Page MenuHomeAleph Objects Inc

X-Axis and Y-Axis not level
Open, Needs TriagePublic

Description

I took several measurements from various places on multiple mini 2s that were experiencing the inconsistent skirt issue as seen in T2220 and several mini 2s that had passed through calibration.

Probe1Probe2Probe3Probe4XTop1XTop2XTop3XTop4XBot1XBot2XBot3XBot4
95.895.2794.3395.5 170.06170.36170.8171.1
95.4295.9694.4595.15170.15170.25170.57170.7
95.1795.8295.5595.5170.1170.18170.05170.27127.75127.97128.16128.34
95.9896.2395.6396.15170.48170.49170.33170.5127.96128.03128.23128.42
95.5595.6494.7396.09169.91170.35170.47170.77
96.0496.6996.2395.84170.53170.5170.29170.35128.07128.11128.51128.59
95.6595.895.295.55170.7170.73170.26170.52128.27128.32128.42128.5
96.0795.9895.1496.26170.14170.24170.29170.53127.81128128.33128.62
95.4795.7694.8295.48170.32170.34170.36170.69127.99128.21128.56128.83
95.5695.7194.9195.7170.01170.13170.12170.56127.64127.9128.3128.43

I measured each of the height of each of the probe disks from a granite block in the order that the machine probes in (BL -> FL -> FR -> BR). I averaged out each side and compared the difference from left to right.

Y Axis Left to Right Diff
0.62
0.89
-0.03
0.22
0.19
0.19
0.33
0.35
0.33
0.47
0.33

We see can see here that the left side of the bed is slightly higher than the right.

Following this we gathered measurements from the granite block to the top of the x-axis rods, taking 4 different points as we went from left to right. Then we averaged the left and rights side and compared the left side to the right side.

X Top Dif
-0.74
-0.44
-0.02
0.07
-0.49
0.19
0.32
-0.22
-0.20
-0.27

The final thing that we looked at is the difference in spacing between the y bed and x-axis and then compared the left and right sides.

Left-Right Comparison
-1.36
-1.32
0.01
-0.15
-0.68
-0.14
-0.03
-0.55
-0.66
-0.60

The first two sets of data are confirmed to be having an issue with inconsistent skirts and we saw that the nozzle would not even touch the bed when leveled to the very bottom of the axis. The last five sets of data are from printers that have already passed through calibrations.

Proposed Solution
Raise the height of the y-axis (by adjusting the y motor mount and y idler mount) by approximately 0.3mm-1.5mm to bring the y-axis to a point where the nozzle can hit the bed and allow the auto calibration to have more affect.

Proposed Test
@logan Suggested using the shims from the foxglove days to boost up the y-axis so that we can test the printers with this current issue.

Data:

Event Timeline

oliver created this task.Fri, Aug 30, 11:11 AM
oliver updated the task description. (Show Details)Fri, Aug 30, 11:18 AM
oliver updated the task description. (Show Details)Fri, Aug 30, 11:38 AM
oliver added a project: kangaroo-paw.
oliver added subscribers: logan, cmerot, DaniAO and 5 others.
oliver updated the task description. (Show Details)Fri, Aug 30, 11:57 AM

Update on the first test: After adjusting the z offset I was able to get a passing skirt (approx. 0.35mm to 0.49mm) on a machine that could not get a passing skirt before.

Since some nozzles are unable to touch the bed in some cases, once we confirm this on more machines, we should get a change order in to adjust the y ends to add ~1mm to the mounting face. This will shift the y axis assembly up 1mm and should give us the additional room for the bed leveling compensation to do its job.
The Z home position of the nozzle is currently at 183mm above the glass, so a 1mm change will allow us to still print 180mm tall and be able to lift and move the hotend away from a full z build height print

zachah added a comment.Tue, Sep 3, 5:26 AM

Update on this from my testing as well. On Friday I was also able to get three additional red tag printers to have passing skirts with the shims. (0.380-0.498) (0.375-0.446) (0.349-0.434)

oliver added a comment.Tue, Sep 3, 7:31 AM

@zachah It seems with the jig we are also able to get the skirt within the old spec as well.

I had one more printer on friday that was 0.35mm to 0.49mm.

logan moved this task from Backlog to Needs Review on the kangaroo-paw board.
logan assigned this task to Steven.

This branch (T9909) contains the model files for the new Bed Corners v2.2.4 (PP-GP0477)

I cherry-picked the commit with the gcode from T9997 so model files and gcode for PP-GP0477 resides on T9909 branch of kpaw. Ready to merge.

Steven reassigned this task from Steven to DaniAO.

This is now merged into master for kangaroo paw so I've removed the tag. @DaniAO assigning this over to you for the hibiscus change order.