Thanks to @Jason_Gullickson for this excellent example of why you should NEVER use SAE threaded rods for your Z axis. Your layer height needs to be a multiple of your full-step distance, or you’ll get artifacts like this in your prints (which are easy to mistake for Z wobble). To choose an acceptable layer height for the screws you’re using, take the pitch of your screws (I recommend M6, with a pitch of 1mm) and divide by the number of full-steps per rotation on your motors (usually 200). Microsteps are not reliably accurate enough, so ignore them for this calculation (though using microstepping will still make them smoother and quieter). For my recommended M6 screws, this comes out to 5 microns. It’s 4 microns for the M5 screws used by the i3, and 6.25 microns for the M8 rods used by most other repraps. For the 10-24 screws used in this example, the pitch is 1/24 inch (1.05833333333mm), so all layer heights must be multiples of .529166666667 microns, and the repeating decimals tell you that you are really screwed (so to speak). In this case, you can get rid of the repeating decimal by multiplying by 3 to get 15.875 microns, which is not nearly as pretty as the 5 micron figure, but will let you get somewhere. If you wanted a .2mm layer height for example, with this screw, you would instead have to use 0.206375mm (39 full steps per layer), or you would see the effect demonstrated here.
Originally shared by Jason Gullickson
Solved the circular overhang problem (cooling was the answer) but even my latest changes (http://www.thingiverse.com/thing:34140) to the z-screws haven’t solved the ribbing problem. To some degree the improvement is simular to the previous change (http://www.thingiverse.com/thing:20147) I made which makes me think perhaps only part of the problem is z-wobble and I should be looking elsewhere to find the rest of the issue.
Regardless, I’m planning to try some proper leadscrews next to further reduce wobble in the Z.
@Billy Pretty sure you’re thinking of something else. I’ve seen properly-calculated layer heights immediately fix this kind of error several times on different machines.
Yes, .254mm layers should be fine, but all the other allowable layer heights have even more decimal places. The numbers aren’t pretty and are easier to get wrong.
As I’ve mentioned several times, the percent error is very small if you’re using relatively thick layers (with many Z steps per layer), but it increases as you decrease your layer height. At .25 the error will be small and will be completely concealed by Z wobble if you have any. If you fix the Z wobble and start trying to print below .1mm, you’ll see this error in your prints.
Mind you, I’ve got some prints that show this effect done on M8 rods with a .29mm layer height, but it’s not really visible except on the ones done using opaque white filament, because the hard shadows bring out the imperfections, where any dark and/or translucent filament will hide them.
I’ve seen that type of mechanical issue cause similar errors. The difference is that those, like Z wobble, have a period equal to the screw pitch, while this error doesn’t (though it may be very close). And, of course, if that was the problem, the error wouldn’t go away just by changing the layer height by a few percent to be a multiple of the full-step length (which I’ve seen happen several times).
