Sorry I am not native English speaker, therefore I can be easily misunderstood.

X rails: pre-machining rails do not count due:
- rails must be parallel - your design is frame from 4 profiles. Forget on straightness (you are assuming this will help with using expensive profiles.) think about other two elements which are connecting both profiles.
You have 2 plates - potentially nonparallel, two profiles Mitsumi + two profiles catted and machined by... as I said check the stiffens of the bearing, lets assume the total error of non-parallelism is 100 microns, and that half of this error must be compensated by bearings, the other half by brackets. This means 25 microns per bearing - now check the stiffness and you will see the load which is in the "system". If you wish to remove error, you need to do machining in the assembled condition...

Check the I (Area moment of Inertia) of profile and plate, most likely 25 mm thick plate is stiffer than 50 mm profile or it is close to it. And whatever will you do: If the profile and plate are not planar, when they will be assembled the result will be non-straightness of the rail "bed" surface.

Y gantry: problem is torsional load not the bending - check the I and calculate displacement applying 100 N force at the end of the tool
The second problem is: Y gantry length must meet X rail width - within micron range otherwise you are facing with additional load to the X bearings... Not mentioned that both side surfaces must be parallel, - if not additional load on X bearings again

Regarding the idea of setting the true angles with bolt - by definition this means you are incorporating elasticity in relative stiff system. Means also you have nice oscillation.

You are right system is relatively rigid. I am escalating problems caused G&T and consequences of those.

regards

regards


,