Lead screw/Guide rail install/pre-load

The recent thread on lead screw bearing designs now has me thinking
about how any preload on the bearing stackup should be approached.

My application is not fatigue driven. What determines an appropriate
torque for the retention nut on the lead screw, i.e. the preload on
the bearing stackup?

Ian Wright points out that, in conjunction with a FIXED or RIGID
double bearing at the non-motor end, a SIMPLE bearing design using a
single bearing at the motor end may allow for limited thermal
expansion along the lead. I don't think my application would benifit
from a pre-load in the lead itself, i.e. my app is slow-speed, low
loads.

The recent thread on lead screw bearing designs now has me thinking
about how any preload on the bearing stackup should be approached.

My application is not fatigue driven. What determines an appropriate
torque for the retention nut on the lead screw, i.e. the preload on
the bearing stackup?

Ian Wright points out that, in conjunction with a FIXED or RIGID
double bearing at the non-motor end, a SIMPLE bearing design using a
single bearing at the motor end may allow for limited thermal
expansion along the lead. I don't think my application would benefit
from a pre-load in the lead itself, i.e. my app is slow-speed, low
loads.

With this layout in mind, how should I do the 1" guide rails? I had
planned to drill all guide rail support holes in the support plates
at the same time, to ensure a common axis, and parallel rails. I
planned to drill interference fit support plate holes for the guide
rails. I'm hoping to use 1" drilled rails, and to simply screw
through the guide rail support block (on the motor end only) and into
the guide rail. I plan to leave the other end (away from the motor)
simply sitting in it's interference fit hole, and not bottomed out in
the hole, in order to allow for thermal expansion.

My thinking is that this arrangement for the lead screw and support
rails will provide adequate support while preventing thermal effects
from deforming the linear elements.

Does this sound reasonable? Any alternate plan suggestions would be
sincerely appreciated.

Best regards,

Ron