Update.
I found some adjustment issues with my saddle's shimmed slide plates (say that 3 times fast!).? Their mounting screws had loosened up a bit over time, permitting the saddle to ride up a bit over the Vee as the saddle was twisted.? Tightening them up eliminated the indicator variation due to that, so I reinstalled my 3/4" test rod and performed RDM.? I saw about .0005" variation at the headstock end and zero to .004" as I rotated the spindle, so I believe that comes to misalignment a bit less than .002"? over a distance of about 10 inches.? The polished guide rod is very smooth so it's pretty good for this kind of test.? I didn't observe much evidence of bed twist as I moved the indicator along the bed.
I haven't performed RDM on the vertical axis yet.? I really need to switch over to using one of my DTI's for these measurements, now that I know the spindle is fairly well aligned.? I have a 15-0-15 and a 4-04 (.0005" and .0001") DTI I can use for this.
Visual examination of the saddle V revealed that about 25% of the V really is in contact with the bed V (based on the polish marks on the V).? Rather than scraping to improve the contact area, I'm thinking about using PSA-backed sanding paper strips attached to the bed to refine the saddle "V".? Pretty much as described by a document in the Files section.? The procedure doesn't alter the bed surface.? But as it stands the lathe does pretty well on steel so why fix something if it ain't broke??
