I've got my WorkBee 1515 assembled and running but I'm not using it yet. I'll cut mostly acrylic with it, and some aluminum. For both, the minimum goal is to match the performance of the Shapeoko XXL I'm replacing. The aluminum is the biggest concern. I read about the flex/twist inherent to the 1500mm gantry, and sure enough there's quite a bit of twist in my stock configuration, especially in the middle of the beam. I've also researched ways to address this, and what I think I'd like to do is double up the C-beam (back to back, bonded) and add linear rails to the X and Z axes. I'll incorporate a couple of other changes as well but I'd appreciate some feedback before I settle on anything. How much better can I expect this double-beam configuration to be? Will it reduce twisting by half, or more? I placed my rails on the top & bottom of the c-beam instead of on the face - my thought is that it'll reduce the cantilever of the router off the beam & & help a bit. Is there any disadvantage to doing it this way? The start of my design is below. None of the Z axis stuff is there yet. Thanks!
First of all, you're asking stuff which deals with more than basic structural calculations, and (for most) requires using a FEA software to calculate. Now, skipping accurate calculations and such. I'd say you gain negilible torsional strength from such a combination of c-beams, seeing how all the action happens on one of them, and more importantly, it's nigh on impossible to make the bond between profiles "perfect". I'd say a steel backplate or filling the profile would both give you more of a benefit (there's discussion on the subject of filling the beam on this forum). Best solution strength-wise would of course be to replace the c-beams with steel and/or concrete. Another thing I'd like to address is your choice of mounting the rails. Mounting them like that is prone to problems with the rails being parallel, as you need many surfaces and angles to be accurate. The original design is pretty close to the 'okos in rigidity, though I don't think anyone has made tests or calculations on the matter. But for your intended use, you should be ok with the original. Remember to use a light spindle to mitigate the flex.
I read the thread on filling the beam & within it is talk of doubling up the beams & the improvement from it. It's discussed in some other threads I've found as well. I've seen enough here that I think it's worth doing... I just don't know how much of a benefit I'll get from it yet. I mocked up something rough to see if it would have any noticeable effect: I moved the z carriage all the way to the left, clamped a bar vertically in the center of the beam, then I pulled on it. Then I clamped a piece of 2" x 4" x 47" long aluminum tube horizontally onto the exposed face of my c-beam (a 'pretend' double beam). I put the vertical bar back and pulled on it. Even without being connected at the ends, and with nothing more than four clamps holding it, there was noticeably less twist. If my janky test made a difference, I feel pretty optimistic about the fixed-end, fully bonded double beam. I'd still feel better if somebody who's done this chimed in & could say how it helped, but regardless, I'm doing it. Someday when it's done and in use I'll share it here.
