(the dial indicator reads in metric, with 1um resolution)
radially the spindle chuck-flange shows ca 14-15 um of runout.
Axially there is more wobble, ca 20um.
The Z-rails are very accurately aligned in the X-direction, I measured ca 6um error.
Measuring in the Y-direction, there seems to be a 40-50um jump around 10cm from the chuck. I'm not sure if that is important, since the rail-blocks don't make contact with the rail top surface. I think I need a way of measuring this again with the machine assembled.
When measuring the rails, I would either put the indicator tip in the ball groove, or on the bearing block. The ball grooves might be machined in a separate operation from the top and side surfaces, so measuring those surfaces might not be accurate.