Well, I did find a slight issue with the machine over the weekend. I built a dual dial indicator tramming aid that fits in the quill. It is one of those where you zero each dial on a single piece such as a gauge block, and then bring the apparatus in contact with the table to observe spindle tilt. After using it to check the table, I was very surprised at the variation of readings over the table. In some areas of the table, there was as much as .009" variation in readings over the 5.500" between the two indicators.
Let me start with my process. I initially checked in the middle of the table, on the front-most of the two wide slabs on the table. (my table has a narrow slab on the front and rear with two wider slabs in the middle, all of which are divided by 3 t-slots, please forgive my terminology here). Then I proceeded to move the table throughout it's range of motion and check the readings in various locations. I took pictures of each position and will upload them shortly. This variation really bothered me, so I immediately tore down the entire machine to inspect everything I could.
I don't have a surface plate so I can't really check for relative flatness. About all I can check is relative thickness. I was impressed that my table was within .0003" on thickness on the ends, with only a slightly (.0015") narrower measurement in the approximate middle of the table. I might add that the front corners measured approx. .0008" thicker than the rear corners.
Next, I checked my saddle. What I found is that the front right corner was approx. .004" higher than the rest of the corners. I was measuring the distance from the top of the saddle (where the table would slide across) down to the ways on the base. So, I removed the saddle and put a micrometer on each corner, and found that all 4 corners were within approx .0003" of each other in relative thickness. So I stoned the bottom of the saddle as there were some aggressive scraping marks there which I assumed was for oil, and placed the saddle back on the base ways. Three of the corners sit flat, but when pressure is applied to the front right corner, I can fit a .006" feeler gauge under with essentially no rubbing. This makes me think that my saddle is slightly twisted.
So, then I reassembled everything, put my vise back on, and squared it up. Then I put a piece of 12" long bar stock in the vise and made a relatively thin cut on each side of the bar. I then measured the bar overall and what I found was surprising. I found that approx. half of the bar had exactly the same thickness as measured in increments of .5" from end to end. Then, approx. half way across, the thickness starts to increase to a maximum of .0012" thicker than the other side.
I believe this is due to the weight of the table shifting from side to side over the range of motion, causing the saddle to flex downward or upward. I also am led to believe that this may explain the variation of the tramming tool readings over the table range of motion.
What do you guys think?