Mini Kossel Layer Shifting

80 views
Skip to first unread message

Kamil Jiwa

unread,
Apr 23, 2020, 4:50:59 PM4/23/20
to Delta robot 3D printers
Hello. I'm bringing this discussion over from the Duet forums since I had started it there initially thinking this was a software calibration issue. Now I'm certain that the issue is mechanical.

I've been printing a 20mm calibration cube and I get layer shifting at junctures where the layer patterns change. For example, if I print with 3 shells, then the outline of the 4th layer will not perfectly match the outlines of the 3rd layer. Here's a picture where you can see the shifts and other artifacts:

IMG_20200328_211902973.jpgIMG_20200328_212011171.jpg


I tore the printer down and rebuilt it twice, but the layer shifting has persisted through it, and always in the same places. My guess is that there's some mechanical binding occurring somewhere. Here are some things I did during the rebuild:
  • I measured all of my beams and used the ones that were most similar in size.
  • I tightened the top and bottom frames with belt clamps to make sure there were no gaps and there was no play during assembly.
  • I cleaned and lubricated my rails.
  • I measured my delta arms and made sure they were equal in length.
  • I calibrated my FSRs and their mounts so that they are sensitive to a light but firm touch.
  • I replaced all of my timing belts.
  • I put rubber bands around the base of the delta arms in case there was any chatter.
  • I tried swapping the X, Y, and Z axes, but the shifting remained in the same locations.
During this process I had some findings:
  • One of my delta arms had a ball joint that had come unscrewed. I reset the joint and matched the length as best as possible to the others.
  • With the tools I have, the best accuracy I can get measuring my arms is to within 0.2 mm. Mine all measured in at 214.4 mm.
  • One of my rails has a carriage that is missing a number of bearing balls. I've ordered some new ones, but the rail still glides well enough for now.
  • The linear guide rails seem a bit beat up (they were secondhand), but they slide smoothly.
The layer shifting persisted after the rebuild, so I'm starting to run out of ideas to try. Here are my remaining ideas:
  • Make some new arms. I'm currently waiting on parts for this.
  • Try magnetic arms.
  • Try new stepper motors.
Finally, here is a picture of the printer. It's my first build, so please go ahead and suggest anything, even if you think it's obvious.

IMG_20200423_134709443.jpg

Any thoughts on what I can or should check next? Thanks!

Kamil

Kamil Jiwa

unread,
Apr 23, 2020, 4:51:47 PM4/23/20
to Delta robot 3D printers
BTW, here's a link to the thread on the Duet forums: https://forum.duet3d.com/topic/15191/delta-printer-twisting-beginning-of-print.
Message has been deleted

Chow Loong Jin

unread,
Apr 24, 2020, 3:11:56 AM4/24/20
to delt...@googlegroups.com
On Thu, Apr 23, 2020 at 01:50:59PM -0700, Kamil Jiwa wrote:
> I've been printing a 20mm calibration cube and I get layer shifting at
> junctures where the layer patterns change.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

That's the important part there. This sounds a lot like backlash, i.e.
when approaching the same intended XY coordinate from different
directions/speeds, you end up in a different actual XY coordinate
depending on where you approach it from.

What you should try is to remove slop as much as possible and maybe
print slower. A typical low hanging fruit is to apply tension springs
between each pair of rods, as close to the joints as possible.

Then try printing at a constant, lower speed, and reducing your
acceleration.

> For example, if I print with 3 shells, then the outline of the 4th
> layer will not perfectly match the outlines of the 3rd layer. Here's a
> picture where you can see the shifts
> and other artifacts:

--
Kind regards,
Loong Jin

Kamil Jiwa

unread,
Apr 24, 2020, 11:59:29 AM4/24/20
to Delta robot 3D printers
Loong, thanks for your response. I've tried printing at slower speeds, but I see the same issues. I've also tried using rubber bands around the arms at both the top and bottom (where the joints are) since I don't have springs. That said, I think the arms may be suspect, so I'm waiting on materials to build some new ones.

Hamp Stevens

unread,
Apr 24, 2020, 1:02:06 PM4/24/20
to delt...@googlegroups.com
Kamil,

Check this guy out. These are the best designed arms I've ever seen.


Good luck.

Hamp

On Fri, Apr 24, 2020 at 11:59 AM Kamil Jiwa <kamil...@gmail.com> wrote:
Loong, thanks for your response. I've tried printing at slower speeds, but I see the same issues. I've also tried using rubber bands around the arms at both the top and bottom (where the joints are) since I don't have springs. That said, I think the arms may be suspect, so I'm waiting on materials to build some new ones.

--
You received this message because you are subscribed to the Google Groups "Delta robot 3D printers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to deltabot+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/deltabot/de50aba0-35bc-4d64-8249-7166af7a88c2%40googlegroups.com.

Edward Simpson

unread,
Apr 24, 2020, 1:15:04 PM4/24/20
to Delta robot 3D printers
Kamil, I'd double check the linear rails for rust or dirt in the groves, those rails look like they were dumped in mud at some-point, so I'd start by checking them first.
Reply all
Reply to author
Forward
0 new messages