The thing's gone wild! In a specific circumstance...

[Humanzee]

Video of behavior: https://www.youtube.com/watch?v=RlIEp4mXtAU

I'm attempting to print a very full bed of individual pieces here. You may or may not be able to see in the video that I'm really pushing the boundaries of the print area, with even a couple of corners going outside the outer circle on the Onyx bed. More commonly sized prints have no problems.

In the beginning, It prints the rings and first layer just fine, and then shortly into the second layer it starts to behave the way it does in the above video. I thought it could be a slicer issue (MatterControl on a MatterControl Touch) as I had duplicated the part in the slicer software, so I made the full bed model in CAD directly. No change. I thought maybe it was making a long, fast move and one or more of the belts were skipping, so I tightened the belts (giving me another bed leveling issue that I don't fully understand the reason for, but will fix later and you may also be able to see in the video), resulting in no change.

Am I simply pushing the edges too far out? If that's the case, why does the first layer print okay? Is the part too physically large for the MCT, and it eventually gets confused? Any other ideas? At this point I'll just be printing these in smaller batches, but it'd be nice to be able to cram the print surface as much as possible.

[IMBoring25]

On the phone now and haven't seen the video, but most slicing profiles slow the first layer. The full speed near the perimeters might be enough for the Rambo to drop steps.

[Humanzee]

I watched it more carefully and caught it in the act of failing, and, yup, as soon as it starts going full speed, it goes off the rails and starts doing naughty things with its filament.

Is there a cure for this other than going more slowly?

[geneb]

When printing along the outer edge of the build platform, you have to slow down otherwise you'll have problems. What type of arms are you using?

g.

[IMBoring25]

Other than going more slowly, there are a few things you could try.

It is possible that a high-resolution STL is contributing, if a lot of short segments are increasing the processor overhead. Using the simplification option of a slicer that has one might help.

Just putting it out there because it's possible, mechanical changes to the machine that would reduce the steps per mm would reduce the number of steps the board has to send at a given speed. You would lose resolution and worsen segmentation artifacts, but it's possible.

The gold standard fix is a switch to a 32-bit board that will have the processing power to handle the delta kinematics and step rates required. I like my Duet but there are other options.