This morning I changed out the test setup for a plate and section of bar that is actually bolted to the slide table instead of c-clamped to the plate. The problem must have been flex in the c-clamps because now I am measuring a deflection of less than 0.00025" with 15.24 pounds of force applied at 10 inches from the centerline of the slide. This translates into a stiffness of 60,000 pounds/inch which is well within the range for acceptable milling of aluminum. If all three axis had worst case 0.00025" deflection at 10 inches from the slide then the stiffness would still be in the 20,000 pounds/inch range (not sure exactly how they might add up but seems a reasonable guess). This is still acceptable for aluminum, especially with a small high speed cutter as the tables can move at many hundreds of inches per minute.
So having determined that the slides are decent for the amount of work (read zero) involved in setting up motors, ballscrews, linear ways and limit switches, I think the plan of attack is to come up with a configuration that gives the most rigidity for the largest travel in all three axis while keeping the overall weight down. I still think the classic c-frame xyz mill is probably the way to go. I am looking around at the few build threads that used linear bearings in a c-frame construction and note that they usually made a box or at least u shaped structure for the column. I have a big plate (36" x 24") of 6061 T651 1" thick aluminum that I would like to use for the base and column. I wouldn't have a problem with filling the column with epoxy granite if vibration dampening was still needed.
I am going to modify my drawing to extend the column higher to encompass 10 of the mounting bolts on the linear slide back, and instead of the forward triangle gussets, I am going to use 1 inch thick side plates to form a U-channel column supported all the way up. It is either this method or try a weldment of thick walled rectangular steel tube, but talk about heavy!