Single wall tower test FAILED: vertical sin wave pattern

ReVeRb

I would like to know what may cause the issue I have. I'm clueless...

I printed a 1 wall thickness tower (50mm * 50mm * 100mm).

the corner are straight and vertical.
the walls make a sinus like wave all the way up.
the wave pattern on the left and right side are as follow: convexe, concave, convexe...
the wave pattern on the front and back wall are the opposite: concave, convexe, concave...

No cooling fan used.
Heatsink fan setup vacuum the air through the heatsink and push it away from the heatsink.
No ambiant air circulation or air draft (enclosed printer).

I had that problem before so I ordered a Duet effector as well as the magnetic CF rods after I realized my old Chinese aluminium effector and carriage weren't dimensionally accurate(by a lot). I though it would fix my problem, but apparently, something else is causing it.

Every extrusion on the printer have been machined inside 0.005 inch. My tower vertical parallelism is 0.4mm or less.

Any idea why there is a "wavy" pattern and how do I fix it ?

giostark

@carl
Mmm... this smell as geometry issue. But...
What material are you using? It suffer of deformation?
Have you tried to turn the part of 45 degree and print it again? (same centered position)
The airflow of the fan is not directional but it expand as a cone. The fans with directional flow have involved blades (are particular , so you notice them quickly if they are like that). Those small fan are not suited for extraction and.. already they move so little air. Like this is even more ineffective.
Before you try the rotated print...
If you try to print some small empty cube on the far part of the plate they come regular? (just 4cm tall , 1 line with , no bottom and top surface).
In my case the extrusion where a little twisted and bended at halfway.
Also bars of different length can produce this result but if you have the Haydn bars and the effector I just can move the error to the pcb fixed on the skates. Are you sure the 3 supports are identical ? (you need a digital caliper, any small differences come multiplied for the length of the bars) Your skates have some little torsion? (try to move them with some force at any height to see if the move or are fixed) .
If you have geometry issue you should have a not so good calibration. What is your deviation after between the first auto calibration and the second?
If you have much deviation between auto calibration for sure you have geometrical issues. (also alternate but similar are not good , the MUST stay more close to each other at any pass, ex 0.025 > 0.022 > 0.027 ecc ecc. The more the number is small the more accurate is your geometry- without points of compensation)
By the way geometry MUST be more perfect as possible. Especially in delta printers.

oliof

I had this pattern when printing single wall containers with ASA-X due to cooldown effects, i.e. you have vertical instead of horizontal warping. wider/more walls help alleviate the issue.

alankilian

@carl If you post your GCODE, I can try on my Duet2/Delta and see if I get something similar.

??? What version of Duet are you using ???
??? What revision of firmware are you using ???

giostark

@oliof
that very small dentation (enlarging the picture is pretty clear visible) depend on the gear of the extruder (not so good) or is because the material is too hot? (or both together)
@carl
Try to print cooling the part. A tiny wall suffer much more for the heat. (is an attempt that will not cost you much). If the part will not detach from the glass and come regular , you are done. I don't see any fan for the part printed . Try to put an air flow from the top. The ventilation should be the more regular on all the side of the part for avoid different tension (just in case).

What is strange to me is the irregular wave on the corner. No warping (enlarging the photo) , the corners on the glass look perfect but deformation as contraction and expansion...mmm...maybe it stick strong on the glass and deform elsewhere? Can be. The printer is all opened.
Or the part is really too hot or is a geometry issue.
Do the deformation for the heat should look like as a twisted and not expanded and contracted as in the picture? Or I'm wrong?

oliof

@giostark I think it's mainly the temperature and the lack of infill allowing shrinkage to the inside. The dual gear extruder effect may play a role, but the machine to test that is currently in a state of disarray.

Phaedrux

Definitely a hot wobbly wall. Print 10% infill and 2 walls and watch it disappear.

Or if you want to keep the print settings exactly the same, print 2 or 3 of those boxes at the same time to spread out the layer time and allow for more cooling per layer.

