DIY cartesian build. issues tuning

So i have built my 3D printer. Frame came from CR10 but there is not much left of it except 3 of the motors (X taken from Y and dual Z).
My printer is running DuetWIFI, has 5" LCD panelDue, IR bed probe, E3D titan aero, NEMA23 on y axis, 420x410x8mm precision alumiunium plate with .5mm PEI on top and 1kW keenovo heater, running on preloaded Bosch Rexroth 15mm rails. X axis runs on LM THK 22mm linear rail and Z axis on IKO 12mm rails. Basically my bed has no slop whats so ever, extruder assembly rock solid with no play and Z rods dont seem to wobble.















I am having trouble diagnosing print quality issues so there might be someone who might help. I have some ringing or ghosting artifacts. My belts are tightened to 70HZ, my Z is braced and all is solid. I have it sitting on 40mm stone tile. I tried printing slow and fast. Setting my jerk from 700mm/min down to 50mm/min Lower values gave slightly better results but nothing considerable. Here are my settings at the moment:

M92 X80.06 Y80.22 Z402.43 E823.54 ; Set steps per mm
M566 X380 Y380 Z12 E120 ; Set maximum instantaneous speed changes Jerk (mm/min)
M203 X10000 Y10000 Z1000 E1200 ; Set maximum speeds (mm/min)
M201 X700 Y700 Z700 E9000; Set accelerations (mm/s^2)

I don't think that's movement related, but rather temperature and flow rate. Try reducing the flow rate by 3% and the temperature by 5°C and see if there are any improvements.

Edited 1 time(s). Last edit at 02/19/2018 09:25AM by jerryjs8.

---

Build Log - DDDE 330x330x300mm Ultimaker

Print a large cylinder- large curves will create far more ringing than abrupt changes along straight runs.

You are throwing a very massive bed and its support back and forth at print speed. It's going to ring. That's why large printers usually move the bed in Z instead of Y.

You can probably eliminate the ringing completely if you set the jerk/junction deviation to zero and reduce acceleration. But that may may introduce other problems (besides slowing printing to a crawl)...

I built a machine with a 300x300 bed and ultimately replaced the Y axis belt with a precision ground ball screw. It almost completely eliminated ringing, but it's pretty loud. All that metal to metal contact is very efficient at transferring vibrations from the motor to the bed plate. That printer is limited to 40mm/sec print and travel speed. Click the son of megamax link in my sig, below.

Edited 1 time(s). Last edit at 02/19/2018 09:39AM by the_digital_dentist.

---

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

If its only in the Xaxis its the belt. Do you have a toothed or flat return roller?
Add a tensioner/dampener at the return roller end.

I'd do a test cube print, to see if its just at the corners or along a whole edge. if whole edge its the belt.
My anet A2 has this problem, it has smooth return rollers. I can minimize it with tension, but it still visible in the reflections.
You could some spectra braided fishing line just to see if thats it.

If its just the corners its the accel/jerk combo.
Be sure your settings are being saved, if you have enabled the eeprom save.
Nothing worse than thinking you are changing a setting and its not being saved.

Finally, if you think all those are right, its like the previous post says, something about flow/retraction at the nozzle.

Nice build btw !

Thanks. So larger motor and wider belt is not helping? I did print bechies with almost no ghosting or ringing, but they are small objects. Ciuld be that something i changed in slicer or firmware or smaller object would have lesser effect?

Its kind off weird. I have more on the y and less on the x but also the shape of it is different. Btw its not on the corners, its along all the way. X is having repeated diagonal lines and its like you said more or less when the light hits ar certain angle. Pla feels a lot less than petg. I will get more photos at correct angle.

OK. I printed large cylinder, well, not that large @ 100mm dia, in PLA. Most of that nasty ringing is gone. I have some faint diagonal ringing (3rd photo) but my thought is that it comes from X. The belt is pulling motor out of square (crappy CR10 bracket reused) and when it goes one direction belt twists to one side and vica versa. Could that be the cause of diagonal ringing?
Next i now have consistent vertical lines as if it was made in sketchup with certain amount of lines forming imperfect circle. This was made in inventor and i had none on my prusa prints?! It does look round from the top








I now will fire up PETG and see whats going on.

BTW, how would you rate that seam?

Forgot to mention, that first grey PETG print was with hardened steel nozzle and i swapped it for brass to see if it makes any difference.

Edited 2 time(s). Last edit at 02/19/2018 01:47PM by agniusm.

That seam looks pretty good, as does the rest of it. I don't think you have a ringing problem...

The vertical lines are probably the STL file's polygonal approximation of the curve.

---

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

Thanks. But can you notice diagonal lines on the 3rd photo? i have another problem which might relate. When i print round object, circle across the bed is not the same thickness. Diagonally its thicker and then narrows. I dont know if its noticeable in the photos below. If it was perpendicular to the edge of the bed steps could be at fault but here i don't know. If i print 100x100x10 5mm tall angle one side is .05mm out and the other 0.2mm.

If the print shape changes depending on the print position on the bed, you may have a belt path issue. Belts should be parallel to guide rails. If they aren't, the steps/mm will vary with the position- most accurate near the center and worse as you get closer to the ends of motion. This diagram illustrates a belt mechanism moving the extruder carriage (the square where the belt attaches) along the X axis guide rail (the big rectangle at the top).



The top is the bad situation, the bottom is how the belt should run. The segments of the belt that connect to the moving part (extruder carriage or bed) have to be parallel to the guide rail(s).

