Extra calibration  

  RSS
eragon9199
(@eragon9199)
New Member

I am having trouble finding a thread related to this, so I am sorry if this has been gone through before.  I have an I3MK3 that I got earlier this year, and I love the printer.  I have access through my job to some high end metrology equipment and was wondering if I could use the values I get out of that to help compensate the accuracy issues with the printer.  I printed a small test peice a couple weeks ago and ran it through one of our CMM's and it was off, not massively but enough that I would like to compensate for it.  

I've worked on a project recently using an optical measuring machine to do some compensation on a direct metal SLM 3d printer which gave me the idea, but due to the nature of the project, I couldn't actually get into what they are doing with the data I got them out of their optical machine, which lead me down the rabbit hole of wondering if it was possible to do this with the Prusa.

My question is has anyone else done anything like this or do they have any ideas of how something could be set up to send use a value I get out of one of my CMM's to send back to the 3D printer and help compensate for dimensional error?  Not really a necessity but something I would like to try and do if possible, it would also let me prove some of my coworkers wrong if it can be done.  Thank you for any help you can give.

Posted : 02/08/2020 8:14 pm
jsw
 jsw
(@jsw)
Prominent Member

A related question that I would be interested in learning is what is the standard for dimensional accuracy that Prusa (and other Reprap and 3d printer developers) use?

A test cube that I did recently using Hatchbox ABS measures 19.93 x 19.97 x 19.94, or way down in the fractional percentage point error.  This, of course, was measured with my digital caliper which I'm sure is not traceable to an actual standard.

Posted : 02/08/2020 9:22 pm
peter.m26
(@peter-m26)
Honorable Member

If you know the dimensional error, why not adjust the printer? Is the error over the whole model?

Calibrate the 3d printer,  with test models, so you can compensate so the model size is good, if possible a small model, or even better over the whole bed.

Use the same filament, by good filament, keep filament dry, measure every roll of filament and adjust extrusion multiplier, keep everything the same if possible. A new nozzle can also help, replace this in a maintenance plan, so you have no wear on nozzle.

Does a smaller nozzle help? Printing time will get longer.

Do you have more info?

Posted : 02/08/2020 9:48 pm
eragon9199
(@eragon9199)
New Member
Posted by: @jsw

A related question that I would be interested in learning is what is the standard for dimensional accuracy that Prusa (and other Reprap and 3d printer developers) use?

A test cube that I did recently using Hatchbox ABS measures 19.93 x 19.97 x 19.94, or way down in the fractional percentage point error.  This, of course, was measured with my digital caliper which I'm sure is not traceable to an actual standard.

See that is part of my problem.  30-70 microns of error is pretty good, but I did a larger part with several different features on it that I measured, and I was seeing closer to 300 microns of error.  I need to do more testing to see if I could determine a scale factor, but it would need a different scale factor in each axis.  

Posted : 02/08/2020 10:39 pm
eragon9199
(@eragon9199)
New Member

@peter-m26

The problem is two fold with that.  You could probably calculate a scale error from a test print like that, I did design a test print with several different features.  I was able to see error in different features.  Being able to use those to determine scale factors could help, but even then I would want to be able to input those scale factors in the background somewhere so I don't have to set them up for every print.  

 

More importantly if something needed to be produced, I want to be able to compensate for the part like we can with a machine tool.  I want to be able to feed values back to tell the printer on a particular part that it's scale factors are different.  I don't want to have to manually compensate every part I want to produce on the machine.  I know with industrial printers and machine tools you can feed measured values back to the machine and have the tool compensate.  I can work on a custom built solution from scratch, but I wanted to see if anyone else was working on something like this so I don't have to reinvent the wheel.

Posted : 02/08/2020 10:43 pm
jsw
 jsw
(@jsw)
Prominent Member

I would think that when we get down to the tens of microns in error, we're actually getting to the point where the diameter of the bead that is being laid down, and even the squish factor, are significant.  That kind of an error should remain consistent between smaller and larger printed items.

If you're getting .3mm error, and that error appears to increase as a function of the magnitude of any one dimension, then, yes, maybe some kind of scale factor can be calibrated.

I'm assuming that when the machine was designed, the calculations for such things as the diameter of the pulley and the fractional degrees per (micro) step of the motor were spot on, and a scale error would most likely be to component tolerances.

I know we had someone question accuracy a month or so ago here, and the reported error was far more than mine has been.

This post was modified 3 months ago 2 times by jsw
Posted : 02/08/2020 11:49 pm
peter.m26
(@peter-m26)
Honorable Member

In prusa slicer you can save for each model a project , use this every time you print, make one good cad model to adjust the print.

In prusa slicer you can adjust the xy size, but this is the whole model.

If you want to adjust a part of the model, I think you need to do this in a cad software.

Posted : 03/08/2020 8:53 am
Share:

Please Login or Register