fit around the base of the mill so dust will be extremely minimal i'm hoping.
I'm using an old Festo extractor so this thing will suck the shag out of a carpet! (love Festo gear!)
With the questions others have asked about repeatability, earlier in this thread I upped some shots I did on MDF and thick cardboard, I routed spiral patterns over the top of each other over several days, and using my magnifying lamp couldn't discern any devaition in the routing, so I would probably have to agree with Luke Chan that it is highly repeatable, although I did all my test spirals in one run without doing a different pattern inbetween, even though they were done over several days.
If however you did one run, then another run of something else, and then went back to the first one, I don't know if this would have any effect on any possible homing drift? I doubt it but it is possible.
Still, for what it's worth, I wouldn't hesitate buying one again as I think it's an amazingly affordable piece of (from what I can tell at this point anyway) machinery!
I'm still trying to get to grips with EMC and Mach3, and creating PCB's using Eagle, so i'm a little way off really getting into anything meaningfull, other than the tests I uploaded shots of earlier.