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

RoBo3D Autolevel via Mike Kelly

Discussion in 'Mods and Upgrades' started by Mike Kelly, Jun 1, 2014.

  1. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 11, 2013
    Messages:
    6,967
    Likes Received:
    2,276
    1) I've been thinking about that quite a bit lately.

    If we assume the x carriage is a rigid body and that when one side lifts they lift in unison, it would be fair to assume that one switch would be sufficient for this case.

    Unfortunately I don't think that's what happens. The linear bearings have some compressibility so when one side lifts it tends to bend slightly. This means that if you're measuring on the opposite end from your switch there will be an increased offset away from the

    2) This is true when using the Automatoic Print Leveling feature in MatterControl. If you're using marlin's built in feature it's more critical to get the offset accurate for both sides.

    Quick activation provides a lot of benefits, such as not putting too much force on the bed causing it to bow and give a false reading.

    Not sure about what you mean by same site
     
  2. lemuba

    lemuba Active Member

    Joined:
    Jun 22, 2014
    Messages:
    127
    Likes Received:
    135
    Hmm,

    I "believe" the x carriage is a rigid body offset wise because its driven by two stepper motors - so once the switch was acitivated, the x carriage moved parallel because both stepper motors were driven parallel the same distance...?

    I meant that both switches possible can be installed mechanically parallel and electronically in series either the left hand- or right hand site of the X-Carriage unit.

    Br,

    Matthias
     
    #62 lemuba, Aug 10, 2014
    Last edited by a moderator: Aug 10, 2014
  3. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 11, 2013
    Messages:
    6,967
    Likes Received:
    2,276
    The thing is the offset is the distanced from when the nozzle touches the bed, the coupling nut unseats, and the switch deactivates. The first two things happen very quickly (assuming not much force is needed for the nuts to unseat).

    The bend happens once the coupling nut unseats and is no longer being supported against gravity.

    Having two switches on the same side isn't necessary. One is sufficient for the sensing.

    This all said I think it's worth trying.
     
  4. lemuba

    lemuba Active Member

    Joined:
    Jun 22, 2014
    Messages:
    127
    Likes Received:
    135
    I have installed now the switches according to the standard guideline.
    Autolevel is working fine, but I´m wodering if it´s normal that the switches are moving this way during printing and the dynamic autolevel adjustments....?
     
  5. Mike Kelly

    Mike Kelly Volunteer

    Joined:
    Mar 11, 2013
    Messages:
    6,967
    Likes Received:
    2,276
    That's caused by backlash in the coupling nuts... Eventually you'll see the results of that by smushed layers and such.

    It's a bit of a double edged sword. You need the nuts loose enough to unseat easily, but not so loose they have backlash like you're seeing.
     
  6. lemuba

    lemuba Active Member

    Joined:
    Jun 22, 2014
    Messages:
    127
    Likes Received:
    135
    My entire Robo seems to be the biggest backlash itself....
    Have attached some guidance to the switch holder now to reduce movement and friction.
    And Yes - the ROBO might print good now - even PLA as I performed/installed until now:

    1. The y-carriage bed upgrade
    2. Auto Level functionallity
    3. Installed the E3D V6 hotend
    4. Ramps FAN
    5. Top mounted Filament Holder
    6. etc., etc

    But now the hobbed bolt seems to be the final bottle neck (mainly during PLA extruding).
    PLA extrudes nicely at all time now if I push/support manually - even after a rest of minues hours, etc. The E3D hotened does a great job.
    Anyway - the stock hobbed bolt has sometimes no grip and this is not related to the Filament itsels or something else - the stock hobbed bolt is just for the garbage.

    Replacent hobbed bolt on the way to Gremany now - one from the US and one from Slovenia - I`m curious who wins and arrives first:)

    To be continued

    Br,

    Matthias
     
  7. Billm

    Billm Member

    Joined:
    Nov 30, 2013
    Messages:
    74
    Likes Received:
    22
    Here are my Z29 outputs after M565 Z-.8 -1 -2 -.5 very little difference and nozzle is on the glass every time
    Start G-code
    G28
    M565 Z-.8
    G29
    G1 Z5 F5000
    I tried with cura and slic3r same thing
    If I home then enter g29 directly I get for the 9 data points
    -.02
    .35
    .20
    .22
    .31
    .24
    .21
    .22
    .10
    Which is ~.3 difference which corresponds to my first layer height of 0.3
    No matter what I try the nozzle is on the glass. Help! image.jpg
     
  8. Invertmast

    Invertmast Active Member

    Joined:
    Apr 21, 2014
    Messages:
    109
    Likes Received:
    58
    Bill,
    Try changing you "Z-.8" to -.5

    The settings i had to try a few different settings to get mine good enough.
     
  9. Billm

    Billm Member

    Joined:
    Nov 30, 2013
    Messages:
    74
    Likes Received:
    22
    do you mean M565 Z-.5?
    i tried that as well as Z.5, Z-1, Z-2 Z-5
    I am sorry that i didnt show on the picture that the top line is the Z settings.
     
  10. Invertmast

    Invertmast Active Member

    Joined:
    Apr 21, 2014
    Messages:
    109
    Likes Received:
    58
    That would be correct... I had a very similar problem with mine in that one of the wires on the auto level switch had broken from the tab it was soldered to. This caused it to do the autolevel probe points as usual, but it was always to low.
    Using M119 in the gcode window allowed me to check that the status of the limit swith's to track down which of the two was the culprit
     
  11. Billm

    Billm Member

    Joined:
    Nov 30, 2013
    Messages:
    74
    Likes Received:
    22
    Where does the ramps fan plug into the ardiuno board?
    Still no luck getting the auto leveling to work. After installing the new y axis I can't get the printer to put down a good first layer. After the G29 The nozzle always starts out on the glass. I tried the MC three point ABLC and dam if it starts out too high!
     
