Have done some extensive tests but now I am certain this is mathematical.
Can’t say exactly when this started, Was it with last firmware upgrade to 1.12? Quite possibly. Now on 1.12.1
I have done extrusion calibration, played with temps and different filament, even changed the hot head, levelled to 11 X 11 but now settled on 7 X 7 to stop the bed cooling too much, it’s not temp, extrusion, or levelling. In have played extensively with my Z height and temp together to get the best layers.
I am getting a weird pattern which is only on the 4th layer of print, every print. Next layers up make it worse, but it is not there on the first 3 layers. See pics.! FB_IMG_1619621135266|281x500
Anyone have any ideas, I am now convinced this is a firmware issue but do not know how to test for that. Can I roll back to see if it is still there in earlier version? If so what is the process?
When I get that pattern, it’s caused by over-extrustion on the first layer (because the print head is too close to the bed). Extruding too much filament in too small of a volume causes the filament to push previous material up off the bed, causing the ripples. Additional layers will slowly cover it, and it’s usually hidden by layer 5 or 6.
But if your layers 1-3 are fine, and it only happens on layer 4, I think your problem would have a different root cause. I suspect it’s the same problem (over extrusion), but I’m not sure why layer 4 and only layer 4 would experience it.
I should add, it has not always find it, started I think with the firmware upgrade to 1.12 but can’t be sure. You don’t see it on small pieces or something that is filled. Only flat items like the bottom of a box. I have pictures of larger items that I did earlier and they are all fine.
You can do a bit of math to figure that out. The first layer is printed at Initial Layer Height thickness. If you subtract that from the Bottom Thickness, then divide by the Layer Height, you’ll know which layer is the last layer of the bottom.
For example, Fast Print defaults to Initial Layer Height 0.3mm, Layer Height 0.2mm, and Bottom Thickness 0.6mm. So the first layer is 0.3mm high, then layer 2 0.3-0.5mm, layer 3 0.5-.7mm, and layer 4 is no longer part of the bottom. Normal works out the same, even if the exact heights are slightly different. But if you use High Quality, Layer 5 is going to the first that isn’t part of the bottom. You can play with the bottom thickness to test to see if changing that changes the layer affected.
i have the same problem and it started to appear after the new firmware release (not 100%, but all prints afterwards show this pattern). it always starts at a certain height, regardless of the slicer in use. looking at the changelog on github i found the FPU commit (could theoretically cause differences in calculations) and a fix regarding the first leveling point (#96 if i remember correctly).
observing the problem for hours on my printer, it even looked as if the z-axis was slightly moving upwards to cause these lines that dont attach properly. might be „fumes induced“ after watching for such a long time … we used to utilize high speed cams (investigating casino aka gambling machines … the legislator did so too) … maybe thats an idea to check?
You may bend the print sheet when you remove the print, which makes the flexible pad is not even. You can rotate the print sheet 90 degrees and then print the same model in the same position.
I flipped the sheet, have not tried turning but will. Might even try to print on glass to compare.
This has happened with 4 or 5 different models, all different sizes. Only happens of course if there is a large flat areas that has no infill, if you have infill or depth then you don’t see it as it gets buried.
It may have been an issue longer than I realise as these are the first items that are thin based.
As I am not the only person seeing this, it would suggest a wider issue.
My test piece is a 150 X 150 X 2mm square as you can see from the picture.
What confused me is that the first 3 layers are fine, it happens from layer 4 and I can’t see why.
I have a few more tests as requested. I started with some narrow strips 150 X 20. I did a 5mm a 1mm and a1.4 mm. These are all perfect smooth really good.
You can see the place the thinner steps have been in the middle which is perfect but either side although the majority is smooth there are rough patches. These will cover in the next 3 layers but then on layer 4 the crazy pattern returns.
I note you said something really interesting. Yes I do flex the the print sheet to ease removal of the objects stuck on them. Are you saying doing so is causing an issue. Should we not be flexing this sheet? But again strange we get the same pattern on different objects turned around.
Have I ruined this print sheet or will it settle back? I will test with a glass sheet next, needs some messing about to get the Z right.
I have done more tests today and it is without doubt layer 4.
First three in latest test were really smooth, had a weird rough patch on one small area on layer 1 and different similar patch on layer 2 with some vibration noise from the head that I can’t narrow down… Then layer 3 was really smooth and layer 4 has the same pattern again.
That last test…does it have the same pattern if you print it twice?
Also, what happens if you turn around the sheet (so it prints on the other side), recalibrate and print again?
Also, a question for all the people that uploaded the surface ‘matrix’ after doing the auto calibration, and even made a graph out of it.
How many Z steps would require to print a straight line from one side of the plate to the other? Could these required z steps be the cause of some visible patterns? I think yes, but doesn’t explain why only see it on layer 4.
And a final idea. A great idea for Luban if i may say it…
Can we do (ok, I KNOW it is possible, and would be a punch in the face for other slicers) a special raft that will compensate the calibration matrix, so after the raft the printer will not have to move Z at all during each layer ?? (if z-hop disabled of course) @brent113 what do you think about this?
And this raises a questionfor me… If you make a cube-like print, will the last layer show the imperfection that the matrix was meant to overcome in the first layer? or will the firmware reduce the corrections over each layer? (another good idea maybe?)
In reply I have done the test 4 times with this particular item (lattice storage box) and used both sides. Each time turned. I have run the test with 2 other objects (shelf 170 x 120 with a lip) and a test square (150x150 X 2mm), all with exact same weird pattern.
The pattern turns so nothing to do with the bed. Also did a test strip 150 X 20 just down the middle both side to side and top to bottom, and both printed beautifully.
Did 2 xyz test cubes they are small but fine.
This started after 1.12 firmware (I think). I have a 120x120x5mm test square I did two weeks ago that it fine, use it as a small cutting board.
I am still really confused.
Others on here and on FB thread report the same, even using a glass bed.
That is what all rafts do. I have posted before about why rafts are the way they are. Specifically to your question, it’s the base layer spacing that give it the flexibility to conform to the surface of the build plate by giving space for extra filament to flow into as a raised area on the build plate displaces filament.
That image is excerpted from this more detailed Cura support article: Support Community
Thanks a lot for the detailed reply (I see you get this response a lot .D)
Ok then, since the raft has a smart design, we don’t have to mess with mesh data…
Does this means that after printing the raft, the printer will ignore mesh data and not touch the z-axis during the same layer?
This could be a good reason to print a raft…and I don’t recall ever reading about this on Cura’s pop-up help.