Just to clarify, what I found to begin with was a spindle running hot, I have since run it in, and still get 160+degrees. I also found what I thought to be play in the spindle, but it turns out is coming from somewhere else. So I am just adding my investigations to this thread so that more experienced people can point out where I may be incorrect or could improve my practices.
I wasn't sure how to determine play/deflection in a manner that would simulate normal usage so I ran a quick experiment. I have a spring that has 10lb/in of stretch ( I picked it up for a different project ) and connected one end to the top of the column and used my steel straight edge to measure the amount of stretch. I then mounted my vise to the table and clamped the indicator snug in the jaws and tested in both x and y directions against the z way ( the big block that the dovetail mounts to ). The indicator is level and is sitting 1.75" off the table.
It only takes a about 5lbs of force to start causing movement in the y direction. The amount of deflection increased to about .001" at about 15lbs and did not increase above that even at 50lbs.
The x direction saw only .0002 of movement even at 50lbs.
Being that the force was applied to the "lever" at 15-3/8" from the fulcrum that would be equivalent to x amount of force at 3-3/4" from the fulcrum. I don't remember how to solve for x. But if I recall correctly, a general rule is that the amount of force applied increases by a factor of 2 for every inch? I will have to look that up when I get home. Either way, with 269oz/in steppers and feedrates limited to 18ipm max I don't think the column itself is the source of play, so I will ry to measure the play between the z table and column next.
I also went ahead and removed the spindle from the headstock to see what condition it was in. It had to be getting -extremely- hot on the inside bearing race/spindle shaft surface because the oil residue has all been scorched dark brown/black and is flaking off. It only has about 5 hours of run in on it, 99% of which was at the lowest speed.
The assembly is a 6203R-oo at the bottom with a spacer that goes up to 2 face-to-face bellville washers that press against the outside race of a 6203z. I assume the washers allow the tension to be maintained regardless of heat lengthening the spindle. I found that when I removed the spindle nut completely, the bearings remained in place, so the spindle fit is really tight. I also found that the spindle nut is not mating flat to the bearing race, one side meets the bearing about .015 before the other side does.
I think 6203 are just deep grooves. That is fine for the top, and I am pretty sure it is ok for the bottom at the low loads this little mill will see.