Consistent Inconsistent/Uneven Layers After Upgrading Ender 3

[JackL 0]

So I've had my Ender 3 for a while, and I've recently finished 2 upgrades for it: I've upgraded my motherboard to a MKS Gen 1.4 with TMC 2208 Stepper Drivers (from Waterott) for X, Y, Z, and E that are currently in the legacy mode (not using one-time-program or using UART), and I've added the Petsfang Duct with a 5015 blower fan, and replaced the hotend fan with a 40x40x20 12V Noctua (I wanted to see how it compared to the 40x40x10 12V Noctua for anyone wondering).

Once I finished, I started printing calibration cubes (https://www.thingiverse.com/thing:1278865) and found inconsistent lines/uneven layers in the prints. However, when I modified settings and printed again, I found that the lines happened in pretty much the SAME places/layers. The lines showed up consistently in those layers as I modified print speed and a few other  settings in Ultimaker Cura 4.0.0. I've also tried printing in a place other than the center of the build plate, and ended up with the same result. I thought my X-axis belt was a bit loose, so I tightened it up a bit, but I still get the same lines, in the same layers in my prints. I've tried adjusting the basic acceleration and jerk settings in Ultimaker Cura 4.0.0 to no avail. What should I try next?

[gr5 2,167]

Dirty Z screws are the usual cause of this phenomenon because it causes the bed to move "not quite enough" or "slightly too much" and this causes under or over extrusion which makes layers stick out or in.

 

Also bad PID values can cause this if the temperature is fluctuating by quite a bit.  By like at least 5 degrees from max to min temperature.  Watch the temp while it prints.  It may be that the temperature algorithm or the PID values changed on the latest Marlin  upgrade.

[JackL 0]

11 hours ago, gr5 said:

Dirty Z screws are the usual cause of this phenomenon because it causes the bed to move "not quite enough" or "slightly too much" and this causes under or over extrusion which makes layers stick out or in.

 

Also bad PID values can cause this if the temperature is fluctuating by quite a bit.  By like at least 5 degrees from max to min temperature.  Watch the temp while it prints.  It may be that the temperature algorithm or the PID values changed on the latest Marlin  upgrade.

Ok, so I've tried cleaning the Z-axis screws and relubricating them, and I've also autotuned PID and set it. However, I still don't see much improvement. I'll try some more things related to these tomorrow and see if anything changes.

Edited May 28, 2019 by JackL

[JackL 0]

Yea.. so after trying a few different things I found that I am still getting the same results.

[JackL 0]

Ok, I know this has been a while from my previous updates, but I got caught up working on another project.

 

So I have cleaned the Z-Axis thoroughly and lubricated it again, and although the effect seems reduced, though still notable. The print on the left is at 20mm/s (no speed settings above that speed) and the right is at my normal 50mm/s default settings. The uneven layers still seem to occur at around the same layers. What should I try next? What else could be causing this? Could this be something like Z-Wobble?

[geert_2 557]

Does this also happen with a slightly larger block with *plain* walls (=without any text or indents)? Just a clean cube?

 

In my UM2 printers, when printing small objects with very sudden changes in layer area, that also shows up as "horizontal lines" in the print, due to the differences in layer cooling time. See the images below.

 

For me, printing slower and cooler helps, or for small objects, printing a dummy "cooling tower" next to the real model. But this will only work for you if this is the cause, of course.

 

I could imagine that an irregular feed also would cause this problem: changes in flow rate, filament diameter, kinks in filament,... Or irregularities in nozzle temperature, or air flow around the printer. Or vibrations in the X, Y and/or Z-axis?

 

Sift for the laboratory sink:

 

Tiny part (see the irregularities in the side-wall in front); the ruler is in mm/cm:

 

[JackL 0]

Ok, I apologize for the delay, I got caught up with a few things the past few days. So I tried printing something bigger, and the effect was greatly reduced. However, the lines still appear to be there on each print, though very subtle. I made something like the ripple test that I was using in Autodesk Inventor, but made it bigger (30mm x 30mm x 30mm) and with no letters or dots in it. I notice that the lines still happened in the same layers between these 2 prints. One of these was printed 10 degrees cooler(printed at 190 degrees), but I cant tell which one just by looking at them. I'll try a print at 180 degrees once I relevel my bed.

