Huge Layershifts / Missed Movements

[luke321]

Hello,

I have a very annoying problem with my Makergear M2 at work, I will try to list everything i noticed.

http://i.imgur.com/MZGDT16.jpg

The shifts happen in both x and y direction at random, often the same part comes out completely fine when printed a second time but it happen 1 out of 8 times sometimes more often.
The shifts are not Gcode related, I did not slice the model again when retrying a print.
I used Slic3r for all of the prints with the exact same settings I use on my second M2 which never has that problem.
It happens with flat parts as well as tall parts like in the Picture, and I normally use Z-Lift so the nozzle will not touch the part.
I had it happen from starting at 4mm up to 50 or more mm in z height, sometime shifts occured in both axis.
In the picture the printer should have moved down to continue the infill, but it seems that after about 5mm in y direction the y stepper stopped.
I heard no strange noise from the printer.
I reduced the speed significantly but still the shifts occured, while my M2 at home with very high speeds only had the problem once in 3 years.
I unplugged all the stepper connectors multiple times so it should be now loose connection.
It's the 12V version of the M2.
No loose pulleys, everything moves smoothly.


I would appreciate your thoughts!

[willnewton]

Gonna just start with a wild guess that it may not directly be the printer if you are using Simplify3d. Judging from the crazy bridging shown in your picture, that is where I would start.

Check your model in S3D in the layer by layer preview window. Is it clean as designed or has S3d inserted odd bridging? Choose to view by print features and bridges show up as yellow. It will be obvious as you scroll through the layers that S3D is getting funky.

S3d has a bug that will generate some whacked out g-code at certain layer heights/features.

Try printing at .199mm layer height if you are using .200mm heights or maybe change your first layer height from 100% to 99%. This small change is usually enough to solve the problem if S3D is the cause.

[jsc]

Probably a slicing error. Possibly the bug Will mentioned, but maybe just a bad model. Try feeding it through http://netfabb.azurewebsites.net

[luke321]

no bridging involveld, what you see standing over the part should actually be the infill.
the model is perfectly fine, I used slic3r, the gcode preview shows no errors.

http://i.imgur.com/MZGDT16.jpg

I updated the list and image for further clarification on the error.

[willnewton]

Have you run the model through Netfabb? https://modelrepair.azurewebsites.net. You may have a leak in the model.

Is it possible to share the file so someone else can take a look at it?

Is it just this model or other models too?

[Jules]

I've seen birds nests like that happen when the nozzle gets whacked out of alignment by a warped spot on the print and just continues printing from there as if it's still lined up. (Although, I've got to say, that's the most interesting horizontal bird's nest I've ever seen. )

I don't know that is what happened there....you might want to check the teeth on the gears in the motors....they might be getting worn after this much time. (Sorry, the name of that specific gear escapes me at the moment...pinion? Start with checking the motor that drives the print head on the X-axis gantry.)

Correct me if I'm wrong here guys.

Note: Found the thread on it:

viewtopic.php?f=7&t=2544&start=10

[luke321]

It happens on all kind of models, even cubes with no overhangs whatsoever, no warping or anything the nozzle could crash into.
Models created with CAD software mostly but also things from thingiverse.
I added fans to the motors and a larger fan to the drivers because I thought it was overheating but the problem remains.

[insta]

If you haven't yet, re-upload the firmware.

The next phase is replacing RAMBo. You've got an actual mechanical / electrical problem. That's not software.

[PcS]

I second maybe Rambo issue. although if you have an lcd on the machine unhook it and test without it.

[jimc]

i think it was determined a while back by josh and jsc that the layer shifting was a problem with marlin. if i remember right that was corrected but as of now i dont know which firmware has the correction applied to it. the pic from the op shows something beyond layer shifting though. thats something different there. a long time ago i went through the whole machine trying to figure it out. i changed out the rambo, harnesses, stepper. none of that worked. what did work was dropping the acceleration a little and changing the xy movement speed in s3d from 18000 down to 12000.