Edward - your very first adjustment/measurement didn't do what you thought it did. Putting a dial gauge in a collet and winding the table back and forth only tells you that the table is flat. You would get the same result with the column at 45deg, say! The only useful measurement is the turn-round/tramming test where you keep the table still and turn the spindle. However, this only tells you that the spindle is square to the table. The column could be tilted, and the head off-square to the column by an equal and opposite angle, leaving the spindle square to the table and you would not be able to detect this. The test to detect this is to tram the spindle, probably by shimming the column base as you have done and using the spindle turn-round measurements, then with your test bar in a collet and without turning the spindle, clamp dial gauge to table and then run head up and down column, note any variation, then run spindle up and down in head (drilling movement). If these match, head is square to column. If not, then there's no simple fix but at least you can double-check manufacturing accuracy.