Kossel XL puts sideways jumps in long straight lines

Hi

Last September I built a Kossel XL kit from [builda3dprinter.eu]. I have been very pleased with it, especially the dimensional accuracy and repeatability. I am using Marlin firmware, printing PLA, onto a heated glass bed with a dilute PVA solution to avoid corners lifting. I design with OpenSCAD and use Kisslicer Pro.

Then I noticed the upper surface of flat prints (50mm * 80mm * 1.2mm) is not as smooth as I expected. Looking at the 'shiny' side, I can see that the long diagonal stokes to make the flat surface have jumps of about half an extrusion width in various places, and this leads to some doubling up of material. The overall size and shape of the printed parts is fine, but the surface is uneven.

I have looked back through parts I printed at the end of 2016, and I do not see this at all. On these, all the first layer fill lines are straight, with no jumps.

I have re-tensioned the belts, checked the pulleys on the slide motors, recalibrated the offsets, and adjusted the motor supply current trimpots, all to no avail.

Yesterday I hand coded a G-code file to make 3 sets each of 5 linear strokes, 100mm long in the X direction, at different Y positions. The one at Y=0 clearly shows this 'jump' on the line. This print is reproducible, with the jumps on the lines in pretty much the same place. So it is not some electrical noise or random mechanical slippage.

During this investigation, I have twice had a situation where I turned the printer on, did a home, and one slide went down not up, and I had to kill the power. On the first occasion it was the C pillar (I labelled them anticlockwise from the front left, and wanted to avoid confusion with the XYZ axes of the printed parts), and the following morning it had fixed itself. Last night it did it with the A pillar. This morning it was not fixed, but after I adjusted the A motor trimpot from 580mV to 470mV, it was OK.

I am puzzled that the printed results have got worse over a few months, even though I have not changed the hardware, and the firmware only for calibration results.

Any hints or suggestions would be welcome.

Best wishes

John

This might be either play in the rod ends or stiction "erupting" at the same place.
Do you see anything binding at the effector or carriers? ( flat spot on the wheels/bearing balls )

Edited 1 time(s). Last edit at 03/13/2017 03:56AM by o_lampe.

What @o_lampe said, plus check all the screws for tightness, especially the screws that attach the rods to the effector and to the carriages, and the whole structure for rigidity... things tend to work loose over several months of printing.

I'm guessing that one of the carriages will reverse direction at the time that the effector jumps sideways, and that play somewhere in that part of the mechanism causes the jump.

My only guess about why your motors are suddenly driving in the wrong direction is that the limit switches are sticking closed? So the software thinks the axis is already homed, and drives it down so that it can pull it up to find the point where the switch closes.Except that the switch never opens, so the software just keeps driving that motor downward.

Hi all,

Thanks for the suggestions. I checked and tightened the rods, and overall structure. I re-tensioned the belts. I was still getting the same, consistent, bad test print.

I then some further investigation on Green. I took off the cabling from the motors/endstops switches to the RAMPS electronics board, and put them back in a different order, effectively rotating that part of the printer by 120 degrees. The test print was still wrong, but not the same wrong as I had been seeing, consistently, over the last month. I then took all the diagonal arms off, and rotated the central 'spider' base by 120 degrees, with the arms swapped round and swapped top-to-bottom. And the test was back in its wrong, consistent, state. I pondered about this, then worked out the only things in the same position relative to the print were the two parts in the central print spider. The site I bought the printer from has downloads for the Kossel printed parts. I printed a new lower part (groovemount), fitted it, no better. I printed a new upper one (the effector plate), and fitted it (quite fiddly), and got a good test print. HOORAY!

I have no idea why the original effector plate was causing a problem, but certainly, replacing the groovemount and the effector plate have fixed the issue for me.

John