So I'm trying to determine what an acceptable amount of error is for my

So I’m trying to determine what an acceptable amount of error is for my MG Prusa.

I print this thing:

And the “cube” comes out 20.22x20.22x10.33 (should be 20x20x10).

I feel like I should be able to do better than this, but am I being unreasonable?

Try a second one of a different size and measure it and compare. If you try to print 30mm do you get 30.33mm or do you get 30.22mm?

You also have to take the layer resolution into account. In yor are e printing with a resolution of 0.27 mm, it will be difficult to hit eg 30mm spot on.

Try printing it double-size to see if the error is additive or multiplicative. If it’s multiplicative, you probably need to adjust your steps/mm. If it’s additive, you may (see next paragraph) be extruding too much plastic.

Also try printing it with a lower layer height. I’ve long distrusted the slicers’ algorithms for calculating the spread of plastic on the outer surface of a print, and have been requesting arbitrary perimeter inset functionality to compensate for years. This error is proportional to layer height, so printing at a higher resolution should reduce the error.

Awesome, thanks guys, I’ll try scaling it up to see if the error scales, and play with layer height to see what impact that has as well…

…it’s awesome to be troubleshooting these problems again :slight_smile:

I measured .20 over printing at .3 layer height on both the 20x20 and the 40x40. I’m going to reprint at .2 LH and see if that the fix. Thanks for asking @Jason_Gullickson i’ve been meaning to get around to this for some time now.

changing the layer height worked like a charm!

Then that’s my next test @Joe_Murphy :slight_smile:

I reprinted at 200% and came out with 40.20x40.20x20.25*, which is probably within the margin of error of my ability to measure with the initial results, so I think it’s an additive issue, not multiplicative (sp? ;). I both cases my layer height was .25 (after an initial layer height of .35).

I did a couple more tests and various layer heights and I can’t say the results were conclusive. The lower heights seemed to be more accurate, but that might have just been because they were easier to measure.

So I switched from doing test parts to real parts (always seem to tell a better story :slight_smile: and there’s definitely something squishing things in the z direction. I’m tempted to use the “move the axis, measure the distance, adjust” technique to fix this but there’s a lot of advice against this on the RepRap wiki and elsewhere.

I imagine that this has something to do with the stupid SAE threaded rod I had to use in the Z drive, but I double-checked my firmware and calculations and it seem correct, but it’s hard to argue with oval shaped circles :slight_smile:

Is there a legit reason I shouldn’t try to measure-and-adjust to fix this, or is this a case where that technique makes sense, to compensate for an error that seems to be evading the purely computational method of determining the steps-per-mm for the axis?

An additive error won’t be fixed by changing your steps/mm. If you make it measure correctly at 20mm, you will have the same error in the opposite direction at 40mm and twice that at 60mm.

Either you are putting out too much plastic (which would not be affected by changing layer height), or the slicer isn’t properly calculating the spread of the outer perimeter (in which case the error would be proportional to layer height).