[maxfuchs]

At one point (I can't exactly figure what changed) my machine started leaving a wavy trace along X axis. The distance between the peaks is equal to the ballscrew pitch. I originally used 1610 screw, now it's 1605 and the distance between peaks changed accordingly:
https://imgur.com/a/eOPONXE

I detached the ballscrew completley and moved the X axis by hand to mill the same pattern and the issue is gone:
https://imgur.com/a/oyjmA5p

The motion system on my machine looks as follows:
https://imgur.com/a/VAtPsx1

I spent several days trying to fix this issue to no avail. I replaced the X axis linear rails, upgraded the coupling to a diaphragm one, replaced entire ballscrew with matched nut (it has C7 accuracy class, so not very high). All parts are original HiWin and TBI. I checked the alignment of the ballscrew with the rails and it's +/-0.05mm at worst. The issue still persists. The tensioning nut is as tight as it gets, but it doesn't help either. It seems that the ball screw must be bent and the issue worsened over time, but how the heck do I prevent this? The misalignment doesn't seem horrible, because I can easily turn the nut by hand and move the axis.

I'd be hugely thankful if someone could help - this issue has exhausted me hugely and I hit a dead end.

[joeavaerage]

Hi,
even if the ballscrew were a little bent, the axis itself should stay absolutely straight as it is retained by the cars/linear rails.

Are the cars preloaded? Is there any side to side play in the axis?

What grade are the screws? Most commonly (cheap Chinese) they are C7 which is 50um per 300mm but the cyclic deviation is uncontrolled. I have seen measurements where C7 screws exhibited 35um
deviation in one revolution. Could it be cyclic deviation?

C5, the entry level ground screws are very much ( as much as ten times) more expensive but the cyclic deviation is controlled and is required to be 8um or less per 360 degrees.

Craig

[cncmakers001]

At one point (I can't exactly figure what changed) my machine started leaving a wavy trace along X axis. The distance between the peaks is equal to the ballscrew pitch. I originally used 1610 screw, now it's 1605 and the distance between peaks changed accordingly:
https://imgur.com/a/eOPONXE

I detached the ballscrew completley and moved the X axis by hand to mill the same pattern and the issue is gone:
https://imgur.com/a/oyjmA5p

The motion system on my machine looks as follows:
https://imgur.com/a/VAtPsx1

I spent several days trying to fix this issue to no avail. I replaced the X axis linear rails, upgraded the coupling to a diaphragm one, replaced entire ballscrew with matched nut (it has C7 accuracy class, so not very high). All parts are original HiWin and TBI. I checked the alignment of the ballscrew with the rails and it's +/-0.05mm at worst. The issue still persists. The tensioning nut is as tight as it gets, but it doesn't help either. It seems that the ball screw must be bent and the issue worsened over time, but how the heck do I prevent this? The misalignment doesn't seem horrible, because I can easily turn the nut by hand and move the axis.

I'd be hugely thankful if someone could help - this issue has exhausted me hugely and I hit a dead end.

As you said originally used 1610 screw, now it's 1605, this could be the reason resulting in intermediate flexural deformation.