Inconsistent layer lines
I'm not sure if this is the same as issue 602 or not, but I noticed the layer lines at several angles are not uniform at all. As you can see in the picture, they are not very consistent. I have tightened the belts, lubed everything, replaced the nozzle, played with retraction, jerk, acceleration, and used multiple filaments, all with the same results. It was a bit worse before I replaced my nozzle, but still not perfect.
I've seen a million solutions on the 602 github, but it seems like most of them are snake oil, while other forums that ask about it don't give many answers.
Any ideas how to fix this?
RE: Inconsistent layer lines
What is the tension of your belts?
RE: Inconsistent layer lines
@tim-m30
X is 287 and Y is 284. The Y belt seems like it's perfect, along with the fact tightening it doesn't really change the tension value. The x belt is slightly looser, but with the tensioner I added, I still can only get to 287. I could MAYBE tighten it by 1 tooth if I really man-handle it.
Here's another picture of a close up of a string test. The issue is very obvious with small circular motion.
RE: Inconsistent layer lines
Not asking what the belt tension numbers are - they tell you nothing about belt tension - you can bend the frame and they won't change much. They monitor motor current when driving the axis during a self test, and that is the only time they are updated. I was asking what is the belt tension (in pound force [lbf]). You can check tension using the deflection method. Move the extruder to the left, the bed to the rear, apply 7 ounces upwards pull midway on the lower section of the belts, they should deflect 15/64" (X) and 12/64" (Y). This is about 6 lbf static tension. You need at least 2 lbf, and no more than 8 lbf (motors are rated 6 or 8 lbf max, depending on version - and where you place the set screw offset).
Hysteresis (ovoid circles) is also common when a drive gear wasn't installed right. Here's a procedure: easy to do and assures the gear isn't a problem.
Flat First: Drive gears must be tightened flat-first.
1) Start with both set screws fully loosened so the pulley is free to spin on the motor shaft.
2) Align one set screw dead center with the flat on the motor shaft, slowly tighten the screw until it fully contacts the flat surface.
3) Torque the flat set screw to spec.
4) Now tighten the jam set screw, and torque it to spec.
Once tightened, never touch the flat set screw unless the jam screw is first fully loosened.
Why Flat First? Set screws have flat ends. If you tighten the jam screw first the set screw on the shaft flat doesn't fully contact the shaft, only one small edge of the screw surface is biting. Reversing torques can easily shift the shaft to a position the flat screw no longer contacts the shaft. This lets the jam screw wriggle loose. And after a while, vibration loosens the screws until the gear is free to rotate.
RE: Inconsistent layer lines
I checked and adjusted the belt tensions to as close as those desired specs without a spring scale on hand. I additionally reset the set screws, tightening the flat face first. There was little to no change in the ringing quality of a string test.
RE: Inconsistent layer lines
I first used a trigger pull scale for my belt testing and later made a beam balance using 200g of steel shot on one side and the other side of the arm pulling up on the belt under test, with a printed belt hook that has two rules, one at the nominal belt spacing, one at the 1/4" point. It's wide enough the top is 12/64 and the bottom is 15/64). It isn't precise, but gets me close enough.
Discrete steps in layer position are hard to explain without a loose gear or belt that has play. The goal post part you posted appears to have them, so the gear made sense. If belts are tensioned and gears are well attached and not slipping, the only other things that would cause the variation is a loose hot end (heat sink clamp is wiggling) or sloppy loose bearings (this happens after use, so it's unavoidable over time).
Try wiggling the nozzle; there should be zero play in any direction, neither bearing wise (the entire extruder wiggles) nor hot end clamp (the nozzle wiggles within the extruder). And try wiggling the bed, there should be almost no play (check for rotational play - a worn set of bearings/rods will have 1/16" or more play as you twist the bed). If you get wiggle that feels excessive, check the U-bolts are snug, and the bearings aren't free to move in their slots (bolts should be snug, but not tight).
It isn't out of the question you received a bad rod, or damaged bearings during install.
Generally the quality is good enough you'll see stepper inducts artifacts, not noise. But as filament leaves the nozzle, as long as the width is 0.4 mm or more, the filament pretty much lays where the extruder positions it. But it can move around since the nozzle isn't precise, the printer is buzzing, etc. And using a smaller extrusion (0.35 on a 0.4 nozzle) will allow more surface noise since the filament can wiggle withing the diameter (0.05 mm) as it is drawn out of the nozzle. You see this in supports.
And there are stepper induced VFA's ... again, these are deterministic rather than random: the first link goes into a test method for finding best setting for the TMC drivers.
RE: Inconsistent layer lines
@tim-m30
It might actually be a loose hotend. It didn't require all that much effort to make the hotend flex by pushing on it. All my bolts were tightened, so I think there might have been a little dimensional inaccuracy, flexibility to the holder, or it melted ever so slightly to allow the flex.
I looked at this page and it says to print the holder out of something less flexible and/or adding something around the heatsink to tighten the tolerances. With one and a half hands, I couldn't effectively get tape around the heatsink like it shows, so I think I might try printing the whole thing out of abs for its rigidity.
https://github.com/PrusaOwners/prusaowners/wiki/How_to_make_prints_better_on_the_mk3
RE: Inconsistent layer lines
Flex at the heat sink generally isn't a problem; and I'd expect it to cause directional issues, not a random surface. But if the nozzle is moving when the plastic around it isn't moving, that is a problem to solve. The nozzle should be rigid within the housing within reason - trying to flex it with a lot of force doesn't say much. But if a finger tip can wiggle it ...
I wish there were a gauge of how much surface variation is typical. Photos don't always portray what is normal - as light can make normal look bad, and abnormal look good.
Here's a part I printed - looks abysmal on one part (with closeup), not so bad on the other. The surface is due to the PLA make up, and probably my laziness in not tuning for the specific mix. It's also a symptom of the wall thickness and markings playing havoc.
And here's other parts that looks reasonably good.
RE: Inconsistent layer lines
Just replacing the hotend PTFE tube fixed this problem for me! See my detailed reply (on January 31, 2023) to this post: https://forum.prusa3d.com/forum/hardware-firmware-and-software-help/inconsistent-layers-3/.