1. Got a question or need help troubleshooting? Post to the troubleshooting forum or Search the forums!

Unresolved Issues with Z Step calibration when using auto level

Discussion in 'Troubleshooting' started by applegeek, Dec 29, 2014.

Thread Status:
Not open for further replies.
Page 2 of 3
  1. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    I'll print both then try to get some good pictures of the prints. Sometimes it can be hard with different lighting and a single color object hehe.
     
    #21 applegeek, Dec 31, 2014
  2. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    I'll try both when I get home and try to get some good pictures of them. I'll do 4 prints one for each infill/autolevel setting.
     
    #22 applegeek, Dec 31, 2014
  3. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,915
    Likes Received:
    7,338
    I was thinking initially (non-R1 user here) that maybe the initial Z offset was not quite right.
    I thought that might help explain the difference in behavior somewhat.

    Not sure I would expect a large a variance in behavior as you are seeing between manual and auto leveling just due to that, but that is one difference I could imagine between the two approaches.

    I know with the beta that until you get that initial (manual) leveling done right, the prints are not going to work right They may be close, but they will not be right. Auto-leveling was supposed to make that less painful. :)

    Something else that you might want to check (but that would not differ between manual and auto leveling approaches) is the extrusion rate.
    Make sure that a 10mm extrusion results in 10mm extruded. That can make a noticeable difference in print quality.
     
    #23 mark tomlinson, Dec 31, 2014
  4. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    Here are the results of two prints. The one on the left was from the file 5mm_CS100 and had the autolevel off and the one on the right was from the file 5mm_CSAL100 and had the autolevel on. That is the only difference between both prints and they are both at 100% infill in the effort to save time of doing another 2 prints lol.
     

    Attached Files:

    #24 applegeek, Dec 31, 2014
  5. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    applegeek,

    I ran your gcode file through a simulator. The gcode itself is ok but the model was only 10 layers of 0.2mm at 100% infill. In other words, it was a solid, flat piece of plastic 2mm high. I am not sure if this gcode was just a test?

    I also looked at your log file. The log looked ok and matched your start code except that there is no G29 gcode in that log. Again I am not sure what the conditions were when that log was created? I assume that you are editing out the G29 in the gcode file you put on the SD card or are removing the G29 from the start code in Cura? Can you describe what you are doing to test with or without the G29 please.


    Anyway the way I see it right now

    1) You have the right firmware version with the Z steps/mm parameter matching your threaded rods.

    2) The G29 auto leveling in Marlin has its quirks but the function itself works well on many Robos (and other 3D models). I do not believe that the G29 function itself is the root cause of your height problems.

    3) You are getting decent quality prints with the G29 turned off - at the right height (which also verifies point 1). I can see you also have a Z ribbing issue on your printer - but that's an entirely different problem with known fixes.


    So where does that leave us?

    After reading your comments about the behaviour of the Z axis during a print I am suspecting that there is a problem with the results of the probing done by the G29.

    For example, if one of the Z limit switches on each side of the X axis were not functioning correctly the G29 would think the bed was sloped across the X axis (probably sloped by at least a few mm). However if the bed itself is actually level (or fairly level), as it prints the extruder is going to be too low on one side of the print and too high on the other as the firmware tries to adjust for the slope which is not there.

    There are a few things to do to prove (or eliminate) this as a possibility.

    - post of log of the output when a G29 is done, and

    - Test each Z limit switch works by manually triggering the switch and using gcode M119 to check the firmware sees the switch is triggered, and

    - Test and calibrate your Z limit switches as shown in this video. But before doing this just adjust your Z height above the bed so it is as level as possible by turning one of the threaded rods by hand up/down and sliding the extruder across the bed by hand to check.

    http://help.robo3dprinter.com/Guide/Homing Your Extruder/7

    The result you want is that the Z limit switches on each side of the X axis are triggered at the exact same height irrespective on where the extruder is on the X axis. You can use the gcode M119 entered manually to check the status of the Z switch.
     
    #25 Ziggy, Dec 31, 2014
    Last edited by a moderator: Dec 31, 2014
  6. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    The log file that I sent you was created from mattercontrol. This did not produce a successful print and it wouldn't let me specify the gcode only the model. The only way to toggle the auto level is by adding the G29 code to the start code within Cura itself. Here is the log file from the start from heating up to where the z calibration is done. I could only get the log file to show up in Pronterface as printing with Cura with USB doesn't show a verbose log.

    Code:
    Connecting...
start
Printer is now online.
echo: External Reset
Marlin1.0.0
echo: Last Updated: Dec 30 2014 17:24:20 | Author: (RoBo 3D, R1 Single Extruder)
Compiled: Dec 30 2014
echo: Free Memory: 3757  PlannerBufferBytes: 1232
echo:Hardcoded Default Settings Loaded
echo:Steps per unit:
echo:  M92 X80.00 Y80.00 Z2560.00 E723.38
echo:Maximum feedrates (mm/s):
echo:  M203 X500.00 Y500.00 Z5.00 E25.00
echo:Maximum Acceleration (mm/s2):
echo:  M201 X9000 Y9000 Z100 E10000
echo:Acceleration: S=acceleration, T=retract acceleration
echo:  M204 S1300.00 T3000.00
echo:Advanced variables: S=Min feedrate (mm/s), T=Min travel feedrate (mm/s), B=minimum segment time (ms), X=maximum XY jerk (mm/s),  Z=maximum Z jerk (mm/s),  E=maximum E jerk (mm/s)
echo:  M205 S0.00 T0.00 B20000 X17.00 Z0.40 E5.00
echo:Home offset (mm):
echo:  M206 X0.00 Y0.00 Z0.00
echo:PID settings:
echo:   M301 P22.20 I1.08 D114.00
echo:SD init fail
Print started at: 21:35:34
echo:Active Extruder: 0
Bed x: 15.00 y: 20.00 z: 0.19
Bed x: 107.00 y: 20.00 z: 1.18
Bed x: 199.00 y: 20.00 z: 1.99
Bed x: 200.00 y: 130.00 z: 1.98
Bed x: 108.00 y: 130.00 z: 1.15
Bed x: 16.00 y: 130.00 z: 0.03
Bed x: 15.00 y: 240.00 z: 0.08
Bed x: 107.00 y: 240.00 z: 1.27
Bed x: 199.00 y: 240.00 z: 2.28
Eqn coefficients: a: 0.01 b: 0.00 d: -0.09
planeNormal x: -0.01 y: -0.00 z: 1.00
echo:endstops hit:  Z:2.28
    Both Z limit switches are in the correct orientation and I test both switches with a multi meter to make sure that they are working properly.

