I spent the last two weeks trying to find the source of the “slanted” printing (see pic above) and solve the problem (also, going insane in the process) , here are the steps i went through to get to the bottom of it:

The problem:

  • Every print, but in particular all those containing circular elements (very visible on this variant of Wade’s extuder gear by NopHead) would be progressivelly offset on the x axis, and the problem was actually getting worse !

    Flickr Tag Error: Call to display photo '5225689935' failed.

    Error state follows:

    • stat: fail
    • code: 95
    • message: SSL is required
  • Also , when looking back at one earlier print failure i had on a bigger gear, the problem might actually have already been present, since the only part that was offset on the x axis was the circular attachment at the top:

    Flickr Tag Error: Call to display photo '5189781538' failed.

    Error state follows:

    • stat: fail
    • code: 95
    • message: SSL is required
  • Giving the extreme slowness of my machine (maximum speed of 2.4mm/s which is about 20 times less than a Reprap, but more on that in another post) , this kind of problem is really bad, since printing takes a lot of time, and the slanting problem is only really visible and can be confirmed after around 5-10 layers .

Methodology:

1. Is it a mechanical problem?

a- Z axis assembly:

Flickr Tag Error: Call to display photo '5252245485' failed.

Error state follows:

  • stat: fail
  • code: 95
  • message: SSL is required
  • Diagnosis: Earlier in the machine’s testing phase i have had problem with the whole X/Z assembly becoming slanted , which affected the prints but certainly not in the way i was seeing now.
  • Solution: i switched the Z axis threaded rods with higher quality ones and redid the motor couplings, then i checked for any alignment problems with a laser pointer (crosshair generator) attached to the assembly:

    Flickr Tag Error: Call to display photo '5225688847' failed.

    Error state follows:

    • stat: fail
    • code: 95
    • message: SSL is required
  • Result : It reduced/ eliminated the slipping coupling issues of the Z axis,but the problem  remained !

b- X axis platform:

  • Diagnosis: During the many stages of mechanical adjustments to the extruder attachments, and due to some original misalignment of the screws holding the roller bearings  in place, the X axis platform had gotten slightly loose (the bearings not exerting correct pressure on the sliding rails), which i had fixed temporarily with …. subway tickets … (forcing them between the platform and the rails reduced the wobble), so i thought the problem might be comming from there. Also , there was slightly more backlash on this axis then on the Y one, so that might have accounted for the problem too.
  • Solution: I disassembled the platform, re-drilled the holes for the bearing screws, improve and tightened the structure to reduce wobble and i added a second captive nut right next to the first one, to reduce the backlash.

    Flickr Tag Error: Call to display photo '5252854208' failed.

    Error state follows:

    • stat: fail
    • code: 95
    • message: SSL is required
    (the weird screw like things is a special positive/negative nut)

Flickr Tag Error: Call to display photo '5252246357' failed.

Error state follows:

  • stat: fail
  • code: 95
  • message: SSL is required
  • Result :While it greatly improved the quality of each layer being printed (these adjustments were long overdue), the problem still remained !

2.Is it a software problem?

  • Diagnosis: The firmware i am using on the dual AtMega 328 chips running the machine is a rather heavilly modified 5D code , and since i work on a basis of “only use what you understand at least a bit” i have started to think that my stepper timing changes were faulty (the timing i use is nothing but a really cheap hack, without register timers…).
  • Solution: Just to be sure , i decided to try a completely different firmware , which does the timing calculations quite differently, so i used a VERY trimmed down version of the HYDRA_MMM firmware V1.4 (which is a lot easier to modify and understand than the 5d Firmware, although it does’nt (yet) support separate extrusion commands)
  • Result :I spent a few hours adapting the firmware, and i made a few test prints … the problem , as ever was still present !

3.Is it a power supply problem ?

  • Diagnosis: I have had problems in the past every since i installed the new heated bed (which uses way too much power), so i wondered if it was “gobbling” up the power to the steppers, and causing a malfunction.
  • Solution: Just to be sure, i did a few test prints without the heated bed, to see if things were any different..
  • Result : Once again,  no change, the prints were still slanted!

THE answer!

I didn’t know any other thing to try, then a friend suggested simply swapping out the X-axis Pololu Stepper driver and that …did the trick!! I was baffled, but i have been printing day and night ever since and have had NO problem at all !

So there you go , perhaps there was a glitch in the connection between the stepper driver and the motherboard, but either way, in case you run into this kind of weird problem:

  • either adjust the pot of the stepper boards
  • or swap the boards out ..
Share and Enjoy:
  • Identi.ca
  • Twitter
  • Facebook
  • Reddit
  • Digg
  • Print
  • Google Bookmarks
  • Add to favorites

Tagged on:                             

2 thoughts on “Doboz repstrap : solving the slanted print issue

  • March 4, 2013 at 22:03
    Permalink

    Have you ever been able to figure out what causes this problem?

    I’m facing exact same problem. If I print a single layer wall, no slant problem. If I print an object that has double wall or have filling, the printed object slants and can be quite bad.

    My mechanical machine is a milling machine which is running fine and is capable of milling aluminum just fine, so it doesn’t look like mechanical problem to me.

    I’m using Replicator & Teacup firmware on a Sanguino board. The Sanguino board drives a 4-axis Chinese stepper driver board which in turn drives the 4 stepper motors. This same stepper motor driver board is working fine while I’m milling with EMC.

    I’m trying to print the 20mm cube as test object which slants quite badly. If I get rid of the G code that’s responsible for filling, the slant problem is less but is still there. If I take this G code and removed all the M gcode (which EMC doesn’t understand) and convert E to A, EMC can draw this cube straight and with no problem. So, G-code, mechanical setup, and stepper motors are working fine. This makes me think the firmware and Sanguino board are the problem, or maybe its timing problem that you mentioned above.

    I’m trying to load different firmware (Marlin) to my Sanguino board to see if it helps any.

    Hope to hear from you

    Reply
  • March 4, 2013 at 23:16
    Permalink

    Just read your post again, and realized you fixed your problem by replacing the stepper driver. I’m puzzled as my stepper driver works well under EMC control.

    Reply

Leave a Reply

Your email address will not be published.