May I encourage you to record the results of your testing method and results both before and after - should you choose to replace your ball nut with another stock replacement - as it would be interesting for us to compare the data.
I became interested in your question, so I just did a quick set up - see photo - trying to isolate just the ‘play’ in the ball nut. When applying 50N of force to the Y axis block, there was about 30 microns of movement.
Earlier, to satisfy my own curiosity (and to gather empirical data), I investigated the amount of deflection I could achieve at 3 different locations on my Woodworker X50. Without getting in to too much detail, When looking at my scatter plots I found that (in my testing up to 50N push/or pull):
- the greatest deflection, not surprisingly was at the endmill (in my case I have ISO20 ATC toolholders, measured at the collet/endmill) with a range of 110-210 microns at 50N
- the deflection varied in the X and Y axis direction
- adding the 3rd rail reduced deflection in all tests and directions of applied force
- the force/deflection curves remained linear (for my test force range)
I have been mostly interested in machining aluminum, and have had great results (for me) using effective and conservative cut parameters and toolpaths. After collecting my deflection data, I used a few programs I collected to look at (theoretical) cut forces for the spindle I have and the chip loads I usually aim for. From my machine and machining data, it seems to confirm that I am able to keep deflection in a range where it has limited impact on the quality of my cuts.
