Hello.
I checked the settings again and noticed a different movement along the Y axis. On the Y axis, I periodically have questions.
Recorded video for X and Y axis.
On the X-axis everything is in order, in my opinion.
A different movement occurs along the Y axis, if make one movement or several in a row.
The movement decreases with each movement as move down the axis Y. And increases if move up along the Y-axis. That is, when move back, in the same step, we come to the starting point.
Y-axis movement is correct only in the first step. That is, if moves only once. For example, immediately by 170 mm.
X axis OK

Normal? That is a question.
I will also join.
Machine VP2500HP-CL32.
SW V2.057
The same behavior, only the axes are reversed.
Y axis - O.K.
X axis - small steps with decreasing offset.

Tested SW
V2.057 from date 10.12.2020
V2.054 from date 6.7.2020
V2.048 from date 20.12.2019
Identical behavior.

Y axis - O.K.
X axis - small steps with decreasing offset.
The offset is created by short movements in both directions.

Perhaps the settings in the servo driver are wrong?

One thought occurs to me.
I had such a mistake with the stepper motor controller when I turned on the Enable Drive late. The first 2 pulses were lost to me.
The long movement of O.K. Short movements accumulated offset.
Very similar.
But I don't want to crawl under machine an oscilloscope.

So it won't be an Enable signal. It is not used at all by me.
The file "VP_Technical_manual_V2control.pdf" >>> pin 1 Enable / only for servo drive, but pin 1 is not even fitted.

I checked the wire connections - I didn't see any problems.
Measured the resistance between phases V-W, V-U, W-U. The resistance is about the same everywhere 2.55 Ohm.

Conducted an experiment. With the machine turned on, I brought the heads to coordinates 100 x 100 in X and Y. On the control board, I swapped the X and Y connectors that control the drivers of the servo motors. And again I moved along the X and Y axes using the "Set Axis Calibration Value" dialog. Now the problem has moved to the X-axis, and the Y-axis is okay. But, I noticed that the error along the X axis is less than that along the Y axis.
Change conn X and Y.png - Bild entfernt (keine Rechte)

From connecting the controller board to the servo motor driver, it turns out that the controller board does not know the real coordinates of the movement, but calculates them from the control pulses to the servo motor driver. Or am I wrong?

I also noticed a different difference in the absolute coordinates for the X and Y axes, if we go from the origin (mechanical zero) to 10 mm along the X and Y axes.On the X axis, the difference is 0.0055 mm, and on the Y axis, the difference is 0.0247 mm. Can you explain why this is so?
X Y 10 mm.jpg - Bild entfernt (keine Rechte)

Also checked the wire connections - no problems.
Inverter connector-soldered, all O.K.
Here again the insidiousness.
According to the Leetro manual the X2 50-pin, connector is actually 26-pin.
Otherwise, the degree of availability for repairers is 9.5 out of 10.
Connectors on the Mainboard again connected differently.
The documentation (VP_Technical_manual_V2control.pdf) is an old version of the board.
Now: Pin 1-CLK; Pin 2-Power; Pin 3-Direct; Pin 4-? (not fitted); Pin 5-GND (not fitted)
Measured the resistance between phases V-W, V-U, W-U. The resistance is about the same everywhere 2.80 Ohm X-axis, 2.57ohm Y-axis.

New knowledge.
1) I'm correcting pin numbers if 1 is a square pad:
Pin 5-CLK; Pin 4-Power; Pin 3-Direct; Pin 2-? (not fitted); Pin 1-GND (not fitted)
2) I brought an oscilloscope and: switching the direction is really earlier than the pulses themselves Step.
So the shorter distance comes from the control board.
This error occurs ONLY in the "Set Afix Calibration" dialog and during short shifts. When scrolling with the buttons on the panel or the cursors in the SW, the shifts are O.K.

And during the execution of the task of mounting components?

This menu is used to calibrate the machine. If you can't rely on this calibration, then why rely on it.
Personally, I think there is a mistake in long moves as well. It is only small in relation to the length of the shift.
That would have to take Johansson gauges and dial indicator. And I don't have Johanssons.