mesh compensation more than 441 points

JoergS5

I had an idea last night. I cannot implement it myself currently, but I want to share the idea:

there is currently a limit of 441 points for mesh compensation. How about

creating several mesh compensations at different areas and storing them as different .csv files
a program in DWC or SBC (DWC preferable, because it works with all Duets) loads all csv files and analyzes to take the regions with the biggest unevennesses
a new csv from the results with the 441 most relevant (uneven) points is created and stored
use those new csv for mesh compensation

o_lampe

@joergs5 I had a similar idea lately. I wanted to use an analog sensor and meander it across the bed without diving up/down. Whenever it reads a different height, the coordinates and height difference is stored.
In the end both our ideas would end in a file that requires to store coordinates to the height. That could increase the file size. ( or at least makes it unpredictable)
There were other proposals, like probing the bed only where the object is. That approach has a better chance to work on all boards. But it requires g-code analysis to find out the outlines of the print object. (skirt-coords?)

JoergS5

@o_lampe you're right, limiting it to the object will be probably more valuable.

There is the approach to be able to remove objects when printing multiple objects, this code should already know the print area. So Cura or similar program can tell RRF which regions to mesh measure.

My proposal may be most valuable with very big printers like house building printers.

JoergS5

@o_lampe said in mesh compensation more than 441 points:

analog sensor and meander it across the bed without diving up/down. Whenever it reads a different height, the coordinates and height difference is stored.

funny, I had a similar idea, taking a video of the nozzle while scanning the lines. The idea is to fasten up the mesh process.

o_lampe

@joergs5 Maybe a dynamic heightmap is the answer for big print areas? The heightmap contains a zillion probing points, but the firmware only loads the currently relevant part of it.

dc42

On the Duet 3 and 3 Mini we have additional RAM available, so we could increase the number of points allowed. What would be a reasonable size? 31 x 31 instead of the current 21 x 21 ?

JoergS5

@dc42 from my side I don't have the requirement, I had only the idea. I was referring to https://forum.duet3d.com/topic/18830/m557-number-of-probe-points and they will probably be interested in increase. But @OBum_2020 used a Duet 2, so this may not be possible.