Layer Shifting

[nitewatchman]

I need help from the Brain Trust.

My RMv2 has started throwing layer shifts into prints at random. The average shift is about 0.4mm and is always in on the Z-Tower, always shifted away from the tower and is always perpendicular to the front of the printer. I can hear your thinking from here, this one is easy since it is so repeatable. So here is what has be checked or substituted to date:

1) Pinter Bed is shifting? Checked bed screws were loose, bed could be shifted. Tightened screws, bed cannot be moved relative to the machine base - no effect, layer shifts still present.

2) Dirty pins in electrical connector? Reseated Z-Column Cable Connectors. Layer Still Shifts.

3) Arm Pivots hanging up? Just changed to SeeMe Ball and Socket Arms, can't find rough spots - layer shifts still present.

4) Bowden Tube and Harness hanging up on structure or printer? Inspected, no hangups - layer still shifts.

5) Belt Tension? Inspected and adjusted - layer still shifts.

6) Damaged teeth on drive belt? Inspected, good teeth - layer still shifts.

7) Looses grub screw on drive wheel? Inspected, tight and retightened - layer still shifts.

8) Flakily -Z- Column motor? Exchange Z-Column Stepper Motor with spare. Layer still Shifts.

9) Bad Rambo Stepper Driver? Moved Z-Column Motor to second -Z- Stepper Driver on Rambo. Layer still shifts.

Next step is to move -X- Column commands to -Y- Column, -Y- to -Z- and -Z- to -X- to see if layer shift follows commands. Also looking at increasing motor current to stepper and lowering acceleration and jerk to axis. Can't think of any other areas to poke around in.

Conspiracy Theories anyone?

gary

[mhackney]

Does it shift back? Or does it stay shifted. And if it stays shifted, can you tell if it does it more than once? And the BIG question, how far does it shift when it shifts?

If it's always in line with the Z tower it has to be something in the Z movement chain. Are you double triple quadruple sure of 7? It could be that it is "tight" but not on the stepper shaft flat.

[nitewatchman]

Nope, it never shifts back. Just keeps going to the front. There will perhaps be a shift after the first 10mm height, prints fine for an additional 10mm, shifts, then another section of a few mm and another shift. Looks like a stair case up the back side of the print. But it never shifts back the other way.

Checked #7 thoroughly, this is what I initially expected to find. Since the screw is hard to get to compared to removing the motor (allen button head screws and long ball end wrench), I removed the motor and checked. Even snapped the end off my allen wrench tightening the screws. One screw seated and rocked out on the flat the other TIGHT on the shaft, a drop of extra thin CCA on the end of the shaft. Replaced motor, got the same shift within 5 or 6mm of height.

This is a new and unseen problem that has been manifest in the last 18 months.

gary

[mhackney]

Now you have my curiosity!

Does it happen only with the nozzle between the bed center and Z tower? Between center and front of machine? at center? or basically anywhere?

[Xenocrates]

A note, there are not 2 Z drivers, the Z driver has a split pinout. there are 3 axial drivers, and 2 extruder drivers, with 6 outputs. Ain't that grand?
A potential temporary fix if it really is the Z driver is to connect it to EX1 (the second extruder port, the first is EX0), and tell the firmware the pins for EX1 are Z instead.

Have you tried changing or cleaning the Z connector? Not just reseating it, but replacing or lightly polishing it?

Did you adjust your firmware prior to the problem occuring? Have you checked it since that? It's unlikely to be the source of the problem since it moves at all.

[nitewatchman]

Now you have my curiosity!

Does it happen only with the nozzle between the bed center and Z tower? Between center and front of machine? at center? or basically anywhere?

Got me stumped also.

Most prints lately have been small and centered on the bed. Once the shift has been made, the part appears as if it has been cut, shifted and glued back together.

Doesn't matter if I slice with S3d, Slic3r or Cura. Also doesn't matter if the layer height is .075, .1, .15, .2 or .25. I have decreased the Jerk and Acceleration, I did rework the Hot End recently, perhaps the mass was increased, I have always run these values rather high. HArd to believe however that a small change can make that much difference and only on one axis.


gary

[nitewatchman]

A note, there are not 2 Z drivers, the Z driver has a split pinout. there are 3 axial drivers, and 2 extruder drivers, with 6 outputs. Ain't that grand?
A potential temporary fix if it really is the Z driver is to connect it to EX1 (the second extruder port, the first is EX0), and tell the firmware the pins for EX1 are Z instead.

Have you tried changing or cleaning the Z connector? Not just reseating it, but replacing or lightly polishing it?

Did you adjust your firmware prior to the problem occuring? Have you checked it since that? It's unlikely to be the source of the problem since it moves at all.

Your correct only five chips. Swapping to the Extruder type needs to go on the list also although my experience is that the drivers either work or don't - not much middle of the road.

Reseated and then replaced the connector on the new motor.

Only firmware change was the Thermistor Type. I had been running an M3 Screw-in Thermistor in my E3Dv6 and the epoxy on these keeps cooking out at high temps. Went back to the original E3Dv6 screw retained thermistor (guess there was a good reason they did it that way to start with) and had to change the type. Other than that no changes in a year.

gary

[ccavanaugh]

I had this same issue once on a cartesian machine. It turned out to EMI on a limit switch. I had to disable limit detection when not homing.

[mhackney]

New thought - tooth skipping on the belt. This can be hard to see by inspecting the belt. But if the skip is consistent ~2mm or so increments, it might be the belt. You could swap the Z belt with one of the other and/or turn it end for end to utilize the other end of the belt the doesn't see the pulley.

[nitewatchman]

New thought - tooth skipping on the belt. This can be hard to see by inspecting the belt. But if the skip is consistent ~2mm or so increments, it might be the belt. You could swap the Z belt with one of the other and/or turn it end for end to utilize the other end of the belt the doesn't see the pulley.

