Question Gremlins on my X-axis?

By using all my magic powers, spells, tips & tricks and shaman background I have reached the inconceivable: +/- 0.02 mm (no typo) tolerance in all my cuts.

Here the bad news:
when I make cuts away from the center of the X-axis the precision progressively drops down to +/- 0.1 mm (10 cents).
The problem affect only the X-axis, the Y-axis keeps his precision unaltered.

What's going on?
It 's a problem related to the wearing of the leadscrew? If so, it should not be evident in the most used zone (the central part of the long screw) instead of the marginal ones?

I'm scratching my head, but I cannot find any simple explaination... any idea? Trolls? Gremlins?


Before you ask: yes, all the bronze spindle-nuts ("spindelmutter") are new.

progressively drops?

Hi julius

Could a very slightly bent leadscrew cause this - might be worth checking with a dial gauge?

Peter

Hello Rory,
yes (actually "drop"). Precision start from +/-0,03 in the center of the X-axis down to -0.1 mm at the left/right end.
To be more precise, the dimension goes always to be smaller of what described by the g-gode. In other words, every shape is 0.1 mm shorter on the X-dimension.
I've made dozens of test in every zone of the working table. As reported, the problem is independent from the Y-location of the milling bit.

I'm scratching my head, but I cannot imagine the reason...

peterg1000 wrote: Could a very slightly bent leadscrew cause this - might be worth checking with a dial gauge?

To check that I should unmount all the stuff... (oh, not that again! ).
Anyway, even if the leadscrew is bent, the effect on effective length of the traveled distance on the X-axis should be minimal, or not measurable at all.

Also, if the bending would be the real reason (i.e. the leadscrew describes an arc, not a straight line) I should observe some effects in the central zone too. If not, then the leadscrew should be bent at edegs only, and it would describe a very very strange "curve".

The nut seems to travel normally along the X-axis. No strange noises, even when moving at the edges.
The only explanation is that I'm loosing steps at the edges, even if didn't notice any particular stress of the machine....

Given a unitary distance on the X-axis, there is an easy way to count the real number of revolutions of the axis motor and compare it to what happens in the central zone?
Any idea?