Problems with curves

It has been a while since I used my printer and recently I decided to try printing with it again. I have a mendelmax 1.5 running with ramps1.3 and using marlin firmware, though I am unsure of the exact version, it was compiled Nov 2012.

When I print curves, the print does not come out symetrical. I was printing a bracer, and as can be seen in the image in the zip file the right side does not line up compared to the left (we are looking at the bottom).

To help trouble shoot I produced a test file with a rounded corner, and the box prints out close enough to the dimensions it should (50x100x10mm) with the exception of the rounded corner, which seems to be getting clipped, the radius should be continous . It makes no difference if the rounded edge is in the bottom left or top right, it still gets clipped, and I have tried printing via pronterface and directly from the sd card. I included the stl and gcode of the test object, along with print images. I couldn't include the bracer as it was too big a file.

I measured the movements of the axis themselves and they seem ok. I looked at the gcode in an external cam package and they seem to have the correct shapes in them, the corner looks nice and round. I used Slic3r to do the slicing, but I don't think the problem is there. Anyone know why my curves are messed up? I am guessing this is a firmware issue, but I am not sure what to change, as the actual linear motions seem to be working correctly.

Your gcode looks good. I would expect you have some serious backlash in your X or Y axis. Check for loose belts or loose pulleys.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

I have looked at the backlash, but don't see anything obvious. When the print is going, I can see that the stepper for the X motion is not moving during the straight part of the what should be a round corner. The stepper is not 'firing' to make the edge round. Surely if backlash was the problem, I would be able to see/hear the stepper trying to move, but not engaging the belt correctly?

Ok, did some testing, and if I lightly push on the x carriage with my finger as it it is going up the 'straight' getting ready to enter the curve, then the curve comes out fine.

Would this still point to backlash? I was wondering if this pointed to the guide rods on the x axis being to tight and therefore causing some momentary resistance before they started moving, or perhaps needing more juice to the stepper motors? I have the X carriage belt as tight as I can get it, so if it is still sounds like backlash, I don't know how else to get rid of it.

I am surpised though that if it is in the mechanics why I am still getting a closed loop. Seems like the steps wouldn't add up around the corners and there would be a gap.

is your x and y exactly 90deg from each other, and can you place a ruler on the table and move 10mm in x dir at a time and does the machine move 10mm each time? does it do this for the x and y axis. use the nozzle as a reference.

Well, I should have read NewPerfection's message a little closer. I went around and tightened all the belts, however I didn't check the pulleys. Turns out the set screw on the x-carriage pulley was a little loose, so I had major backlash. Guess I still don't understand why that wouldn't have lead to a gap, or to a rhombus type shape, but tightening it up seems to have done the trick for now anyway.

Goes to show you can't just play with the belts and think everything is good. I didn't even really think about the pulleys.
I am printing my test print now, and I seem to have nice round corners.

Guess after sitting for nearly a year things can work themselves loose.
Thanks for all the help.