3D Printer Geometry Problem

[Gina]

I've rebuilt my Concorde 3D printer and been testing.  The test print shows a fault I have never had before.  A square frame test piece has come out as a rhombus - 4 equal sides but skewed.  Has anyone else ever come across this?  I have tried with two different sized frames.  The geometry is correct in Slic3r.  Printer control board is the Duet WiFi.

[Gina]

I should add that I haven't changed any settings that could produce this effect since I was last using this printer and it was printing good geometry.

[Dr_Ju_ju]

Ist thought is that you have some belt binding somewhere.....  

After homimg, have you tried manually entering the gcode to move the head in a square\circle ?

[Gina]

No, I haven't tried that - will do.  Thanks.

[Gina]

Sending move X +200mm the Y moves by about -10mm.  Similarly with X moves.   Moving -200mm brings the X carriage back to where it was. 

Both axes seem quite free to move.

Edited June 19, 2019 by Gina

[Gina]

I should probably add that this is Core-XY.  It's as it it isn't getting the calculation quite right but goodness knows why.

[Dr_Ju_ju]

IF sending the movement codes, results in correct movement control, that would suggest that the hardware and Duet configurations are ok, which then leads to the gcode… try and older pre-sliced object and\or a different slicer.

[tekkydave]

Could it be the belt tensions? If one is tighter than the other this can happen.

[Gina]

Using the UI to move the head does not result in correct motion.

[Gina]

2 minutes ago, tekkydave said:

Could it be the belt tensions? If one is tighter than the other this can happen.

The belt tensions are about the same.

[tekkydave]

You need to isolate where the problem is - Duet or printer.

Lift the belts off the motors and make a pencil mark on each spindle. Then use the UI to make a fixed move in X or Y. Both motors should move the same as it is a diagonal move.

If they move the same then the issue is with the printer, otherwise with the Duet.

[Gina]

Both motors turn the same amount - 90° for each 10mm of movement which is correct for 20t timing pulleys.  Tried X and Y and both correct.

Edited June 20, 2019 by Gina

[Gina]

Ah...  In taking the belts off I've noticed the pulleys don't look exactly the same.  I'll take them off and check more closely.

[tekkydave]

Hmmm, looks like a problem with the printer. Either the frame isn't square (unlikely) or belt tensions are unequal.

[Gina]

Well, I have the pulleys side by side and they are not the same make - one has one grub screw the other two.  The toothed parts look the same but I'll get my digital calipers and make sure.

[tekkydave]

Swap them over and see if the shape is different.

[Gina]

There is a small difference so I've counted the teeth - one has the correct 29 teeth but the other only 18t, so that's the problem!!  Don't know where that 18t came from - GT2 timing pulleys near that size are generally 16t or 20t and in that case the difference is obvious.

[Gina]

I have plenty of 20t pulleys so can simply use two identical ones.   I'll do that and test again...

[Gina]

That was it!  X and Y manual movements do not move the other one.  Now I'll try a print test.

Thank you both.  I might never have found the rogue pulley without help - it didn't occur to me.  I'll be more careful in future!

Edited June 20, 2019 by Gina

[tekkydave]

My D-Bot uses 20t pulleys. Check your steps/mm is still correct.

[Gina]

It should be but I'm printing a 100mm square frame so can check calibration.  The config.g is correct at 80 steps/mm.

Edited June 20, 2019 by Gina

[Gina]

It's square now - at least to the eye, I can check calibration and orthogonality when finished.

[Gina]

I would have been struggling for ages with both astro and 3D printers if it wasn't for the fantastic SGL!!  Thank you very much everyone.

Once this printer is set up and properly calibrated I can resume my astro projects.

Edited June 20, 2019 by Gina

[tekkydave]

31 minutes ago, Gina said:

It should be but I'm printing a 100mm square frame so can check calibration.  The config.g is correct at 80 steps/mm.

Just checked my config.g and mine is 100 so the pulleys must be 16T.

[Gina]

X = 100.25mm  Y = 99.66mm  Correct value should be 100.40mm for a 0.4mm nozzle.  So X wants to be 80.12 and Y 80.59.  I'll correct those and then check orthogonality.