[FR] or Mathematical Improvement to Quad Gantry Leveling

Features
1

Hi There,

in the file:

at line 88-91 the new Z-Height will be set doing the following lines:

88        z_height[0] = self.plot(slope_y_s01, self.gantry_corners[0][1])
89        z_height[1] = self.plot(slope_y_s01, self.gantry_corners[1][1])
90        z_height[2] = self.plot(slope_y_s23, self.gantry_corners[1][1])
91        z_height[3] = self.plot(slope_y_s23, self.gantry_corners[0][1])

works so far. But: Usually, if you do something like this (e.g. a Newton Raphson or similar), you add a little damping to prevent the system from swinging. My proposal is, to add something like

 z_height[0] = 0.9*self.plot(slope_y_s01, self.gantry_corners[0][1])

so 0.9 is just an example but might be a good value (0.7-0.9), showing the damping factor. The result should be, that there is little overswing and the speed of convergence should be slightly faster and the system should be more rigid agains false values for the Z-distances. At the end it is just good mathematical behaviour like adding washers to screws… :slight_smile: Or maybe not?

Onother guy asked for QGL-ZTilt improvement but I do not really understand, what is intension is in detail. But looks like the behaviour above.

1 Like