Aboutthat dead space comment that resulted in confusion.
You mentioned that you used two beams in your gantry design so that you would have room for a leadscrew. What I was trying to say is that such channels are nit always required and may not be desirable. The reason being is that your linear rails and the Bearing cars that ride on them, form a channel between the often good enough for a leadscrew. The specifics of the parts used obviously comes into play but by leveraging this “dead space” you can reduce build complexity.
The other thing here is that mechanical you want to avoid a heavily cantilevered leadscrew nut bracket. The bracket then becomes a lever arm that can flex or put an undesirable toque on the saddle.
By the way I’m not saying that you should never build in such a gap as a combination of factors can make it to be the right solution. Rather what I’m saying is that I would avoid designing in such a feature first and instead try to mount everything on one beam. If your components don’t allow for a single beam solution then a dual beam approach might be the right idea. The complexity I’m talking about here is getting the two beams assembled and in the same plane. It can be done but the lure is extra work involved.
As for the extrusion comment:
A lot of people assume that the T-slotted extrusions are perfectly flat out of the box. They aren’t and like steel beams may require machine shop work to flatten. Generally they aren’t bad for wood working but may not be good enough for machining aluminum. It really comes down to your expectations but generally people need better results in Aluminum than would be acceptable in wood.
So the extrusion comment was more about nit assuming that they are good enough flatness wise the meet your goals whatever they may be. There is a whole discussion about the precision one can achieve in simply mounting linear rails to t-sootted extrusions.