Skewed print any ideas?

Stupid question time, have you tried raising your build plate up by 50mm? That would help to isolate the problem, although easier said than done.

It would be interesting to raise the bed 10mm and see if the error-zone is then 40mm or still 50mm

That's a great idea - I now have a new list of things to try this evening - really appreciate the help.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

While you're at it enable bed probing and try the new height map visualizer. I'd be so curious to see how it does or doesn't line up with the symptoms. The ability to see what the printer sees is awesome!

Got a link for that? I assure you my height map file will be quite uninteresting.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Just thinking how to raise the bed it's definitely doable.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

I've built a Kossel XL with similar rod lengths and dimensions ( 330mm rods 6x4 , 40mm apart ) and found it hard to get it working with the heavy diamond hotend.
It seems to me this is the max. length for these 6x4 rods, before it gets ugly. Maybe you're running in similar problems?

[forums.reprap.org] Challenge Accepted, I'd love to be wrong here, but I bet you something is fishy in the visual representation, which of course doesn't REALLY prove anything because a mechanical problem could cause effector tilt or poorly located probing points etc, so a messed up bedheight isn't that probative, my thought is something is causing the bed height to be off, which throws off autocalibration which exacerbates a smaller mechanical problem. Its literally the only theory I have left that you haven't ruled out haha.

Edited 1 time(s). Last edit at 12/22/2016 01:26PM by clearlynotstefan.

Could be but it's the sudden onset of the problem which surprises me. I printed an order for someone 2 weeks ago which was cylindrical and he remarked on how good the parts were. So it was printing really well now very much the opposite.

Thanks for the link. Definitely will try these new versions and the height map.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

As promised my really rather dull (as its quite flat) height map.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Is that red peak -x? I have no perspective on it

Well +x in this diagram is towards left of the grid shown and -x towards right, so the high spot is near the y tower - I think, but even if it was, I am printing another leaning tower right now with grid disabled right in the middle of the bed. I am going to try DC42's suggestion of just moving up a small amount checking the coordinates then homing and coming back to the same place to see if its different.

I see where you are going with this - maybe that odd point is throwing off my calibration? I could set the calibration radius to just 100mm and just calibrate off the centre of the bed which is very flat and try again.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
dc42
Thanks, I'll try printing it during the weekend. Meanwhile, two more suggestions:

1. Is the bed clipped to the bed support? If so, could it be moving?

2. Change your slicer end gcode if necessary to not home the printer, just raise the head a little. When the print has finished, jog the head until the nozzle is lined up with the edge of the top of the print and read off the X coordinate. Then home the printer and repeat. How do the two X coordinates compare? Do they match the position of the edge of the print given in the gcode?

I upgraded to 1.17RC3 and webcontrol 1.14 grid compensation off. My config.g config.g

1. The bed is normally clamped although since I've removed it and replaced it a few times to try to ensure it is as level as possible, I have the clamps off at present, but printing with clamps on/off makes no difference - same skew. It doesn't move around, and certainly not the amount it would have to, to create this skew.

2. I tried this repeatability test first using the 100mm high object I sent to you, and after homing the nozzle returns the exact same place at x3.3 y3.3 z100. However at z100 the object is straight and has been since z50.

So I cut the object at 40mm discarding everything above that, and tried again, at this height the object is skewed. When I asked the nozzle to return to the last printing coordinate after homing, it returns to the same position.

Sorry for big edit here I missed a minus sign. No the coordinates return the nozzle to the same position on the skewed object after homing, so I guess we can rule out calibration errors.

This is my m122 after printing this last object m122.txt

This is the gcode for that object LargeKossel_axesorthogonalitycompensationobject2.gcode
I don't know with any great confidence how to see whether those coordinates match the gcode, I'm not very confident deciphering gcode.

Edited 4 time(s). Last edit at 12/22/2016 06:08PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
clearlynotstefan
Is that red peak -x? I have no perspective on it

So I changed my auto-calibration to only calibrate using the centre of the bed at 100mm radius rather than 135mm.

Printed a leaning tower again. So even if I have one rogue point which is higher than it should be (and we're only talking 0.25mm here) its not throwing the calibration off.

Next to see if I can raise the bed 50mm and try again. Come to think of it I do have about twenty 50mm cylinders lying around, albeit they're slightly leaning.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Don't forget to check the area around your printer for some hidden cameras, this is looking more and more like someone is playing an elaborate prank on you!