---

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

You See, i cant grasp around this concept. motor pulley is teethed and that angle does not do anything in my opinion. I have slight bend on the belt but the artifact is across all 5 samples. I have checked and my beds edge is not perpendicular to x gantry. i am not sure if its only the med or motion system as well. That would definitely have so impact of axis being skewed,
I am checking if i can compensate that in firmware and see if it makes a difference. I lean towards it cause changing steps would render more changing steps, i cant tune it in. That would explain why. If the axis are skewed , changing steps is no linear any more.

Unfortunately, the math doesn't care about your opinion.

This picture illustrates what happens when you offset the belt anchor point and cause the belt segments not to be parallel to the guide rails.



The difference in the path length (C+D) when the carriage is at different locations matters because belts don't like to stretch. The tension will increase as you get toward the ends of the motion, maybe enough to cause the motor to skip steps. The other thing that happens is the controller tells the carriage to move 20 mm, but since the belt isn't parallel to the guide rail, it won't move 20 mm it will move less than that. As you get closer to the ends of the axis, the belt gets further from parallel, as the 20 mm movement gets even shorter. This creates distortion in the prints.

If you want the carriage to move the exact amount it is instructed to move, the belts MUST be parallel to the guide rails.

If circles are not circles, the two most likely causes are that the belt paths are not parallel to the guide rails or the axes are not square.

Edited 1 time(s). Last edit at 02/20/2018 05:29PM by the_digital_dentist.

---

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

Thank you. Sorry i have not responded sooner. Kids
I might have a slight misalignment. I reused creality CR10 pulley and it is larger than the one on the motor side. I have teethed one so i might swap it. Y axis i think is better. Here are some photos of the belts

The idle section of the belt doesn't need to be parallel to the guide rails, just the sections that the carriage is connected to. So you may be able to correct it by moving the pulley slightly. In the second photo it looks to me that the left hand pulley needs to be raised and the right hand pulley lowered in order to keep the lower sections of the the belt parallel to the rail.

Edited 2 time(s). Last edit at 03/16/2018 12:18PM by dc42.

---

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].

Thanks David. So basically my round objects being not round is more of a belt being not parallel rather then skew of the XY as i did print XYZ skew compensation sample and did offset it but it made little difference?

If the X belts that connect to the carriage are not parallel to the X rail, then I would expect the X size of the cylinders to depend on the X position at which you print them.

Edited 1 time(s). Last edit at 03/16/2018 04:52PM by dc42.

---

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].

The thing that bothers me that its not related to certain axis. Its more diagonal to XY axes and that is probably why i had hard time calibrating X and Y steps. I would change one axis and the other goes out of tune. One hint is that printing a tube does not make it oval, it probably is slightly oval, but the wall thickness goes from thin to thick, like in/yan diagonally so its not tied to cerain axis. I have now squared my machine to best of my abilities And i think i get slightly better results(?). Tomorrow i will try to square my X with some padding or rather get it out of square the opposite direction and see how the print reacts.

I tried moving my right pulley up so it is parallel to X and printing 3 round tubes was rendering most left tube to be roundest which is counter intuitive.

Does it give the appearance of printing the circles at constant speed, or does it speed up and slow down depending on the position within the circle?

You should not need to calibrate the axis steps/mm because they are defined to fairly close tolerances by the pulleys, belts and leadscrews that you use.

Edited 1 time(s). Last edit at 03/16/2018 04:53PM by dc42.

---

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].

Nothing noticeable but i will look and see if it does.

OK, something weird. The squareness of the axes does not have dramatic impact on the print. I have spaced out right side by 1mm and printed a circle, then paused a print and transferred 1mm spacing to the left side and resumed to print second circle and they looked same, same problem in the same place.

Printed 9 circles around the bed to see if the belts where to blame but they have same artifact all over. I have made a video to show the problem. Running our of idea here. I have set my XY steps to 80 (X80.06 Y80.22 before) and that did not do anything.

[www.youtube.com]

You can see it has that bulge diagonally.

I will try to disable pressure advance now

P.S. Disabling pressure advance - no effect
Aligning Y pulley - no effect

Edited 3 time(s). Last edit at 03/17/2018 04:17AM by agniusm.

Which slicer did you use? Some slicers produce much more consistent gcode for circles than others.

---

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].

I use Simplify3D or slicer

This is a shot where i spaced out right side by 3mm:



Also changed Y steps from 80.22 to 85 and circle came out oval with the same yin yang artifact (could i have just came up with the name for 3D printing error? )

Edited 1 time(s). Last edit at 03/17/2018 04:45AM by agniusm.

Quote
dc42
Which slicer did you use? Some slicers produce much more consistent gcode for circles than others.

OK, Went back to slic3r and the problem is gone. Weird, never would think it might have to do with slicer. It also stutters printing skirt in s3d and i read about it on duet forum. Really weird. So much work cause of 150USD slicer ?
Really glad i know where the problem lies as i was starting pulling my hair

Did it print the two perimeters in the same direction, or in opposite directions?

---

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].

Quote
dc42
Did it print the two perimeters in the same direction, or in opposite directions?

Wasn't paying attention, think opposite.
Whats weird is s3d was generating files over MB when slic3r 300kb

The reason I asked is that I think backlash in the Y axis would cause that type of pattern if it printed them in opposite directions.

Edited 1 time(s). Last edit at 03/17/2018 10:56AM by dc42.

---

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].