Article on Z axis artifacts

spapadim wrote:

Page 29 of "Taxonomy of Z axis artifacts". Z wobble is #1, and article has a nice overview of common causes. Wish I had seen this earlier.

Sadly, it is all wrong. This article has been brought up numerous times and IMHO completely debunked.  If you search the forum, you'll see that the effects it describes don't match the banding on the solidoodle (wrong frequency and amplitude), and printing with 1/180" layer height doesn't make all the banding go away either, which it should if banding were caused by microstepping roundoff.

Hm, should have searched first... But isn't what you're referring to the "Z ribbing" part , as he calls it (i.e., rounding error -- which I'm not so sure about its importance either)? The "z wobble" is different, he clearly states that z ribbing has period different from thread pitch (and is more relevant in high resolutions), but z wobble has period equal to pitch, and the primary cause of z wobble is basically an over-constrained z axis which makes platform deflect in a circular trajectory on x-y, esp. when rod is bent. The particular causes may be different on SD (guessing lack of flexible shaft couplers rather than too many rods or axis constrained on both ends and/or with bearings), but general ideas seem right to me..

spapadim wrote:

Hm, should have searched first... But isn't what you're referring to the "Z ribbing" part , as he calls it (i.e., rounding error -- which I'm not so sure about its importance either)? The "z wobble" is different, he clearly states that z ribbing has period different from thread pitch (and is more relevant in high resolutions), but z wobble has period equal to pitch, and the primary cause of z wobble is basically an over-constrained z axis which makes platform deflect in a circular trajectory on x-y, esp. when rod is bent. The particular causes may be different on SD (guessing lack of flexible shaft couplers rather than too many rods or axis constrained on both ends and/or with bearings), but general ideas seem right to me..

Yes, but if you look at the banding on a part printed on the solidoodle, the banding matches the pitch, but is symmetrical around the center of the part, meaning that a ridge "out" on one side of the print is matched (mirrored) by another ridge "out" on the other side of the print, not a valley "in".  This matches the description of what the article describes as "z-ribbing", but with the frequency of what it describes as "z-wobble".  If we were getting the circular "z-wobble" described in the article, we'd expect layers to be offset like a stack of identical (but irregularly centered in a helical pattern) coins, rather what we see, which is a perfectly centered stack of coins of periodically varying size.

Instead, everything to me seems to point to a nut whose looseness makes its position sensitive to any lateral force due to an imperfect, bent or off-center rod.  This would give the needed combination of varying layer heights at the thread pitch frequency, also explain why adding anti-backlash devices, flexible couplings, and thinner, finer z-rods alleviate banding.

solidoodlesupport wrote:

The Solidoodle's banding is caused by a variety of issues with several physical mechanisms on the machine. We're working on a fix, but it's definitely not a software/firmware issue.

What types of fixes have you tried or are you looking to try?

But your products are starting to not be so "low cost"  there are several printers with as big or bigger print area starting to hit the market that are less expensive then the SD3.

robo3d for one.  12-16week delivery.  Though they should have started delivering the first group of printers now.
Rigidbot

and i did say starting.