Also, in the middle where the print transitions to rounded corners, there seems to be a small shift on one top square. One side (in the image below, it's the left side) seems to have shifted slightly into the print where it should be lined up with the side, and on the opposite side of the print, it's sticking a little bit outwards, like a really small layer shift. This shift happened on both prints on the same layer. Another thing I noticed when observing these prints is that the printer seemed to go a little bit passed the corners when rounding the square corners, as seen by the slight outwards bump on the corner in the image below. This happened on all of the square corners on both prints.

I'm more worried about the layers than the corners, unless they have some sort of connection to the problem. What should I try next? I could find out which axes the shifts are happening after I level my bed again, as I wasn't thinking about the orientation as I was taking these off the printer. Should I try something like recalibrating my axes? Or what should I try next?

[geert_2 557]

The overshoot on the corners is most likely mechanical: oscillations and vibrations because the printer head can not immediately stop and change direction. Plus it will extrude a bit more material per mm while the head is slowing down, compared to printing straight lines. This effect can be reduced by printing slower, and by changing some settings, although I don't know which ones (never experimented with that).

 

I have no idea about the shift. Can you see that shift in the layer preview in the slicer, when zooming in a lot?

 

I guess if you move that transition from straight to rounded corners up and down, the shift also moves?

 

If you look close while printing, does the printer change direction at that point, from clockwise to counter-clockwise, or so? In which case it couls also indicate a mechanical problem such as too much play?

 

[JackL 0]

Ok, so I've been trying a few different things and started printing things because I can kind of live with the very small uneven layers on bigger prints. I can tell that it is not the temperature that's the cause of anything. I also downloaded the Creawesome mod for Cura and that seemed to reduce the issues and get rid of that tiny layer shift. Even though the Creawesome mod did reduce some effect of the lines on the test cube (not the one I made), the lines were still there.)

Yesterday, I designed a stand for a floppy cube (yes I am a cuber) as seen below.

However, on the finished print, I did notice something I had not seen before. When printing something tall and skinny, it didn't really print straight but rather printed in a kinda wobbly manner when printing upwards. Another weird thing is that this happened in pretty much the same place on the 2 prints (left is first prototype, right is the most recent prototype).

Could this be a clue as to what my problem actually is? Maybe Z wobble? If so, how should I go about fixing it? If not, what should I try for next?

Another thing I was noticing while printing was that the PTFE tube in the extruder fitting slid in place with retraction, but never actually popped out while printing. I tried printing without retraction (so the PTFE tubing wouldn't move in the fitting) to see if this movement could be the cause of the issue, but the lines still appeared on the test cube, so I don't think this is the main issue.

[geert_2 557]

I don't know your printer, but this looks like too much mechanical play in the system? Things wobbling around? But I would rather guess the X or Y-axis.

 

[JackL 0]

I apologize for not posting in a while. I haven't been able to work on this since it was close to the end of the semester at my school, and then I got caught up with a few things in summer. However, I did manage to get a 3D printer I've been building (I've been building my own BLV MGN Cube, with a custom frame), and while running test prints, I ran into the same problem that I did before, with consistent uneven layer lines. However, this time I know it's most likely not attributable to faults with the Z-axis (for those who didn't click the link, this printer has a dual Z-axis motor setup) because the lines appear in the same areas with scaled up parts. I printed a ripple test cuboid (I'm still tuning slicer/printer settings), scaled it up to 150% size, and printed it again with the exact same settings. To make sure speed wasn't a big factor in this, I printed it at 60mm/s and 20mm/s. But surprisingly, there was little to no difference between the prints at different speeds. Here are the prints:

 

20mm/s

 

60mm/s

 

If you look closely at the uneven lines, you can see that they happen in the same areas on the cube, despite the differences in Z-height. With a few more of these prints, however, I found that putting the seam in one corner seemed  to reduce the effect a little bit, though there were still some lines in the prints, in the same areas.

 

Seam set in the corner

 

What should I try next? Is there something I should try with the seams to further reduce the effect?

[jzzNN 0]

Im exactly with the same problem in the same test Cube.

