Holes and circles doesn't come out round.

Hey! Finally my prints starts to look beautiful on my bigger version of Prusa i3.

But the last thing I need to deal with is that circles doesn't come out round.

I suspect it is something mechanical?

Setup:
IGUS Drylin bearings
Nema Astrosyn Dampers
Steel reinforced belt
High torque(long body) nema 17 stepper motors.
DRV8825 stepper drivers on 1/32 microstepping.

Can you guys tell which axis i'm having problem with (or maybe it is both) and come with suggestions to the source of error?



Edited 2 time(s). Last edit at 11/15/2016 09:10AM by Hatmpatn.

Ok, so I have done some troubleshooting today:

1. Tightening the Y-belt, so tight it makes a low note when flicking it. No difference
2. Tightening the X-belt, so tight it makes a low note when flicking it. No difference
3. Removing the Y Astrosyn damper. No difference
4. Removing the X Astrosyn damper. No difference

Now I'm leaning towards that my IGUS Drylin bearings are somehow biting too hard on the shafts, would this explain non-circular circles?

It could be if the friction is too high between the bearing and the rod
Almost looks to be a particular point on the X axis.

Edited 1 time(s). Last edit at 11/15/2016 06:04PM by FA-MAS.

Building on Fas point.
Is there anything stressing the hot end such as cables, Bowden or filament path? Is the hot end loose. I am wondering if you are twisting the hot end at certain points in the X travel. It looks like something is pushing the hot end in the Y direction then at a certain X travel it releases.
Assuming you have a Pruisa style machine.

Post pix of the printer...

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

After another day of troubleshooting.

-Been making sure all the belts goes perfectly perpendicular to the the y-and x-axis. Corrected some, No difference.
-Removed 1 out of 4 drylin bearings on the X-axis. Result: Now good
circular movement where the X-axis moves in the circles.

But I still get these corners where the Y-axis makes "the turn", in the upper and downer part of the circles.

I think that the Y-axis is hard to start and therefore when the Y-bed i "makin the turn" doing the circles, the X-axis gets a little ahead while the Y-axis is still starting up.

Just thinking about this I think the only sources of errors could be backlash in the Y-belt or the bearings are making too much friction and the printer can't handle this. I'm actually thinking about doing the forbidden... I'm going to grind the drylin bearings down with fine grit paper. What do you think?

Edited 2 time(s). Last edit at 11/17/2016 01:27PM by Hatmpatn.

8825 are know to have VERY BAD slow and low current movement, try putting those in fast decay see if that help or if you have 4988 switch to them to check and rule out the 8825

Okay, Ill go ahead and swap for a 4988 right away.

Okay, so I tested the 4988 instead of the 8825... aaaand no difference.

Although, I'm kind of relieved that the drivers weren't the issue since the noise the 4988 makes is just unbearable.

I'm still leaning towards grinding the IGUS bearings down a little bit to leave more clearance for the shaft to move, but I'm not too fond of the idea. I just don't think the bed moves that sluggish to that extent that the printer has a problem with it. Oh well. Any more suggestions to a solution?

It's worth adding that yesterday I tried printing four of the same circles in the corners of the bed, and all of them came out with the exact same problem. They even matched each other when placed on top of each other.

Today I sanded the bearings on the y-axis and ofcourse there was no difference in the circles. I cant imagine the bearings are the problem anymore. I tried slowing down the printspeed to 40% and still the exact same pattern. Im starting to lean towards Cura or pronterface having some kind of bug (both versions are from like 2013-14). Maybe ill try to update both... other suggestions?

Have you ever checked that the axes are orthogonal? Print a square and measure the diagonals. Here's a test part I designed and a companion spreadsheet that can tell you the amount of error (and correction required) based on the measurements of the printed cube: [www.youmagine.com]

When you print the cube, all sides should measure 100 mm. If they don't your steps/mm calibration is off. Opposite diagonals in each plane should measure the same or the axes are not orthogonal. Enter the measurements into the spreadsheet to calculate the required correction.

If something is interfering with the filament feed from the spool you can have this sort of problem...

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

It's difficult to tell from the photo exactly which problem you're looking at. Is it that the circle has a flat area as the axis reverses direction? If so, that's a classic problem. This issue is that there is a lag between commanding movement of the motor and generating enough torque to overcome any friction in the bearings, which is compounded by any lash in the drive chain which includes any stretch in the belts, however small. Circles are the worst case scenario for this as each axis slows to 0 and then does a reversal, during which all the lash has to be taken out of the system and then sufficient torque generated to start the axis moving again. A straight section instead of the smooth curve is the result. Drylin bearings are nice, but I suspect they are pretty much guaranteed to have more static friction than a ball bearing, so my suggestion would be to try a set of ball bearings and see if that resolves the problem.

Thank you The Digital Dentist! I'll make sure all the axis are orthogonal before doing anything else!

Thanks for you answer JamesK! Yes, exactly that is my problem, the circle has a flat area as the y-axis reverses direction. If the orthogonal-axis test fail I'll try to switch to my LM12UU bearings instead today.

So I printed two layers of the calibration "cube" to measure the X- Y- and diagonal lengths.

This was the result:
100mm on X was 100,17mm.
100mm on Y was 99,91mm

Longest diagonal was 141,50mm where the ideal was 141,48

I interpret this as that the axis are pretty much orthogonal.

Now on to switch the IGUS bearings to LM12UU.

So I switched to chinese LM12UU bearings. Now the circles are perfect! Thanks for the help guys. IGUS bearings suck. Don't buy it.

That's good news, thanks for letting us know. It's not quite that the Igus bearings suck, it's just that the characteristics are different, which leads to different trade-offs. And there sure are a lot of trade-offs to make in 3d printing