Layer shift? What causes this?
Since I finished building my CORE One, it seems to have layer shift issues on large or high infill prints. What could be causing this?
Example (70% gyroid infill).
Belt tensions are good, per the tuner.
You can see it had a couple small layer shifts, then the BIG one. Also, the layer before the big one, looks like it was printing on a curve...? Part was just a cube, all the previously layers were vertical/horizontal lines.
Help! Please and thank you.
Best Answer by hyiger:
- It moves completely freely. Don't now how to verify squareness other than visually, but it certainly looks square.
- Interesting. When I built the printer, both passed. Today, because of this thread, I ran it again-- X passed, Y failed. That certainly feels like a clue!
- Where is that?
- Interesting. I've been running that for load bearing parts because I read it's the strongest. Guess I'll move to other infills!
- Completely loosen the left and right tension pully screws until the belts are slack (don't need to remove the screws completely). Move the gantry to the front of the frame. Check that both sides hit the Y axis block at the same time and check for resistance. Hold one side against the frame and push the other side, is there any play? If there is, wedge something (like the universal wrench) on the side without movement and manually (and gently) slightly bend then other side until there is no more play. Next loosen the 4 screws on either side of the gantry holding it to the y-axis bearing holder. Move the gantry up and down along the y-axis several times stopping at the front, then tighten the screws. Then tighten the belts. Turn the left and right screws equally. This is important. Two turns left, two right until the belts start to gain tension then, 1 turn left, 1 turn right then when in range 1/4 turn left/right etc, then 1/8 left/right. After tensioning the belts, make sure the gantry is still parallel to the front frame.
- Y calibration failures can happen from a bent gantry (see #1), incorrect belt tension, or problems with the smooth rod bearings binding.
- Set the PrusaSlicer to "Expert Mode" then go to "Print Settings" -> "Layers and perimeters" -> Quality -> "Avoid crossing perimeters". Make sure it's off. It's generally off by default.
- Suggest testing with Grid infill to start with.
RE: Layer shift? What causes this?
Is your gantry completely square? Does the print head move freely to all corners and nothing is binding?
RE: Layer shift? What causes this?
Forgot to add. When you run X and Y calibration, do they still pass?
RE: Layer shift? What causes this?
This will be interesting, not only does it print the gyroid infill on a curve on top of a layer of straight lines, it follows the same curve consistently in that layer, that consistency throws me off. If the part lost adhesion I would expect much more of a mess, not consistent curves, if it bumped into something I would expect a shift and continued straight lines. If the part is oriented straight on the print surface my understanding is that both x and y steppers need to move go get a straight line, it is almost like a stepper is not able to keep up periodically, in any case it will be interesting to see what ends up being the root cause.
/Anders
RE: Layer shift? What causes this?
Is the "avoid crossing parameters"option set? Aside from the basics already discussed above, this seems to be the major culprit for layer shifts as it apparently generates such high speed directional changes that the printer loses it.
RE: Layer shift? What causes this?
Is the "avoid crossing parameters"option set? Aside from the basics already discussed above, this seems to be the major culprit for layer shifts as it apparently generates such high speed directional changes that the printer loses it.
But there are no perimeters to cross on a single cube, why would that logic kick in? Otherwise I agree that setting has caused me pain on more complex objects and it makes sense to turn of until fixed properly.
/Anders
RE: Layer shift? What causes this?
If you are building the perimeters from the inside out (standard behavior if I remember correctly) you would. That said, I don't know enough about the slicer to exclude any effects on a cube... I know that that's a bad param to choose 😁
RE:
"Gyroid" infill shakes the printer badly. It is rated among the worst offenders on the open-ended 3D printer Richter scale.
Is it simply the nozzle hitting the print, repeatedly?
RE: Layer shift? What causes this?
"Gyroid" infill shakes the printer badly. It is rated among the worst offenders on the open-ended 3D printer Richter scale.
Is it simply the nozzle hitting the print, repeatedly?
So far I've only seen layer shifts when my belts were grossly out-of-tune after I had just assembled the printer. And another time when I was printing PC and it severely warped causing the print head to crash into it.
RE:
yes, it's unlike the shaking alone would cause layer slip (unless something is loose, like belts as you said). My hypothesis is that the "dynamics" of the infill add up with other unwanted random mechanisms that bring the nozzle too close to solid material, such as too high extrusion factor, warping, etc.
RE: Layer shift? What causes this?
yes, it's unlike the shaking alone would cause layer slip (unless something is loose, like belts as you said). My hypothesis is that the "dynamics" of the infill add up with other unwanted random mechanisms that bring the nozzle too close to solid material, such as too high extrusion factor, warping, etc.
But those curved infill lines, several of them with the same consistent curve does not look random but rather systematic, how could a slip or collision cause that? Then maybe I am too much stuck on them being significant, I just can’t figure out what caused them.
/Anders
RE: Layer shift? What causes this?