 

I build my ratrig vcore kit (like hypercube) and Im facing this same issue. Did you manager to fix it?

[xbobys 0]

Hello there, I got something similar.

Printer: Ender 3 (age - 4 months, regularly cleand and maintained)

I printed this pinecone in the picture below and have noticed these strange, pretty random layer shifts.

E-Steps are calibrated.

 

After I continued with test cubes and troubleshooting which is shown and described below.

1) XYZ Cube printed few months ago - pretty flawless

2) XYZ Cube printed now, sliced with Cura

3) sliced with PrusaSlicer

4) Cleaned Z rod and nut (Cura sliced from now on)

5) tightened pulleys and belts

6) cleaned frame and wheels touching it for all axis

7) switched motherboard (from silent back to "nosiy") + moved object from center to corner of the bed

8) PID autotune + swithced motherboard (from "noisy" back to silent)

9) scaled model to 300% - no misalignments at previous problematic layer heights

10) changed filament

11) scaled Z dimension to 500% - misalignment moved up with the letter, plain walls are perfectly aligned

12) rotated 90° (X Y flip) - misalignment rotated as well

 

I have run out of options and ideas. Did You managed to figure it out somehow?

Referring to troubleshooting i did, shifts are not present at plain walls, but on walls which contains any kind of detail (pinecone is all messed up, lot of details). As layers with details appears, whole layer is slightly moved (best shown in picture 8,10,11 - left wall at about the X letter starts).

Thanks for any ideas

[DarRAven 4]

On 6/24/2019 at 5:47 PM, JackL said:

 

 

I guess the cooling seems weak It prints each layer top of other layer(as everybody know) if bottom layer is not cool enough this could happen

Edited May 5, 2020 by DarRAven

[JackL 0]

From what I've gathered from this and from my experience, these "consistent inconsistent layers" seem to not be from the fault of the printer. In fact, I've observed from the Lulzbot Mini 2, Prusa i3 Mk3 (to a lesser extent), and my BLV MGN Cube that I've built. So, I tried printing a few things and showing them to people like family and friends and nobody commented on the lines. So, I just accepted the fact that I was being a bit of a perfectionist and stopped caring too much about it. Some prints I've found really aren't effected by it that much, and I have gotten some pretty good prints that happened to come out without much lines, if at all (no significant changes in the slicer settings).

 

As I have mentioned in the post, they do seem to change with the placement of the Z-seam. You could try other slicers such as PrusaSlicer to see if it results in changes that satisfy you. Here's something else that might help: This video by CNC Kitchen (https://youtu.be/Hvw3DrVAeTA). Basically, the video talks about adjusting the "Maximum Resolution" and "Maximum Deviation" settings in Cura. Because the lines are related to the Z-seam, I believe that the problem is related to the blobbing issue in general. I have tried this with some success in my free time, but haven't looked into it too much due to being busy with other things.

[Volacar 0]

Same here...basically commenting as a "metoo" thing and to keep this thread alive 

I hadthe issues with my HEVO, HEVO-RS and my BLV Cube (one printer,  modified in hope of getting rid of this s**t)

Images are from tst pieces placed in all four coners of the bed to maybe pin it to one part of the printer but nope

[tank423 0]