    Watching the print as it first starts it appears that the first layer is very even and uniform in height and adheres well to the bed. If you want me to use a specific program to try to get a print log for you let me know and I can try to get it working but MatterControl doesn't work for me at all.
     
    #26 applegeek, Dec 31, 2014
  7. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    When I use the M119 code is this supposed to show both z limit switches or just an overall z?
     
    #27 applegeek, Dec 31, 2014
  8. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    I think what the main problem is is that the z switch on one side triggers at a different offset than the switch on the other side. How would you recommend calibrating/adjusting these so that the same amount of movement is used for each switch.
     
    #28 applegeek, Dec 31, 2014
  9. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    IMO the best free host printer software is Repetier Host. RH works well, has Cura built in and you can upload gcode to the SD Card (albeit slowly). I think you will find having Cura integrated will make it much easier for you to get from stl to printing.

    http://www.repetier.com/download/repetier-host-windows-1-0-6/

    You will need to set up the Robo printer parameters in RH eg bed size and comms speed. And you will need to transfer over your Cura settings.


    I can see from your G29 log there are some fairly big variations in the Z height being probed with about a 1.8 - 1.9 mm slope across the x axis. Only slight slope on the y axis.

    Are you absolutely sure both Z switches are being seen correctly by the firmware? The best test is to use gcode M119 which will show if the firmware actually sees the switch state.
     
    #29 Ziggy, Dec 31, 2014
    Last edited by a moderator: Dec 31, 2014
  10. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    M119 will show just overall Z but each switch can trigger separately.
     
    #30 Ziggy, Dec 31, 2014
  11. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,915
    Likes Received:
    7,338
    werd.
     
    #31 mark tomlinson, Dec 31, 2014
    2 people like this.
  12. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    Yes I suspected there was a problem there and the ~ 1.8mm slope across the bed being seen by the G29 is fairly high. The video I posted above is supposed to explain how to set up the switches (but it's not too useful IMO). Maybe @Mike Kelly has something more useful.
     
    #32 Ziggy, Dec 31, 2014
  13. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    I had to look up the urban dictionary for that one :)
     
    #33 Ziggy, Dec 31, 2014
    2 people like this.
  14. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    So I think I may have found the problem. It looks like one of the magnets on the slide for the heated bed is mounted higher in the plastic than the other three. This is causing one corner of the bet to be much higher than the other three. Any suggestions on how to fix this? image1.JPG image2.JPG
     
    #34 applegeek, Dec 31, 2014
  15. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    I think also that the other issue that I'm having is that the vertical rods are moving based on where the print head is and that slop is causing both switches to trigger at different times. I'm going to print some rod stabilizers and see if that helps with the leveling.
     
    #35 applegeek, Dec 31, 2014
  16. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    Well that should definitely not be happening. The vertical rods should be solidly mounted. Possibly one or more of the vertical rod mounts is cracked or screws not tightened?

    I can't really help you fix the bed magnet problem. I don't have that type of bed. Maybe @Mike Kelly can help sort that?
     
    #36 Ziggy, Dec 31, 2014
  17. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    So here's what is happening. The fully homed position of 0,0,0 has both z limit switches triggered when zeroed. This is fine. If I move the print head to the full right side of the platform. The Right contact switch is now pulled up off the nut by a good 2-3mm It seems like the metal rods on the x carriage and on the z plane have enough slop in them to not keep everything square.
     
    #37 applegeek, Dec 31, 2014
  18. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    This will definitely fool the G29 into thinking there is a slope on the bed when there is not.

    Is the extruder nozzle scraping across the bed when you move it from 0,0,0 to the right side?
     
    #38 Ziggy, Dec 31, 2014
  19. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    Yes, thats with the nozzle touching the bed all the way across.
     
    #39 applegeek, Jan 1, 2015
  20. applegeek

    applegeek New Member

    Joined:
    Sep 6, 2014
    Messages:
    29
    Likes Received:
    0
    So design/firmware question. How are the Z limit switches wired? Are they parallel, series, or on their own separate circuits so the firmware can level the x carriage? Second, does the firmware adjust the Z stepper motors independently when homing to z allowing one side of the carriage drop when it shouldn't or does it stop as soon as one of the two switches hits the z limit? Third, if this all is the case, is there a way to have the G28 home which is in the center of the bed when it homes Z find the level for the X carriage and then move both steppers at the same time to prob the bed without allowing the carriage to come out of level? I know you may not be able to answer these Ziggy but hopefully someone might be able to.
     
    #40 applegeek, Jan 1, 2015
Page 2 of 3
Thread Status:
Not open for further replies.

Share This Page