So far I have not seen any recommendation with a Blanking Time of 0 but you might have good reason to set it.
Generally speaking:
Hundreds, if not thousands, of Ender 3 are running fine with Klipper
None of your mentioned components is “special” (One of my printers also run of a SKR1.4T and on another I use an Orbiter 2.0)
The PA calibration settings are not a good reference point, since a too low SCV will also cause issues: It is exaggerating speed change effects that is why it is used for this calibration
I still believe that you have mechanical issues or to put it differently: I think it is not caused by Klipper or its configuration.
I have an oscilloscope here as well but no current probe yet…
But I did not test if there is any noticeable difference between different settings. I’m using those settings on all my Orbiters even the Voron V0 with the standard tool head.
Would you maybe share your settings although its offtopic here - maybe via PM?
Will not make much sense since I do not use the LDO but a Moons stepper. The LDO died after 2 weeks and I had a Moons lying around.
TBH, I do the tuning for pure fun of it. Never noticed too much of a difference to the default. Maybe a bit less noise but this is anyway not my paramount concern. Hard to judge. Seems like a bit of black magic
spend some days to figure it out. Made about 30+ Lochwürfel with different settings one by one.
I also calibrated the Extruder before the test marathon and changed the Phaetus Dragonfly BMS for a Dragon HF and dual z after it because the artifact is still there after the tests…
BUT the artifact is going away when i ramp up the outline perimeter speed to above 40mm/s. Then the corners are a bit bulgie. With PA the artifact shows up again. It seems to me that the little underextrusion before the corner needs the excessive material after the corner for a perfect look. Because i can get it to work without the artifact before the corner and a real nice straight line but then i got excessive material after the corner…
So that means to me there is a delay in the movement of XY and the extruder caused by the heavy toolhead and bed i think. I also checked the tightness of all screws and did another resonance compensation test with my adxl. Now it suggests mzv 65hz for x and an acceleration of max 12700mm/s². But i still use 2100mm/s². For y i do the resonance compensation test later.
I think its because of the heavy bed and toolhead and the only one 6mm belt for each axis that i see those artifacts.
Any suggestions?
Maybe im asking too much of my Ender 3 and should build a Hypercube
EDIT: Also got backlash of 0.1mm on Y. The belt travels straight between the pulleys. Checked everything on y multiple times and also changed the pulleys and even the motor. Dont know how to get rid of this. X is absolutely precise. Maybe it has something to do with the backlash?
Very strange…
Can you upload photos of your print head and the whole printer?
In what mode are the TMC2209 running especially the feeder, StealthShop or SpreadCycle?
Even if you would be running the printer too fast, the artefacts should occur after a corner and not before.
Maybe the decelleration conflicts with PA.
Hmm, the print head is a bit bulky and its center of mass is quite far away from the X axis but nothing you can’t print with.
However the fan duct would profit from having the good Sunon fans rotated 180° around Z so they are facing backward not forward. But this is not related to this strange behavior.
All axes are moving smooth and easily when being moved by hand?
The issue is also present with a 45° turned cube is it? If I recall correctly.
If nobody else has a brilliant idea yet I would suggest grab a new MicroSD card and reinstall Klipper from scratch with KIAUH and only the minimal config.
Just setup basic parameters like motor current, rotation_distance so everything is properly moving and then reslice the Lochwürfel and give it a try.
Depending on your slicer just check the flow preview that it has constant flow around the areas of interest.
Thank you Brian for your thoughts, i appreciate it Can you tell me more about the fans turned by 180°? Whats the benefit? Im using the HeroMe stuff. But maybe i should design something by myself.
Super smooth movement on all axis, hm. I also ordered different bearing balls from Kugel Winnie to eliminate play at the rails. 45°turning - check, you recalled correctly.
I wiped my complete Raspberry sd card, BUT i never used a minimal config after the wipe.
Will try that, thank you
My hope is we will find the root cause of this issue and can finally solve it! But it is not so obvious…
Regarding the fan duct, maybe you saw my video on 3DDC in the “[Tipp] Bauteil-Kühlung” thread:
As the ducts themselves point backwards the fans blow their air stream against the duct wall and not directly “into” the ducts. This will cause turbulences within the air stream inside the duct making it less effective.
In my opinion it is better to mount the fans backwards so the inclined air stream is parallel to the duct walls but this would require redesigning the fan duct.
And it would move the center of mass more towards the X axis.
That is a bulged corner. Is PA enabled and calibrated?
Are all corners like this one?
Where is the Z seam in this print?
Is SCV still 5? Those corners will become better with higher acceleration.
But your main issue seems to reduce flow way before the corner and then increases it with the corner. As if PA or your feeder or something is not synced with the rest of the kinematic.
Just wanted to chime in here- banti im having the exact same issue of an indent before my corners and you I agree with your thinking that something is out of synch witht he XY motors or steppers. The indent /underextrusion before corner is clearly not a PA issue to me, and in any event I have also test with/without PA, input shaper ETC. I’ve found a few others have posted about this issue, with nobody diagnosing the cause. I will probably try swapping out my XY motors/steppers one by one and see if that fixes the issue.
Trident, so coreXY.
Fundamentally this issue/artifact is caused by either the XY motion system being off, or by the extrusion being off.
It persists throughout no, proper, and too high PA, with properly functioning hotends and extruders. I believe I saw on another thread someone replaced their extruder and hotend- and still had this problem. So I can think of no reason why the issue would originate with extrusion, although that was my initial guess. A closeup of a one wall cube with this “indent” does also look non square I would say.
Does the defect as shown on the white cube propagate up on the entire z-height, or is it just the first layer?
What are the accelerations and velocities with which this example have been printed? If printed with very high velocities / accelerations, does it change when lowering these value to be rather conservative?
