Height Map issues

sudodelta

I have been struggling with the above height map ever since installing a smart effector. I even took apart my whole frame and reassembled it, only to have the exact same issue. The high points are near the pillars and the low points are in-between the pillars. I am using the PCB magball mounts on the carriages.

I have check the calibration of my frame and confirmed nothing is loose anywhere. I am currently on firmware 3.3.

Below are some pictures of my printer as well as config files. Any help would be appreciated.

bed.g config-override.g config.g homedelta.g

T3P3Tony

@sudodelta what happens when you run G32 a number of times. you should see the error reducing to a "steady state".

sudodelta

@t3p3tony The deviation on the G32 never gets below .130 (after running it 8 times) and the height map issue persists. I am beginning to think my endstops are to far out of sync, M666 X-4.079 Y0.542 Z3.536 A0.00 B0.00, i thought the g32 would have compensated for it but maybe it's too severe.

jay_s_uk

@sudodelta 0.13 is very high for a delta with magball arms (even if they are home made) and a smart effector.
do you have much play in the carriages?

T3P3Tony

@sudodelta G32 does compensate for the endstops as part of the 6 factor calibration. As @jay_s_uk says that's very high deviation, try bringing in the outer probe points in slightly as well as checking for play and ensuring the probing sequence does not cause anything to hit onto bed clips etc

sudodelta

@t3p3tony Apologies for the delay, had a baby last week
I can't find any play in anything and no clips are being contacted. If I bring in the outer probe points the deviation does improve though.

T3P3Tony

@sudodelta congratulations!

So what deviation are you getting when repeatably probing with G32 now?

sudodelta

@t3p3tony If I reduce my Delta calibration radius from 110 to 95 I will get a deviation of .03 but my height map will still show those highpoints that it doesn't seem to be able to compensate for when printing.

T3P3Tony

@sudodelta 0.03 deviation is pretty good. What does the heightmap look like after you have got down to that deviation and then saved the new M665 and M666 values (either by sending M500 and using config-override.g, or by copy the values in config.g

sudodelta

@t3p3tony So I noticed that with the pcb mounts for the mag balls when i was close to the tower it was too close for the rod ends and they were lifting away from the mag balls. I was able to replace the carriages with ones that put the magballs at a downward angle and now my G32 returns a deviation of .015 all the way to the edge of the bed. I am still working on getting a good height map since so much changed with the carriages.