Is your gantry completely square? Does the print head move freely to all corners and nothing is binding?
It moves completely freely. Don't now how to verify squareness other than visually, but it certainly looks square.
Forgot to add. When you run X and Y calibration, do they still pass?
Interesting. When I built the printer, both passed. Today, because of this thread, I ran it again-- X passed, Y failed. That certainly feels like a clue!
Is the "avoid crossing parameters"option set? Aside from the basics already discussed above, this seems to be the major culprit for layer shifts as it apparently generates such high speed directional changes that the printer loses it.
Where is that?
"Gyroid" infill shakes the printer badly. It is rated among the worst offenders on the open-ended 3D printer Richter scale.
Is it simply the nozzle hitting the print, repeatedly?
Interesting. I've been running that for load bearing parts because I read it's the strongest. Guess I'll move to other infills!
RE:
- It moves completely freely. Don't now how to verify squareness other than visually, but it certainly looks square.
- Interesting. When I built the printer, both passed. Today, because of this thread, I ran it again-- X passed, Y failed. That certainly feels like a clue!
- Where is that?
- Interesting. I've been running that for load bearing parts because I read it's the strongest. Guess I'll move to other infills!
- Completely loosen the left and right tension pully screws until the belts are slack (don't need to remove the screws completely). Move the gantry to the front of the frame. Check that both sides hit the Y axis block at the same time and check for resistance. Hold one side against the frame and push the other side, is there any play? If there is, wedge something (like the universal wrench) on the side without movement and manually (and gently) slightly bend then other side until there is no more play. Next loosen the 4 screws on either side of the gantry holding it to the y-axis bearing holder. Move the gantry up and down along the y-axis several times stopping at the front, then tighten the screws. Then tighten the belts. Turn the left and right screws equally. This is important. Two turns left, two right until the belts start to gain tension then, 1 turn left, 1 turn right then when in range 1/4 turn left/right etc, then 1/8 left/right. After tensioning the belts, make sure the gantry is still parallel to the front frame.
- Y calibration failures can happen from a bent gantry (see #1), incorrect belt tension, or problems with the smooth rod bearings binding.
- Set the PrusaSlicer to "Expert Mode" then go to "Print Settings" -> "Layers and perimeters" -> Quality -> "Avoid crossing perimeters". Make sure it's off. It's generally off by default.
- Suggest testing with Grid infill to start with.
RE:
^^ see above - IMHO the gantry-squaring topic is totally neglected in the build instructions. They -really- should explain the problem during the build (sheet metal parts not being "sufficiently orthogonal" for CNC machine precision requirements. The error generally canceling out as it's symmetric but then only "generally":
E.g. on my printer (built with extreme care along instructions that skipped over this topic) the gantry was several millimeters off.
Square gantry and belt tension are mandatory. Don't continue until this is rock solid. The RC 6.3.4 FW has an excellent LED strobe belt tuner. The earlier one is just hopeless, at least with my mobile phone (I guess others got more lucky). If not comfortable with RC, just update it for mechanical alignment, then flash it back to the production version, takes only a minute.
Regarding gyroid, of course, long term expectation is that it works like any other advertised feature. But with fundamental issues, I'd stay clear of any unnecessary settings away from the trodden path. Establish a working baseline, then debug one small step at a time.
RE:
- It moves completely freely. Don't now how to verify squareness other than visually, but it certainly looks square.
- Interesting. When I built the printer, both passed. Today, because of this thread, I ran it again-- X passed, Y failed. That certainly feels like a clue!
- Where is that?
- Interesting. I've been running that for load bearing parts because I read it's the strongest. Guess I'll move to other infills!
- Completely loosen the left and right tension pully screws until the belts are slack (don't need to remove the screws completely). Move the gantry to the front of the frame. Check that both sides hit the Y axis block at the same time and check for resistance. Hold one side against the frame and push the other side, is there any play? If there is, wedge something (like the universal wrench) on the side without movement and manually (and gently) slightly bend then other side until there is no more play. Next loosen the 4 screws on either side of the gantry holding it to the y-axis bearing holder. Move the gantry up and down along the y-axis several times stopping at the front, then tighten the screws. Then tighten the belts. Turn the left and right screws equally. This is important. Two turns left, two right until the belts start to gain tension then, 1 turn left, 1 turn right then when in range 1/4 turn left/right etc, then 1/8 left/right. After tensioning the belts, make sure the gantry is still parallel to the front frame.
- Y calibration failures can happen from a bent gantry (see #1), incorrect belt tension, or problems with the smooth rod bearings binding.
- Set the PrusaSlicer to "Expert Mode" then go to "Print Settings" -> "Layers and perimeters" -> Quality -> "Avoid crossing perimeters". Make sure it's off. It's generally off by default.
- Suggest testing with Grid infill to start with.
Did all of the above, and it seems to be holding now.
Also, test passed now!
Thanks!