Circles not round

Hi,
I'm having trouble with round circles at any scale, they're about 0.5 mm oval difference at a 45 angle to the axis and seem to happen when the print direction reverses. The inner circle prints round then reverse direction prints the outer circle oval. (See Pics)
I'll explain the config:
I'm using an H Bot G2 belt config on a custom "Uconduit" style FDM printer. My previous printer was 200x200 and I didn't have this issue. The current printer is 300x300 and now I'm having this issue.
I'm using 40 tooth alu drive pulleys tight on the shaft.
Motors are not skipping or getting hot so I assume the current is set right.
Belt is snug but not too tight.
The frame is square.
The Marlin config for calibration is 40 for both X and Y (default_axis_steps_per_unit). No bearings are dragging. The belt runs true. I don't see any jerk or hop in the motion.

I've run out of ideas how to fix this. Does anyone have a suggestion?

Thanks!

Edited 2 time(s). Last edit at 03/02/2018 04:08AM by MarkG.

Does your hotend tilt? Does it get bettter, when printing slow?

Looks to me like severe backlash in one axis, probably the one that is the up-down direction in your images.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Hi O_lampe.
The hot end has no wiggle and seems firm in the mount which is also firm on the bearing and rods. Speed doesn't seem to effect it at all

Edited 1 time(s). Last edit at 03/03/2018 01:15AM by MarkG.

Hi dc42,
I thought the same but I can't see where the backlash is coming from. All parts seem firmly attached,. The motion seems smooth when I push it by hand and I don't feel any binding. I've tried more and less tension on the belt. My X and Y axis is forward/ back and side to side. The Z axis is up and down so it's somewhere in the X and Y. Heck... I've even tried adjusting the step settings in the firmware to adjust it out but that just made things more oval in the axis direction.

FYI...
I've also tried another simple circle design file from Solidworks saved as an .stl file using different slicers and software like Simplify3D / Slic3r / Pronterface / Reptier / and Cura with the same results.

What about Jerk settings?
How would that setting in the firmware affect it? What would I adjust and what should I expect?

Edited 1 time(s). Last edit at 03/03/2018 01:23AM by MarkG.

Maybe you should try printing [www.thingiverse.com] ?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Considering it's an H bot, have you considered this could be from racking of the gantry? I don't think it quite matches when the gantry should experience racking, but depending on how your printer is constructed it's possible the racking action has loosened and created some slop in your frame or gantry that is then causing these issues. That being said I don't have any experience with H bots so this is just a guess.

Definitely looks like a mechanical issue though, but it doesn't quite look like backlash to me. The print looks skewed, whereas with backlash issues I've seen people printing circles that have flat spots where the direction changes (i.e. at the max/min of the x and y axis travel).

Edited 1 time(s). Last edit at 03/03/2018 08:55AM by Trakyan.

Reminds me a bit of the "skewed tower" problem DJ had with his Delta. He was using steel core belts, were one belts steel core was broken.
Are you sure your belts are OK?

Hi,
Sorry for the delayed response. I've been checking and tweeking everything to find the best settings for the best print and get the best results and pictures. I've printed this ALOT.

Rotating the part counter clockwise, Img_1 and Img_3 show the travel along the Y axis. Img_2 and Img_4 show the travel along the X axis.

You can see the most dramatic event on the X travel axis Img_2 and it seems to be just one line only. The opposite side has a near perfect result at 0.0 (Img_4) which I find hard to explain.
For travel along the Y axis, Img_1 and img_3 shows the best results at 0.12. The differences are barely noticable.

Would this reaction be causing the un-round circles? So what does this tell me??? Since I'm using Simplify3D I don't have a software backlash adjustment settings. Is there some adjustment in the firmware??

Thx
Mark

Edited 1 time(s). Last edit at 04/03/2018 01:59AM by MarkG.

Hi Everyone
I still need a little help with this.
I've gone over everything including printing the suggested backlash tests suggested by dc42 and can not find the cause of my circles out of round. The backlash tower test from thingiverse provided a clue tho. If I pring with the object square to the table my Y prints well and smooth by the X has a profile all the way up.If I rotate the object 45 degrees to the table all surfaces are smooth.

Any have suggestions at this point would be helpful

Thanks!

A picture of the printer and it's belt paths might be useful.

Are the parts being printed on the center of the bed or closer to one corner? How fast are you printing? Do you slow the machine down for small perimeters?