  12. jonebersole

    jonebersole Member

    Joined:
    Jun 27, 2014
    Messages:
    54
    Likes Received:
    16
    I had way too much play (backlash) in my coupling nuts, so took some scotch tape and cut it to the height of the part of the coupling nut that slides inside the x carriage and wound it around the coupling nut a little bit at a time to remove some tolerance (space) between the coupling nut and the x carriage. I think I ended up with about 2.25 revolutions. I had to compress it on pretty well, and then move the x carriage up/down on it a few times to make sure it was smooth enough and didn't have too much resistance. It significantly reduced the backlash on those coupling nuts. The scotch tape provides enough slip that it still activates the stop switches as it should, but also reduces the tolerance so it doesn't move back and forth as it prints. I think it helped my prints a bit.
     
  13. lemuba

    lemuba Active Member

    Joined:
    Jun 22, 2014
    Messages:
    127
    Likes Received:
    135
    I attached each one self made plastic guide side plate to the switch mount which overlaps now to the x-carriage mounting - can attach later also a picture...
    This works very well and is de facto "self adjusting" by gravity.
    I think it would be also a good idea to include/add it to the printable version of the switch mount. I will go for it as soon I can operate Sketchup accordingly...:)
     
  14. neomage2021

    neomage2021 New Member

    Joined:
    Feb 14, 2014
    Messages:
    34
    Likes Received:
    28
    Having some issues here. No matter what i do it seems to always scrape the glass. Acts like it isnt using the offset or something. It doesnt seem to matter if it is -0.7 or .7

    My code:
    G28 ; home all axes
    G1 Z0.9 ;Adjust Z offset
    G92 Z0 ;Define new Z home
    G29 ;Autocalibrate bed
    G1 Z5 F5000 ; lift nozzle

    [​IMG]
     
    #74 neomage2021, Aug 20, 2014
    Last edited by a moderator: Aug 20, 2014
  15. Billm

    Billm Member

    Joined:
    Nov 30, 2013
    Messages:
    74
    Likes Received:
    22
    I am having the same problem.
     
  16. neomage2021

    neomage2021 New Member

    Joined:
    Feb 14, 2014
    Messages:
    34
    Likes Received:
    28
    When I run the code manually in the terminal there is a definite gap when I do G1 Z0.9 after homing. Gap between the nozzle and bed looks great, but trying to run a print with the custom start code it just scrapes the bed no matter what I try
     
  17. neomage2021

    neomage2021 New Member

    Joined:
    Feb 14, 2014
    Messages:
    34
    Likes Received:
    28
    If i change it to:

    G28 ; home all axes
    G29 ;Autocalibrate bed
    G1 Z0.9 ;Adjust Z offset
    G92 Z0 ;Define new Z home
    G1 Z5 F5000 ; lift nozzle

    it at least prints but doesnt look like it is compensating for levelness
     
  18. jonebersole

    jonebersole Member

    Joined:
    Jun 27, 2014
    Messages:
    54
    Likes Received:
    16
    Question: I apologize if you already mentioned this somewhere (couldn't find it). What software you using to print with and auto level with? My auto level feature inside MatterControl only probes 3 points during the bed level configuration. And it is the outside 3 corners, so it doesn't take into consideration the bow of the glass in the center. I know this throws my prints off a bit. It would be nice to get it to recognize at least 9 points. Thanks.
     
  19. Billm

    Billm Member

    Joined:
    Nov 30, 2013
    Messages:
    74
    Likes Received:
    22
    I have the same problem. If either of you figure it out please let me know. Up to now, everyone has been telling me to keep increasing the negative amount in M565. When I home the nozzle it takes .78 mm to get the equiv of a piece of paper to slide under the nozzle. So by what everyone says I should use M565 Z-0.78 before the Z29. I have used repitier host and matter control with the same M and Z and I always get the start of the print scraping the glass. I used the mike Kelly auto-leaving code to flash the ardiuno board and it definitely does the auto leveling prior to printing but it still scrapes the glass. I don't know enough about gcode to trouble shoot the problem. I feel that I must be doing something stupid but I don't have a clue as to how to fix it. I have tried -0.2, -0.4, -0.6, up to -5.0 and still scrape glass.
     
  20. Ziggy

    Ziggy Moderator
    Staff Member

    Joined:
    Feb 20, 2013
    Messages:
    707
    Likes Received:
    530
    I use this gcode to make sure my auto leveling remains calibrated

    G21 ; milimeters
    G90 ; absolute coords
    G1 Z10
    G28 ; Home all axes
    G29 ; probe bed
    G28 X0 Y0 ; home X and Y
    G90 ; absolute cordinates
    G1 Z5.00 F4800 ; get ready to move
    G1 X100 Y110 F4800 ; move to centre
    G1 Z0.3


    Basically this gcode homes the axes then probes the bed.

    The last move is to the centre of the bed where the extruder should be positioned 0.3mm above the bed. It should be possible to then slide a sheet of paper under the extruder nozzle and move the extruder down manually until it is 0.1mm (or one paper thickness) above the bed. Note that both the extruder and bed should be at their working temperature when this test is done.

    If there is any problem with the extruder height, it is easy to work out whether to increase/decrease the Z Offset in Marlin and by how much.
     

Share This Page