I cannot really help you directly. I used to test some of my old cncs with a 45° linear interpolated movement. I would fasten down a straightedge to the table, at an exact 45° angle, and then write a single line interpolated movement to make the table move along this line. With a dial indicator in the spindle, and resting against the straightedge, any delay in one axis would result in a displacement of the indicator needle. If both axis were identically tuned, then the needle would not move at the start, during motion, or end of the test movement.
I wish I could remember what it was that I adjusted to fix this, but they were old analog drives without any sort of digital PID settings, so it may have been a combination of tweaking the gain or maybe the balance pot. Balance pot adjustment was actually an initial setup adjustment, where no command signal was sent to the motor, I simply adjusted this so the motor did not drift either direction when it was supposed to be at zero speed.
Perhaps this will create some new methods of testing your own setup.
When I see the terms "rolled ballscrew, no backlash" in one description, I see a contradiction I do not think that rolled ballscrews have enough precision to operate without some sort of floating backlash adjustment, because otherwise they would bind up here and there in the travel, or be loose in other places. So even though the illusion of no backlash may exist, this does not mean that high accuracy is ensured. You need high accuracy ballscrews to have a hope of doing high accuracy circular interpolation.
First you get good, then you get fast. Then grouchiness sets in.
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)