---

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

Depending on what the slicer does, changing the orientation of an object may change the path of the nozzle, and therefore where any backlash errors appear.

If you use auto-leveling and your bed is very out of level, backlash in the Z axis can manifest itself in X or Y axis errors, depending on the path of the nozzle.

Hi the_digital_dentist,

There are some pics in the first post that show the offset circles. Here are some pics of the printer and the latest test print. The X/Y drive is a typical Hbot / Core XY type with a G2 belt setup. I've tried different locations on the bed, same results. Bed is level belt is tight and nothing is loose.
I'm using Simplify3D slicer and printing this test tower in vase mode which is painfully slow and single wall. Printing fast gets the same result only slightly sloppyer.
Whats puzzling me is the out of round is always the same 45 degrees from the XY axis. I do notice that the circle is round as long as the print continues in the same direction. When it reverses direction the pattern shifts 45 degrees in the +/-X +/-Y direction. for a total offset of about 1mm. This test tower is supposed to be a backlash test and when printed square to the bed I see the Y axis smooth and the X axis offset on the +X side only.

Any suggestion is greatly appreciated

Edited 1 time(s). Last edit at 05/23/2018 12:52AM by MarkG.

Hi frankvdh,
Yep I thought the slicer G code would be different as well just wanted to see if the offset followed which it didn't. The bed was manually leveled before the print and the first level is good. Z axis on this printer is for the bed.

CoreXY and Hbot are different, HBot has racking/skewing issues which could be causing what you're seeing. Do squares come out square and with straight sides?

Quote
MarkG
Hi the_digital_dentist,

I do notice that the circle is round as long as the print continues in the same direction. When it reverses direction the pattern shifts 45 degrees in the +/-X +/-Y direction. for a total offset of about 1mm.

This really does say to me that there probably is play in your mechanism. A couple of thoughts; how are your belts tensioned? If done with springs, that's where your backlash is hiding. With the motors powered on, try moving the nozzle around by hand... if you can move it, then there's play. Don't forget to check the grub screws that hold the pulleys on the steppers, and the stepper mounts, and any other pulley mounts.

The belt paths look OK.

I have to agree with Trakyan- since the problem crops up when the direction is reversed, it sounds like backlash caused by racking - the undriven side of the Y axis is lagging behind the driven side. Try turning the motors off and manually wiggle the driven end of the Y axis. Does the opposite side follow exactly?

The X axis guide rails are separated vertically- probably not the best arrangement for an H-bot because you really need the lateral stiffness that spacing the rails horizontally would provide.

Edited 1 time(s). Last edit at 05/23/2018 09:51AM by the_digital_dentist.

---

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

Hi I’m in no way a “professional “ just a hobbyist, but have had the same issue with my printer
Printing pistons and cylinders. I noticed that they were stretched along the x axis so I scaled it down on the
X axis 8% in the slicer (cura) and came out much better. Not ideal but I think it’s a firmware problem.
Hopefully a better solution comes through eventually

Edited 1 time(s). Last edit at 04/23/2020 05:07PM by 79FRANK.

8% is a YUGE error!

---

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

Hi folks!
Thanks for responding. Since I fought this error a while back without a reasonable solution I moved on to a different printer and let this go. The original printer is a core XY single belt setup. I didn't seem to have this skew trouble with an 8" bed but when I modified it for a 12" bed the error became obvious. It seems to be a general point of view that this error is an inherent part of this single belt function due mostly to backlash issues. BUT, in my quest, I stumbled across a potential solution on Thingiverse. Search "Skew Test" or "Skew Calibration". If I remember correctly you can compensate using G-code. I did not try it but maybe it will get you where you need to be.
Good Luck!

Hmmm. It isn't that hard to get the axes square. Why would you try to compensate for skew instead of fixing it?

---

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

Hi,
The axis is square that's foundational. As I stated excessive backlash is an inherent problem with a single belt CoreXY. One belt driving two axis in alternating directions may have enough slop to cause the nozzle to drift a bit. This will show itself in a repeating pattern. Think Man!

How do you make a corexy mechanism that uses 1 belt? Do you have a picture?

---

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

I believe the single belt corexy being referred to is an H-bot. The attached image makes it clear, the systems are very similar, but H-bot printers indeed use just one belt. This error sounds like it is caused by the y-axis racking when reversing direction, which would cause the axes to be out of square and consequently print ovals instead of circles.

Then it's H-bot, not corexy.

---

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