PLEASE don't kill this thread I have been experiencing the same issues after upgrading my ender 3 currently testing to figure this out and could defiantly use some added insight to the issue just a hint tho.... this is driver related picture are on the way but i first began experiencing this after upgrading my ender 3 pro with the SKR MINI E3 V1.2 board with the 2209 in UART after trying everything z screw related and even equipping the printer with the 3d fused Lin rails BTW there not a great upgrade but there ok, I finally realised this had to be firmware related so i tryed many different marlin configurations running the steppers in 1/4- 1/8 stepping with and withought the 256 interpolation changed the r sense current and was seeing small improvements believe it or not i couldn't !!! SO i have since purchased the SKR MINI E3 DIP, SKR PRO V1.1 and the SKR 1.4 BOARDS as well as a set of SKR 2130 V3.0, PANUCATT TMC2209 eaasy uart and of course a bunch of A4988 drivers and hears where I'm at to date.     Using any of the 3 STEP-STICK SKR  boards outfitted with A4988 drivers I get great prints or at least i do not have layer shifts with the TMC2209 drivers on the XYZ and a A4988 driver on the E0 I get good prints but I have to tune my extrusion down to 88% for the outer wall and 95% inner wall in cura with the overlap at 10% all of this I never had to do when my Machine was stock I have re-calibrated the extrusion steps played with the jerk setting and acceleration and have since tried the full set of 2130 drivers on all 4 axes with these tuned extrusion setting and am finally seeing results but noting compared to the amazing quality i was seeing with my stock ender 3 my opinion is that these boards are not properly hardware configured and the change in directions in back feeding the extruders stepper using an A4988 in 16x micro-stepping requires more voltage/ current to move a step or to hold making the tiny fluctuation have less effect on the extruder BTW i see this with my multi-meter and oscilloscope on the board so its not a question of weather its happening its weather its causing the e stepper to loose accuracy. I have spend some money trying to solve this issue and now that i know others are experiencing it too maybe we can figure it out AGAIN pics are coming I just stumbled on this thread and I'm hoping its still alive anyway hope it is 

[GregValiant 1,245]

I'm old.

I started with analog computers, and punch cards.  The g-code went on 1" wide paper tape with 1/8" holes punched in it.  The paper tape was fed into the CNC machines and they moved and worked all by themselves.  It was magic.  (No...they were not steam powered.)

Over the course of time I came up with some observations that have worked for me.  I call them rules, but they're more like guidelines.

 

#1.  When you're dumb ya' suffer.

#2.  Never upgrade an operating system.

#3.  Everything affects everything.

#4.  Never blindly trust a Gcode file that was created by someone else.

 

There are a few more (yes, some involve the opposite sex) but I won't waste anymore of your time with them.

[Volacar 0]

Got some infoormations for you, additional to Gregs wisdoms (experienced #2 waaaays to often btw.)

I made some changes that got rid of all banding issues, that you can remove - so currentlly only having the "benchy hull line" apparent:

1. tightened the Bed leveling screws: a 400mm Bed on three m4 screws was not the smartest thing to do but still workking IF the springs are really compressed.

2. PID-tuned the Bed: Large AND thick aluminum bed with springsteel on top makes thermal expansion an issue and as the standard in Marlin is Bang-Bang,  it really moves quite a bit

3. my X carrier rail sled was extra loose, meaning a possible wobble of <1mm

 

hope these points might help others

[RizHassan 0]

Has anyone figure out a solution to this problem?

 

I am printing PLA on Ender 3 V2, with default settings.

 

I have done extruder calibration, xyz cube, and bed level.

 

A simple thin wall cube prints perfectly, but the minute I try to print large model I get really bad results as shown in picture.

[GregValiant 1,245]

I can just make out the flat area on the bottom and that looks like there are ridges between the nozzle marks.  I'd say it's over-extruded.  The walls, I don't know.  It looks to be a mechanical problem.  Down low on the right - is that a layer shift?  Layer shifting is often caused by loose belts.  They should twang like a guitar string.

If you aren't using Accel and Jerk control - turn both on and set Accel to 400 for everything and Jerk to 8 for everything.  You may have to unhide all the settings.  There are a lot of Accel settings and Travel Acceleration and Initial Layer Travel Acceleration don't update automatically.  Softer starts and stops may help.

Another thing to check is if there is any wobble in the print head.  It should roll back and forth and you should not be able to rotate it around any axis.  Same with the Y table.  You may have to revisit the eccentric adjusters to make sure.

Edited January 13, 2021 by GregValiant

[RizHassan 0]

@Greg,

I have gone thru mechanical alignments and backlash issues using a dial indicator (old fashion way), but  will recheck over extrusion again.

What I cannot understand, is why only this particular part creates ridges and layer offsets compare to a simple extruded thin piece? Is it because its compounded curves that the printer cannot handle?

I will also look into acceleration and dec profiles and jerk control. I am not familiar with these parameters. 

