Graduated bed compensation

From RepRap
Revision as of 03:49, 4 July 2016 by Thinkyhead (talk | contribs) (??)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

"Graduated bed compensation," as the name implies, augments bed compensation by gradually reducing the amount of tilt or Z adjustment being applied as the print proceeds in the Z direction. This is better suited to Mesh Bed Leveling than Grid-Based Leveling because MBL has no effect on X or Y position, only adjusting Z to compensate for bed irregularities.

The theory behind graduated bed compensation is simple. For a very uneven surface, it makes no sense to transmit the bed irregularities to the entire object. Instead, why not gradually and imperceptibly reduce the amount of compensation being applied over, for example, 10mm of the print? Thus the remainder of the print can proceed with compensation effectively (or actually) disabled, reducing both the amount of irregularity transmitted to the object, and the amount of work done by the machine's processor.

There are tow proposed approaches to reducing the amount of compensation that needs to be applied throughout the print. Used in tandem, these approaches can eliminate the need to perform extra transforms or interpolations in the majority part of many print jobs.

The first approach is as described above, simply a gradual reduction in compensation, so quickly the machine returns to flat, uncompensated printing. (Testing bears out that this produces an obvious curvature in extreme situations, so don't do that!)

The second approach applies the unevenness in the bed to the amount of extrusion done in the first one or two layers. In this procedure the Z axis remains level, but more material is extruded in the "troughs" and "valleys" of the measured mesh, while less material is extruded at the "peaks." The result is similar to what you get if you let liquid settle anyplace on Earth. The bottom may be as uneven as you like, but the surface is nice and flat, and should be as smooth as glass.

Currently no firmware implements either of these approaches. But as these algorithms are developed and the idea catches on, we should see these concepts implemented more widely.

(No complaints. The base of your object was already warped!)