Quote Originally Posted by Jonathan View Post
That concept is excellent since it will have good stiffness for forces parallel to X and excellent torsional stiffness.

In case you've not already spotted it:
http://www.mech.utah.edu/~bamberg/re...e%20Design.pdf
Yup seen that a while ago, it is probably that which stuck in mind when I had the idea.

I'd probably stick to using one cross section throughout, most likely box section.
I had thought about that but I was getting concerned with the weight, the design as it is will probably go 80kgs with spindle&Z etc on it. Using all big box section would probably add another 20kgs to that. I was starting to chicken out running two screws to move a 100kg plus gantry on a single 3NM motor. :(

Am I misunderstanding how trusses work? I thought that the struts functioned by effectively transfering force to the main beams by being in a combination of compressed/tensioned members when it is loaded.

Do the struts really need the mass of the main beams or were you suggesting that for ease of construction?

Your main concern will be making the surfaces on which the Y-axis rails mount sufficiently accurate, i.e. parallel. It's much easier if you mount both the rails in the same plane (i.e. on the 'front') since you can then just use epoxy granite if necessary to obtain the required flat surface. Using that orientation does however increase the overhang of Z, but I think that's not a big problem given how strong the gantry will be.
It isn't shown on the design but the beams that the SBR is going on I was planning to attach using a bracket and bolts. I was hoping an eccentric bolt would be sufficent to give me fine adjustment and get the two pieces parallel when fitted. The steel angle seemed like a good choice as I could get the fine adjustment finished and then epoxy the steel angle in place knowing that it wouldn't move out of alignment. The angle also allows me just enough room to run the ballscrew inside the gantry.

Any discussion of structures failing is hardly relevant, since in a CNC machine we are aiming to obtain a certain level of stiffness, not just aiming to ensure the material does not fail. That is why to the beginner the designs can look over engineered, since the size of material used is comparable to much larger structures. The critical difference is in civil engineering the allowable deflection is substantially greater - nobody cares if a stage truss deflects by a couple of millimetres, but if your gantry deflects by just 0.1mm you could be in trouble.