calibration issue with new drivers

I recently replaced my drivers because one gave up the ghost, with DM 452A units, on Larken's recommendation. It was easy once I got the wiring straight. Thank goodness for optocouplers.

I have 3 distinct and repeatable errors happening, as I tried this run 3 times using fresh G-code from Vectric V-Carve (thus the sets of 3 photos).
1 - The photos show an unevenness between the left and right edges, and 2 is a strange hump consistently showing up at the base of the vertex. Finally, the Z axis is not matching between tools. I set the top surface to Zero after each tool change and that has always ensured consistency. However, now suddenly, I am getting a disagreement of about 0.005" between the 90 degree V bit and the .0125 endmill. Finally 3 is that it looks like the 90 bit is starting off about 0.005" lower than the top of the workpiece, as the 3 stars across the middle are not crisp edged/pointed like they normally are.

I have not had this sort of error before. I am baffled and frustrated and clueless on how to solve it. I have cut a few other items too, testing, and had some odd/puzzling actions occur, again focused on scaling.

rocko,

The errors you are referencing do not appear to be a calibration problem. I would suggest that you machine a 3" square in a piece of waste material and then measure for error in the width and height of the square. Testing calibration on an intricate pattern is not the way to get a definitive result.

RFB

I concur that using an intricate piece is not the way to calibrate as well, but since all has been well up to now, this is how it came to be known. And I just kept digging and digging at it.
I made a new set of code using the origin at the center of the workpiece and got a bit of a different result. The hump issue is now being considered a mechanical cause, maybe indicating I have a bad bearing under the table in the X axis. The Z is still funky. As shown in the new attached file (4rth cut of the same item!) I found that I had to lower my Z 0.05" to get the broadest stripe to match bottom with the 90 bit that bevels its edge. But that might be because the 90 bit was 0.05 too low. The 3 stars now have a definite variation in depth.
Overall, I am convinced this has nothing to do with the new drivers. They just don't matter. They simply do the amplifying of the signals for the steppers, right?
This must be a controller issue. I recently bought a new controller chip from Larken as he recommended due to a bug identified that was causing a misregistration in the Y by about 0.1".
I am going to make some more scrap cuts as suggested and try to find what the pattern is.

Thanks for the suggestion!
Rocko