If i have model that is .2mm high on the z and my initial layer height is .2mm why does:
a) the skirt print at .2mm high
b) the model print at .4mm high
doing this the model only has perimeters, solid infill and external perimeter
I am using prusa
ironing is off which adds a layer in the slicer)
I am using a true micrometer to measure (not calipers)
is this one of those conventions like where 2x4 wood is not actually 2x4 inches?
am i doing something wrong?
yes i know this somewhat academic as my 20mm and my 40mm xyz cubes are pretty darn perfect, i am just one of those folks that like to understand how things work and why from first principles.
What height is the gcode calling for. Is this a machine or slicer problem.
Side note, micrometers put a lot of pressure on the filament and in my experience is about as accurate as a caliper since it’s easy to press very hard. My mitutoyo ratchet is too tight for plastics. Good repeatability though, just no idea what the true thickness is.
I was measuring the change in filament diameter with moisture content, it was great for relative changes like that.
Metrology is such a pain like that. What IS diameter even
there is no change in z height until we get to gcode end
not only is the object 0.4m high, extrusion is shit, here are many i have printed (not all in the same location) - i have tried changing just about everything (nozzle temp from 180 to 240, bed temps, recalibration, lowering z offset, trying 0.4mm size object and 0.4mm extrusion height.
Note this is absolutely related to the xy coords of the object on the bed and is IDENTICAL on a glass surface, the snapmaker stock surface & and an energetic PEI smooth plate too. aka this thread (thanks to @Artezio for all his help, and just plain listening to my crazy!) io should note the 'good squares in the 5x5 matrix are ~0.22mm - aka what I would expect.
i cut the model to be just the top center square to speed up testing iterations.
happy to rehash - i have calibrated many many times, this same issue occured BEFORE the y rails were replaced too.
one calibration i went close on the pinch feel than i like - that just made the issue worse
the per-print z offset, i tried several changes, no better
from a z rail perspective i don’t understand how the z height can be ok for one row of sqaures but not another (i don’t see any evidence of bed wobble up and down, and again, these are a new set of y rails and did not solve this issue (did solve the horrible slope)
oh and i tried manual calibration too on the glass.
oh and i have tried with z backlash both on and off
can you think of something else i should i look at
Oh I understand more what you’re saying, the bed seems to be varying relative to the toolhead. Sorry, slow.
With the bed at print temp and the nozzle cool if you jog to z=.1 or something and then jog all over the bed (or at least where you’ve been trying to use) checking with the calibration card is the bed leveling accurately following the surface of the bed?
(as the first 3 lines are what a print job does) then i jogged to roughly same area where the single problem square i am testing (back center of bed)
at X160 Y300 to Y320 -the head is too close the white card has quite a lot of pinch
oh scratch that (sorry typing in real-time)
This is VERY weird as I jog backwards and forwards from X160 to Y300, Y310 and Y320 the head sometimes pinches the card and sometimes doesn’t in each of those 3 locations.
@brent oh and here is the real weird part - the skirt around the object is always perfectly ~0.2m high - this only seems to apply to the object itself
printing a whole benchy in that location now (already printed benchy base in that location as part of the test that produced these, they were scattered all across 1/2 the build area)
I may be wrong, but that might be attributed to the fact that it’s not actively printing so the firmware is not compensating for the bed mesh. But if its varying on and off of pinching the card in the same spots, that is weird. Is the head mounted tight? It almost sounds like something is loose. Is there any play in the X axis whatsoever? X axis mounted tight to the z rails?
My understanding is at startup with the 3dp toolhead the mesh is turned on. It can be verified with M420 I think, the output should include something about leveling being enabled.
It should be enabled at all times unless specifically disabled with an M420 S0.
@scyto it definitely sounds like there’s something not matching up between what the controller thinks is flat and what reality is. You could try disabling at leveling with the above command however be aware that it will suddenly drop like 7 mm, so don’t do it with the head near the bed.
I can’t remember if it was with you or somebody else I was talking with a while back but the leveling was actually double compensating. For example if the bed in a given spot was .2 mm low the mesh was reading .4mm making it high.
I would suggest continuing investigating how the mesh is interacting with the surface of the bed. Obviously the end goal is that at every place on the bed the calibration card has the same tension on it. If automatic leveling is not getting that on its own maybe some of the mesh points need to be manually adjusted or if you’re using a glass bed maybe leveling needs to be disabled entirely.
@brent113 Hmmm, @stewl by chance did you happen to notice if the bed mesh was supposed to be always active even if jogging? Just wondering if you happened to notice when you discovered the double compensation, I have no idea if it is or not tbh.
Well in the home position doing that doesn’t change the head position at all.
However the display coordinates on the touchscreen show a change in Z from 330.01 to 335.16
So i can confirm the mesh is on at all times if the screen is to be believed.
With the mesh off i still get the same issue.
I enabled backlash compensation (it has been off for all of the tests) which sets the Z backlash to 0.06 based on tests a while back using a dial meter. = Same issue.
checked for any looseness on the x and z rail mount points and head mount points, can’t find any looseness.
thanks, thats what i thought you would say, but it is so subtle i didn’t want to lead the witness
yes brand new, i wonder if the old ones were doing it too as the same issues were happening then too…
and it makes me wonder if the x and z ones could do the same…
i think i am done for the night, this week sometime i am going to take the platform off the carriages and see if this right side unit does the same when it is unloaded / not in tension with the other rail via the platform.
in the mean time here is a benchy from around that spot (though it was printed from left to right, and so that may mitgate some issues) but it makes me wonder if this is an academic issue or not…?
Any of the rails can have this happen, now is the part where you contact support, it requires opening the rails and if you do it without them telling you to, it’ll void your warranty. I kept thinking… this sounds so similar to loose carriages when you mentioned the card pinching. Snapmaker can’t even seem to tighten them to their own spec, which I feel is a fool’s errand beyond paper.
For sure that abrupt camera movement at the end I thought the snapmaker was going out the window.
That’s understandable since the spec requires adjusting it to like a hundredth of a millimeter, it’s a poor specification.
Honestly if you’re just careful to not rip the micro switch out, like do not force anything if something’s binding, it’s really easy to work on if you’re at all mechanically inclined. Just get the rollers so that they have slight tension when the carriage is inserted.
Around the 5:40 mark you can see where the micro switch is where it can bind
@brent113@Artezio thanks, I have an open support call on the layer issue that’s been going back and forth for a week or two , they have links to this thread so I will wait until they tell me to open the rails. Maybe they can ship me some pre-tested and adjusted ones from china. The last ones came from a US drop shipper.
