New Railcore II Bed Leveling Issue



  • I just received my preassembled RailcoreII from ProjectR3d. I have a bed leveling issue that I have been working through with them but thought someone here might have seen something similar.
    I'm running from the "factory" firmware 2.02, DWC 1.22.6, WiFi server 1.22.
    config.g
    0_1563653535172_8eaa7c7c-5981-4838-a293-59c1b0027c0d-config (1).g
    and bed.g
    0_1563653621180_d056d52d-fdab-4ca6-ba90-c961e5d3c125-bed (1).g
    bed (1).g

    Started out by doing a g28, 3 g32’s ( bed.g with 3 points, one at each z rail).
    This converges to .001 before and 0 after. So nothing weird. If the bed was flat this with 3 z motors should fully correct bed. So I print a 150mm circle. Toward the back if the bed, the print doesn’t adhere. I resand the wham bam pex, wash with soap the alcohol. Same thing. Raise bed temp to 70 then 75. Same thing. Rotate the pex flex sheet. Same thing. Rotate infill angle. Same thing. I put a borosilicate plate with pei on it. Reprint. Same thing.
    I pick 2 points in the middle of the bed, one near the front of the bed, the other near the back, but only 180mm apart. I rerun the g32’s. I then move the nozzle over one of the points. G0 Z0 then manually jog down until it grips a .02mm cig. paper. -.08 Then I do the same at the other point at the back of the bed where it hasn’t been adhering. I have to go down -.20 (it's dead on at the center of the bed). So it's clear why it's not adhering. They did test prints before it was shipped but lots of bouncing around between there and Seattle. Anyway with a .12mm difference no way will the print adhere at the back.
    I spent all day yesterday working with the ProjectR3d folks and they had me try a bunch of other things. They have been very helpful and patient.
    Anyway so is the bed twisted ( it's a mag bed and apparently the magnets create havoc with the bltouch so no good multipoint mesh)? I tested all this with cold bed. I had a piece of borosilicate glass I use on my ultibot and laid that on the 713maker mag bed. I ran all the same tests with virtually the same result. Doesn't this imply the mag bed is not twisted? The bed is lower in back than in front. I've tried using all 4 corners of the bed in a revised bed.g. Same result. It's almost as if the G32 is not seeing the same tilt or the mechanics of the printer are not functioning properly in responding to the correction.

    Can I command each of the 3 Z screw motors separately? I'd like to move the front screw away from the extruder by .06 and the back screw toward the extruder by .06.

    Any thoughts much appreciated.



  • can you post a picture of the mesh bed levelling result?



  • 0_1563687604436_1DD5E96F-060C-49A6-BFCD-1A845A9EA299.jpeg



  • that is not a flat surface. so you need mesh bed levelling to print on it.

    have you got the bltouch with the plastic pin?



  • Yes. It's a mic6 bed from 713maker. Its pretty flat. this is I think an issue with rails or the way the kinematics/mechanics are playing with the reprap algorithms.



  • Just a thought as I had a similar problem, Do you have G29 S1 in your slicer start gcode?


  • administrators

    My guess is that one or more of the points you are using for the G30 commands in bed.g are being affected by the magnets. Are the magnets in particular places? If so, can you adjust those point to avoid them? Although we suggest one probe pont close to each leadscrew, the exact positioning of the probe points doesn't matter because the levelling algorithm takes account of where the points are.



  • One way to test if the magnets are the problem is to change the probe type to P0 for manual probing where you are presented with a jogging window to manually lower the nozzle until it touches the bed. For the 3 points needed for the mechanical auto leveling it wouldn't be too tedious to test. That at least takes the probe out of the equation.

    Do you have a straight edge or a square that you could use to eye ball that dip to see if it's actually flat or not? 0.3mm isn't big enough to see just looking at the plate, but it's enough that there should be a visible gap.



  • So I took the flex sheet off and tested with a small piece of steel to find the magnets(they are covered on the underside by the heater). I moved the 3 bed.g points a little(projectr3d ships with a plastic probe bltouch are these still affected by bed magnets?). I’ll run more meshes but if magnets are really a problem then that’s fraught with error. Testing a print now.
    I decided the flattest thing I had was my large macbook pro so turned out the lights and tried it. I do get a small gap between front and middle of bed and a slightly larger one between middle and back of the bed. Perhaps that’s part of the issue but rail adjustment must have something to do with the way the map falls off linearly toward the back.



  • Nothing changed. Prints still don't adhere past middle of the bed. New g29 after 3 g28’s yielded.
    0_1563751490288_image.jpg



  • Do you have the mesh compensation active for the print? Is it trying to adjust for that dip or not?



  • Tried with and without. After just the g32’s no way. After a 16 point g29 it improves a bit but doesnt correct enough for the dip. When I used mesh with the 256 point I did last night, i got a 225mm circle to look great except for one small section at the far end of the drop off.



  • Here is the 4 point g29 they had set up for rail adjustment. It doesn't look bad. Obviously you miss the dip. But does this indicate rails ate reasonable?0_1563754349393_image.jpg



  • @dc42 Looking at your post above, you said that it doesnt really matter which 3 point are probed fir g32. I realize you said that because if the bed is flat, any 3 points will determine a plane. I then thought why not use that to disprove flatness. I have tried using sets of different points (avoiding the magnets) and can get very different g29 results after the g32’s. Is that not proof that the bed is not flat?



  • G31 X2 Y42

    can you double check that those values are correct?



  • @Veti those are the values I have been using. Basically I am using the config.g, bed.g and firmware I received from Projectr3d.



  • Finally got a quality machinist’s large square. The bed has those dips in it. Software rules!


  • administrators

    @4lathe said in New Railcore II Bed Leveling Issue:

    @dc42 Looking at your post above, you said that it doesnt really matter which 3 point are probed fir g32. I realize you said that because if the bed is flat, any 3 points will determine a plane. I then thought why not use that to disprove flatness. I have tried using sets of different points (avoiding the magnets) and can get very different g29 results after the g32’s. Is that not proof that the bed is not flat?

    Yes, or that the trigger height of the sensor varies with where on where on the bed you probe.



  • It turns out they had some bad beds. Sending a new one.



  • @4lathe Did the new bed fix the problem?


Log in to reply