You guys ever see z wobble this bad? A friend just brought these prints to me.
It’s so regular and defined its almost a feature not an issue lol. He’s using m8 threaded rod in a Lulzbot Taz clone.
Не хорошо, можно и лучший результат получать 
Yes, I saw the same issue on the lulzbot mini (pre release, the versions sold don’t have that issue) because they were using imperial rods and not metric. Did they just replace the rods recently? Has it always printed like this?
That looks more like a badly-bent screw than an imperial/metric step size rounding issue.
Yeah, but on the taz, the Z axis smooth rods are the constraints not the threaded rod. It would have to be bent enough to move the smooth rods a bit too. Although, imperial rods aren’t the only thing that causes this issue. Having z steps calibrated wrongly can do it.
Some clone companies really are dumb enough to add constraint to the top of the screw, but yeah, I see what you mean.
@Jay_Babin are you constrained at the top?
Nope… when we built these clones it was already known not to constrain the tops with bearings… they are free
I believe it’s @Whosa_whatsis who showed why imperial screws can have that effect.
It’s so regular, I wonder if it’s not incorrect steps/mm on the z-axis or poorly chosen layer heights leading to cumulative error.
…or a very bent screw, I suppose.
Either way, that’s a real problem. Might want to fix that.
bent lead screw so it wobbled the bed as it drops or sticky beatings on the Z axiz doing the same. quite easy to see if you raise or lower the bed. unless those are the layers in which case it over extrusion.
@Mark_Rehorst Properly set layer heights can minimize the cumulative error effect of imperial leadscrews driving metric distances to the point where it’s effectively invisible. But improperly set layer heights (ie layer heights in even mm sizes and not in even step-turns of the screw pitch) will cause something very much like this.
I’ve demonstrated it to my own satisfaction on my own imperial leadscrew z-axis printer.
I’d be looking at layer heights too. Another thing that can cause this kind of banding that should be considered is overextrusion leading to a periodic overflow (as mentioned by @Michael_Scholtz ).
Diagnosis can be methodical, so opinions don’t have to enter into it. If the banding wraps around the part in a helical pattern in the same pitch as the screw, it’s Z axis wobble caused by bent screws and insufficiently rigid guidance. If the banding is the same pitch as the screws, but not helical, layer heights are wrong, whether due to imprecise inch-mm conversion or some other mathematical error. If the banding is non-helical, it may or may not be overextrusion. Of course, all three can be present at once.
Banding this severe won’t be exclusively Z steps/mm unless it’s so wrong as to be imitating overextrusion. That would show as short parts.
The screws don’t have to be imperial to cause this. Some of the worst banding I’ve seen was on my Cupcake (already modified to fix its inherent Z wobble problems), and it was caused by a poorly-chosen layer height. Switching to a value that was a multiple of the full-step length for its M8 screws made the problem go away.
Even so, unless you’re not using any microstepping or have really poorly-spec’d motors, it’s unlikely given the amplitude of this error that it is caused entirely by that.
What type of couplers are you using to attach the screws to the motors? Using the wrong kind can sometimes be just as bad as constraining both ends of the screw. You might also consider a move to M6 or M5 screws, which will have less rigidity relative to that of your linear guide rods. Of course, if your bearings are loose, all bets are off.
@Mark_Rehorst Microstepping is also a contributor. If your motor/driver combination gives you very accurate microstepping, the effect will be smaller, but machines should not be designed to depend on that. There are also many other types of error that can swamp this one, and it’s sometimes so subtle that you won’t notice it unless you put it right next to another print without the error. Other times, it looks like the picture here. You may not notice it on your machine, but it has been thoroughly demonstrated on lots of other machines. I don’t warn against imperial screws because they will definitely always produce banding all the time on any machine that uses them, I warn against them because they are one of many potential sources of error that are better avoided.
Based on what you list @Mark_Rehorst , assuming you’re using 1.8 degree steppers and 1/16th microstepping, you would get a full number of steps per layer at .2495mm rather than .25mm. That makes each layer a full 55 steps, rather than 55.1268-ish at .25mm layers.
If you use .25mm, you accumulate .23mm of error over the course of a 10cm distance. Which isn’t that bad, at least relative to other printers, like the printer with an imperial leadscrew that I’ve experimented on. It has a .94mm error over the same distance.
But I agree that the size of the banding on this print probably means something more than just bad layer-height values is going on. That may be contributing, and after the bent screw or stuck bearing or whatever else is wrong here is fixed any remaining banding may be down to the math of the situation.
As with most print quality problems on home 3D printers, there’s probably more than just one error at work.
1/2 inch pitch = 1/2 inch table travel per rev. Talk about high speed z movement…
@Stephen_Baird This type of error does not accumulate. The Prusa Calculator is wrong on this fact, and I’ve tried to explain it to @Josef_Prusa repeatedly without success. This is non-accumulating error, meaning that rather than causing the error to build up over the course of 10cm, you’ll just get a rounding error at some regular interval that will return your Z height to the expected value, getting rid of all the error that has accumulated since the last rounding error. When the rounding error occurs, you get one layer that is too thick or too thin, and thus over/under-compressed, which is where the banding comes from. The bigger the difference in height between this layer and the surrounding ones, the more over/under-compression you have, and the more pronounced the banding will be. Large numbers of steps per layer make the single-step rounding error smaller relative to the height of the layer, so the banding is less pronounced.
Thanks guys for all the great info. It gives me a good path to go down when I get home.
SO, A Little background info - This machine is a Taz Clone. I have not finished calibration and have only printed about 7 objects so far. I have never gotten a great part but I am getting close.
Today is the 1st time I have seen this banding.
I did just transport it back and forth to MidWestMakerFest.com
While at the show I printed a really nice part - I do not see this kind of effect on that part.
This was the 1st print since then. I did re-level the bed.
I am using the same Z-Height as I had on the 1 good print that I had. Its a different material but not sure if that changed variable could have anything to do with this.
I just grabbed the parts and ran this morning so I have not been able to check to see if there was visible Z-Wobble happening. Also I will check for a bent threaded rod.
It’s definitely a bent z screw! The x axis was probably nudged/ hit pretty hard when loading it up after MWMF. It wasn’t printing like this at the event.