I wasn't sure how deep to make the epoxy but 5mm, maybe a touch less, at the deepest seemed to be about right while leaving an adequate depth at the shallowest points. I wonder if the roughly 7 thou variation is due to the change in depth along the length of the rail - there is a slight dip in the epoxy surface where the epoxy is deepest. Might be coincidence - don't know. I was concerned that the epoxy should be deep enough that the self-levelling would work; I had a feeling that the epoxy might not flow quite as freely if it were too shallow. Again, this is conjecture and not tested. I was not concerned about depth as far as tapping was concerned; the epoxy drills easily (rather more easily than the steel, anyway) and I would not expect unfilled epoxy to hold a decent thread. Although I used 3mm box section, there is a 25x5mm strip inside the box section to give depth for threading. I drilled the box section and plug-welded the strip in place using an improvised wedge arrangement to hold the strip against the inside of the box for welding. In retrospect, I should have used Eddycurrent's "glue it in place" technique. Once tapped and bolted, it ain't going anywhere...

Another variable to play with is epoxy strip width. I used draught excluder tape to make the dam and with the aid of hot-melt glue to plug any joins, it worked well. However, it leads to a narrower strip. Njhussey's technique with MDF walls glued to the sides of the box section clearly gives a wider strip. Personally, I believe that the narrower strip (as long as it is wide enough to support the rail) works fine and is perfectly adequate but it does make it essential to remove the meniscus which has to be done without damaging the surface that will take the rail. A wider strip means that the meniscus will not interfere with the rail mounting, and in any case there is more room to hack at the meniscus without damaging the surface. My way uses less epoxy, and I didn't have any problems taking a coarse file to the edges but that doesn't mean that it's the best way.

I'm sure that the epoxy technique would work well for a differently-shaped area. My experience suggests that the epoxy does not level itself absolutely but it might be better over a square area rather than a long thin area. Additionally, I wonder if large variations in depth affect the top surface so doing the filling in two passes - the first gives a "nearly there" surface and the second corrects any minor residual error - might help. Or be ready to shim out the odd thou or two at the end if you need a really flat surface, but the epoxy will still get you closer than any technique other than machining or grinding. Even a slab of Ecocast will need to be mounted carefully to avoid deflection when bolted down.

Quote Originally Posted by beone
NASA calls it "conformal shimming"
...and I shall be doing that with epoxy putty when it comes to mounting the gantry feet to the plates bolted to the tops of the X carriages!