I think Peter said in another thread that it's better to spend 500 hours in CAD than to spend that time in the workshop and thousands of dollars to figure out that something is wrong with the design.
Since my decision to mount the rails on the side I've investigated the accuracy of the machine that will make those sides parallel. It turns out the machine isn't that accurate; across the whole length they can't guarantee parallelism of more than +- .3mm. If my gantry was made from aluminium that could probably work, but with a rigid steel gantry I think that error would put a LOT of pressure on the bearings. I could shim the rails to make them parallel but what was the point of machining them?!
So I've came up with another design. I'll still use their machine (since they'll do it for cheap for me) to make my construction flat for mounting linear rails. I think this is an interesting design. It prevents a big problem most routers have and that's sawdust and chips falling on the rails and ballscrews. Here they are completely covered.
But I have a dilemma. I hate that I have this cantilevered portion of the table top. I intend to use 1" - 1,1/2" aluminium plate here (20-30mm). Is this going to make some problems that I'm not aware of? Only during drilling operation you have a straight down force on the table...
I could raise the top an inch more and bolt some angle iron underside the table to give it some support against flex and vibration...
What are your thoughts?