 

One last question I have, if my printer global tolerance in Cartesian system is 1mm, should my part thickness be more than 1mm so these layer steps are reduced? Hypothetically speaking

[GregValiant 1,245]

When a part is exported from it's native CAD software as an STL file it can be from coarse to very fine resolution.  The resolution is usually an option in the exporter.  When you print a cylinder with coarse resolution you can see the facets around the circumference.  The length of the lines that Cura calculates to go around a curve are dictated by the size of the polygons in the model.  Big coarse polygons, coarser print and small fine polygons mean a finer print.

As far as what the printer can handle, you may be able to make an adjustment to the Maximum Resolution setting in Cura.  For my Ender 3 Pro the default in Cura 4.7.1 was .05mm.  That was too fine for the printer to handle and caused blobs.  It's much happier at .5 resolution.  In 4.8.0 that has been resolved by making the default .8.

I don't understand your term "global tolerance in Cartesian system".  Generally, if you make your parts multiples of the layer height tall (the Z) they will be more accurate.  If your part is 93.75mm tall and your layer height is .2 then the final dimension can't be hit exactly.  But if you can live with a part 93.8mm tall and design it that way, then it should end up real close.

The X and Y are different.  Cura builds up to the outer wall and leaves the inside as whatever dimension is left over.  93.75 wide with a line width of .44 and 3 walls should end up at 93.75 wide.  The walls will be .44 * 3 on each side and the interior will be whatever is left.  Thin walled models are different because there isn't any inside area to take up the slack.  The new Arachne version that is being worked on addresses that.  Using Adaptive layer heights can help make adjustments in the Z.

 

Acceleration control places a limit on the acceleration of the print head and of the bed.  The higher the acceleration, the faster printing will be.  The cost is hard stops and starts where any sloppiness in the drive train will result in vibrations and target over-runs ending up in the print.  Jerk is in mm/sec and an analogy would be to think of it as how fast you go around a corner.  With Accel and Jerk control turned off the print head basicly tries to come to a stop, and take off again instantaneously.

 

[RizHassan 0]

@Greg, my solidworks export option is set as "Fine" Should I change to custom and reduce tolerances (See snap shot)? 

 

Where can I find the resolution settings? When I typed resolution in the search field there are Quality, Support, Mesh fixes, and experimental features comes up. If I am not mistaken you are referring to Mesh? Mine is set to .25 for resolutions, and Max deviation .025. Should I set Mesh resolution to 0.5?

 

Global tolerance in Cartesian system meaning - Since its a XYZ gantry (Printer) and each axis has its tolerance, so if you take root sum square you end up with a number that the gantry can achieve theoretically. No need to discuss this any further 😉

 

Yes, I can live with 93.8mm compare to 93.75.

 

Not sure if I understood thin wall analogy. I'll revisit that again later.

 

Acceleration makes perfect sense should that be half of the speed as a starting point? Oh no its set as 500 mm/s^2.

 

Should Jerk be set to less than 6mm/s?

 

Thanks, again

 

 

 

 

[GregValiant 1,245]

Your .25 resolution numbers should be OK  If it stutters going around curves you might want to bump it up a bit.  Your Accel and Jerk numbers are reasonable as well.  For the Enders, if you get Accel much over 500 and Jerk much over 8 and with a fine STL, the line segments are very short around curves so there are a lot of Gcode lines to process.  The combination of resolution, accel and jerk, the speed of the printer/processor, and any slop in the drive train cause issues with print quality.  The trick is to find the sweet spot where you get good resolution, you aren't over-taxing the physical machine (bad corners and layer shifts), the print head isn't stopping because the processor can't keep up (intermittent blobs), prints have good finishes and are dimensionally correct.  What's needed is a calibration model that has curves and lines and circles.  The cube is a decent tool but can't show you the whole story.

 

Back off on the resolution of your model.  These are decent printers but they are by no means a high end printer.  I know I can print some models very fast to the point where the hot end's ability to deliver plastic is the limiting factor.  Those models don't look so good, but most of what I print is functional details rather than art so I don't care.  Layer adhesion and part fit is usually more important to me than finish quality.

 

Come to think of it, post the STL of that model (unless it's proprietary or a paid-for model) and your gcode file as well.  I am just finishing a print and there isn't going to be much left on the spool, and I always use my leftovers for practice.

 

 

Edited January 15, 2021 by GregValiant