Z Axis - precision and accuracy

The reason your couplers only move backwards/forwards sometimes is that the friction is high and enough to usually stop the couplers moving when the power is turned on. However the microstepping error is still there because every time the power is turned off the A4988 drivers (and Marlin) forget where they were.

 

As I mentioned in my earlier post it would be very helpful for someone (or a few Robo'ers) who is NOT having Z Ribbing issues could print the test.stl I attached above in PLA at 0.2mm layer height and post a picture. The test model only takes about 15mins to print.

This should give us some clues as to why 20% of Robo'ers are not having problems with Z Artifacts.

 

Here ya go.
I have never had this problem as I indicated earlier. Perhaps all of the set-screw tightening/loctiting and other tweaks we did when the machine arrived helps, or not.

 

N

No. That is simply layers. The sides are glass smooth.
It is PLA and the layers are very visible, but there is zero deformations on the sides. No ribbing at all. Since it is translucent you can see the layers better than a solid color (that is all internal). The first shot is at an angle to the light so you can see how smooth it is. It is that way all the way.

 

Interesting.
Kickstarter backer myself, one of the first October batches. Mind you, it had LOTS of QA issues (if there was a screw installed that was tight originally, I could not tell you which one it was).

 

Thanks Mark,

I am aiming to print Mark's gcode on my Robo to prove one way or the other whether there is as issue in the gcode or my machine.

Mike and Axislab

I created a new thread Z Axis Artifacts - Known Issues and Fixes.

I'm hoping we can keep this thread up to date and edit/delete stuff as we get a better understanding of what works and doesn't.

 

No you are right. The Z axis power off issue won't be the only problem causing Z artifacts of course.

I suggest you print the test.stl in PLA at 0.2mm and take a careful look at exactly what is happening. First question - are the artifacts Z Ribbing or Z Wobble?

I deliberately made the test.stl model round so it would be possible to see if the ribbing was on one axis or uniform right round the model.

 

Montravont

I suggest that you also check your threaded rods are seated properly in the couplers. Easiest way to check is to lightly push down on the threaded rod and watch the coupler. If the coupler compresses at all, the threaded rod is not seated properly.

 

While the Z Stepper disable is one cause of varying layer height = Z Ribbing, it's obvious that it is not the only cause. For example I still have some residual Z Ribbing in my prints. It's a lot better but the problem is not completely solved.

So my next step is to focus on

- Is there some difference in the gcode between my test prints and test prints which don't have any Z Ribbing? I now have known good gcode from Mark Tomlinson so will try to prove that one way or the other.

- Maybe the Z Steppers are not actually turning the correct number of microsteps because of friction or lack of power or maybe firmware?

- Maybe there is an issue with the M8 threaded rods I am using? Probably unlikely as the Z Ribbing issue is so widespread.

- Maybe the helical couplers are winding up and releasing in a pattern due to friction?

Frankly I am so pi$$ed off by the Z Ribbing I'm willing to spend the time to get rid of it. Would be good to get some support from the Robo tech guys as well.

 

Just an update

I ran Mark Tomlinson's known good test gcode but could still see the residual Z ribbing in my prints. So I can't blame Slic3r for the ribbing.

I think I now know how the residual Z ribbing is happening but not yet why.

I attached a very light protractor on the top end of the threaded rod, ran my test print and video'ed each turn of the Z screw against the protractor.



Note I changed the Z height to 0.125mm for this test because I wanted exactly 10 increments per turn (360 degrees) of the M8 rod 1.25mm pitch. Normally you would expect to see 36 degrees for each increment.

The chart shows a plot of the incremental degrees of each turn of the Z Screw during the print.




So I think it pretty clear that the Z Screw incremental degree variation is responsible for the residual Z ribbing. As you can see, the plot shows 6 waves in 6 revolutions of the Z Screw - exactly what the print Z artifacts look like on the test model.

The chart also shows that the forward/backward jumps due to the Z Stepper being disabled between moves have been eliminated by always keeping the Z Stepper powered up.

This really helps narrow down where to look for the cause. My protractor measurement setup will also give me a way to check if I have found and fixed the cause.