PMSL - might be whats happening.... anyone capable of sneaking in here and causing this fault is a highly skilled prankster and should be taken extremely seriously.

however there is progress - finally a test which yields some change. Good thinking Nebbian -

So I raised the bed up temporarily by 40mm, using some wonky cylinders as legs, knew they would come in handy.

Caution doing this as you really have to check you've reset the homed height and the probing distance or you get a nasty head crash, trust me.


Reprinted a 40mm version of the leaning tower and its only leaning for the first 10mm, then its straight - don't know if the photo does it justice.

So I think I can definitively rule out software being the cause in any way. It must be mechanical, still totally perplexed at what can be causing such a massive skew in just the lowest 50mm of the printer but the field of potential causes has narrowed considerably.

Edited 1 time(s). Last edit at 12/22/2016 07:36PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

That's progress! Well done on that test, it's a great result.

Are you using a flying extruder? Could it be binding when the nozzle is low?

I'd also check the lower section of your linear rails (especially the Z rail)with a forensic attention to detail.

Edited 1 time(s). Last edit at 12/22/2016 08:11PM by nebbian.

It could also be some belts that are stretched in certain areas (but not in others). Just throwing it out there as another thing to eliminate.

OK serious test time. Move your bed back down to where it should be. Now jog your effector down to 0,0,0.

Put a square on the bed, so that it is touching the effector. Jog the effector up 50mm and see if it pushes the square along the bed.

Edited 1 time(s). Last edit at 12/22/2016 08:29PM by nebbian.

It can't be the belts, as the skew happens all over the bed, so the belts are in different positions all the time.

I like the idea of mechanically checking the effector as it moves in height. Because that answers the question if it is the hot end of the effector.
Next thing to try is disconnect everything and only check movement. With an empty effector.

In the end, as it is universal and looking at the test with the 40mm columns, it simply has to be mechanical. But that again doesn't make sense. If it was localised to 1 column and 1 spot on that column, you would see a different slant on different areas of the bed. Baffling, for sure.

Just imagine the relief when you finally figure it out!

Is there any way to very accurately measure the angle of the slant on different areas of the bed? See if there is ANY variation.

Is there a chance the effector balls are not exactly round or maybe their shaft is bent by accident? Maybe the rods bind somewhere?

Good point about the belts Lykle.

O_lampe, it can't be inaccuracies in the balls, because this wouldn't change with Z height.

This is a real head scratcher, for sure.

Could one of the carriages be rotating as it moves up and down its rail, because of a bad fit on the rail or missing ball bearings? Or even because of a loose screw?

Edited 1 time(s). Last edit at 12/23/2016 08:34AM 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].

DJ

Lykles idea re the belts reminds me of something, your's look to be the Steel cored ones if so the cores do fracture one I built for a friend we put them on and it lasted about 2 weeks before they gave way (we used 16 toothe pulleys so was a tight radius) it may be worth trying the std Glass black belts (easier than a full strip down?)

HTH

Doug

Edited 1 time(s). Last edit at 12/23/2016 09:05AM by dougal1957.

Yeah the steel cores went on the belts on my corexy I bought some polyurethane/kevlar black belts, I've got some spare I'll swap some onto the large kossel its an easy test. I have 16 tooth pulleys. Might explain it being the bottom 50mm this is the most likely part of the belt to be worn out.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Still doesn't explain why its anywhere on the bed tho but still a worthwhile test I think I have a 5 m roll of that Steel cored stuff upstairs that may just go in the Bin?

Doug

I'm only half subscribed to this theory in that it took nearly a year to fail on the corexy. Not 3-4 months.

I suspect with 20t pulleys it would last longer.

Edited 1 time(s). Last edit at 12/23/2016 09:54AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
I'm only half subscribed to this theory in that it took nearly a year to fail on the corexy. Not 3-4 months.

I suspect with 20t pulleys it would last longer.

Indeed but on the one I built they lasted no more than 3-4 weeks suppose there is different Quality even on them?

It will be interesting to see but Somehow I am not holding my breath that it fixes it but at this stage anything is worth a try?

Doug

No its fixed now all I need is to print everything on a 50mm tall raft.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
No its fixed now all I need is to print everything on a 50mm tall raft.

what youve fixed it or it is still evident?