Starting to think that it might be something like that or a bearing. Increasing the current and slowing the rapid seems to have lessened the problem but the -Z- Axis motor (new motor is still in place) is running warmer than the other two. Seems like extra drag might be in there somewhere. The skate (new SeeMe Molded Skates) seems to be free with no tight spots or roughness but it is tough to tell with it in circuit. When I changed the motor out however , it did free fall when the belt was released.

I think tomorrow I will tear down the entire the entire column and swap out the components with spares and invert the belt.

gary

[nitewatchman]

I had this same issue once on a cartesian machine. It turned out to EMI on a limit switch. I had to disable limit detection when not homing.

Interesting, may try rerouting the wiring to this limit. It is the longest and does pass the Power Supply, 24VDC to 12VDC Buck/Boost and the SSR.

gary

[bvandiepenbos]

I had this same issue once on a cartesian machine. It turned out to EMI on a limit switch. I had to disable limit detection when not homing.

Interesting, may try rerouting the wiring to this limit. It is the longest and does pass the Power Supply, 24VDC to 12VDC Buck/Boost and the SSR.

gary

I once built a RMAX and happened to run the extruder motor wires inside the column with the endstop wires. That machine had weird random layer shifts. Re-routed wires and problem went away.

[nitewatchman]

I had this same issue once on a cartesian machine. It turned out to EMI on a limit switch. I had to disable limit detection when not homing.

Interesting, may try rerouting the wiring to this limit. It is the longest and does pass the Power Supply, 24VDC to 12VDC Buck/Boost and the SSR.

gary

I once built a RMAX and happened to run the extruder motor wires inside the column with the endstop wires. That machine had weird random layer shifts. Re-routed wires and problem went away.

This is stranger though, been running 18 months without an issue like this.

[ccavanaugh]

Interesting, may try rerouting the wiring to this limit. It is the longest and does pass the Power Supply, 24VDC to 12VDC Buck/Boost and the SSR.

gary

I once built a RMAX and happened to run the extruder motor wires inside the column with the endstop wires. That machine had weird random layer shifts. Re-routed wires and problem went away.

This is stranger though, been running 18 months without an issue like this.

EMI does that... I see it regularly in fluid metrology applications. A machine runs fine until an operator literally lays their hand on a machine frame and it wigs. Ground loops tend to be the culprit.

Better yet, when my garage door opener light bulb goes off, my car locks its doors... I guess it's afraid of the dark.

[nitewatchman]

Okay, while this will likely jinx the printer from this point forward, the layer shift may have come under control. Just finished a 12 hour print with no shifts. This along with a hand full of shorter duration prints is encouraging. The fix doesn't seem reasonable but this or something jogged by the fix seems to have been the issue.

Oh, what was the fix. Removed the -Z- axis belt and turned it over. And before the wiseass comments, yes the teeth were facing the gear before. I reversed the belt end for end bringing the end normally above the skate and not used to drive the effector into service. While the belt was off I went over it in detail looking for damage going so far as bending it backwards to flex the teeth open on the outside of the bend. Did not find anything other that polished surfaces where the teeth ran on the gear. Put it back on, tensioned and the problem retreated into the dark cave of printing monsters and demons.

I know, not logical, not reasonable or believable but there it is. I don't understand what just happened.

However as I have said before, being a simple man, I don't understand how Color Television works but that doesn't keep me from watching it.

[mhackney]

It could be that the teeth on the belt in that area were "soft" from running over the pulley and flipping the belt eliminated that. It might have only been a couple of teeth on the belt. It's a hard thing to inspect and find. But at least you might have fixed it.

[mhackney]

Now I have an unexplained layer shifting problem! All was well until today when I attempted to print this QR extruder mount. I let the print finish so I might learn what was going on. Checked everything and started another print. Got exactly the same shift at exactly the same layer. In my case, the shift is towards +X and -Y. Close up it looks like there are several very small translations over 6 layers - a stair step effect - rather than a single shift. Very odd.

[mhackney]

Forgot to add - when I say shift in X and Y I mean Cartesian X and Y! Not towards the X or Y towers. That's what's perplexing.

I'm doing another print at 50% speed to see if it happens again and maybe be able to observe something odd.

[nitewatchman]

Now I have an unexplained layer shifting problem! All was well until today when I attempted to print this QR extruder mount. I let the print finish so I might learn what was going on. Checked everything and started another print. Got exactly the same shift at exactly the same layer. In my case, the shift is towards +X and -Y. Close up it looks like there are several very small translations over 6 layers - a stair step effect - rather than a single shift. Very odd.

The attachment FullSizeRender.jpg is no longer available

Gypsy Majek!

You need to squeeze a ripe tobacco worm over each stepper motor while saying "There's No Place Like Home" backwards.

My layer shift (it was also a stair step type flaw) hasn't returned, keeping a spare worm just in case!

gary

[teoman]

And to whom will the shifting problem migrate to??

[teoman]

Aaa shiiit.

[mhackney]

No!

[mhackney]

Well the 50% slowdown print turned out perfect. So now I'll try 75%. The original is my normal speed so there is something odd going on.

[nitewatchman]

Well the 50% slowdown print turned out perfect. So now I'll try 75%. The original is my normal speed so there is something odd going on.

Is this happening on a file that you have printed before?

gary

[bvandiepenbos]

You might try changing position of parts on table and re-slice. Could be a certain move(s) causing enough jerk in direction change to lose steps?

I have had my engraving machine lose steps and shift position only on a certain custom logo. Fix was I went back and adjusted a few nodes in my digitized graphic, then problem went away. Vector path had a sharp direction change in one spot that made stepper lose steps.