ReVeRb

Ok, wow !! I'm submerge with questions and I really appreciate !! I'll try to answer all your informations at once and give you more informations about the printer itself.

Before I go further, please note that 2 weeks ago, I had the same problem. I tested with cooling fan and different speed, in an enclosed box and had the same result, but here is a short resume of what I did and why.

It all started when I tried to reach the best dimensional accuracy I could get from my printer. I started by adjusting the flow, the temps and the speed with different tests. Then I tried to adjust the X, Y, Z dimension accuracy through the M579 command. I struggled to reach my goal and printed different small parts (never that high tho) so I ended up printing that big tower to confirm I had a tuned filament flow(layer width). After I printed that 1 wall tower and discovered the wavy pattern. Here is when I tested with cooling fan and different speed, then I searched on this blog to understand that if I adjust too much through the M579, I would have geometrical problem, which lead me to dismantle the whole printer to find THE physical problem I didn't see when I assembled my printer, then realize my Chinese fisheye and carriage weren't dimensionally accurate. They were the only part I never checked the dimensions, so I though I found the problem. Then I ordered the Duet effector and the Haydn rod and installed them and reset the M579 values to zero, and test print again... ....and same problem!

printer specs:
All extrusions have been machined 0.005 inch. All 3 tower are the same length. All cross section are 300mm (bottom and top). I use aluminium extrusion corner suited for 2040 extrusion on top and 2030 on bottom.

Tower parallelism (measurement from both notch of the vertical extrusion(where you screw stuff), measured with a Neoteck 200mm caliper:
XY:
top: 63.28 and 28.72
mid: 63.20 and 28.72
bot: 63.04 and 28.48
YZ:
top: 62.81 and 28.36
mid: 62.90 and 28.40
bot: 62.95 and 28.46
XZ:
top: 62.93 and 28.46
mid: 62.99 and 28.51
bot: 62.81 and 28.34

PCB heat bed with a 3mm glass. Direct installation with mini kossel bed printed part, no spring.

Linear bearing
Duet effector + Haydn bars (3x 288.06mm + 3x 288.07mm) which I configured as 1 pair of 288.06mm, 1 pair of 288.07 and 1 pair mixed. I set them to 288.065 in the firmware, which is the average (Haydn himself who told me to do so).

wanted Printed part:
50mm x 50mm x 100mm
1 wall thickness at 0.3mm
40mm/s
no cooling fan
260°c head
67°C bed
PETG

Result:
Mostly perfect finish
Corner are straight (except where it's wavy, they are a very very little wavy too)
layer adhesion is perfect
Wall thickness between 0.29mm - 0.31mm
X: 50.02mm
Y: 50.08mm
Z: 100.01mm

I joined the config file as well as the override file.
config-override.g
config.g

ReVeRb

Also, 2 weeks ago, when I tested with cooling fan or no cooling fan, I did a test with 3 tower side by side with 20mm in between and the wave was on all the towers, but at different height from the one in the middle.

Seem like the pattern is also on the horizontal.

I'll print it again to show you what I mean.

alankilian

@carl Can you post or link to your GCODE?

giostark

@carl
Here seems that there are two way of interpretation (missing all the real data) . Too heat or geometrical issue.
i lost ages on my delta so I would be sure (for my printer) , that those stuffs are correct.
Try to do what Phaedrux suggest , at least to exclude the heat side.
If you have cooled the part and the result is the same should not be the heat but I can be wrong.
The temp seems a little too hot , is suggested by the vendor?
https://www.simplify3d.com/support/materials-guide/properties-table/
This material should have less deformation against the others but...

Ok the deviation between several attempt of auto calibration? (run 5 x G32 in a row and copy past the results of the line reported in the console)
In the bed.g try to set a perimeter that is big enough for calibrate the extreme without crash on the extrusions :
http://www.escher3d.com/pages/wizards/wizardbed.php
10 external points , 6 middle points , 6 calibration factor (not 7 or more , you won't calculate the length of the bars that are already measured by Haydn , the software didn't calculate them properly).
Please try to zero the endstop in your override M666 (let the printer calculate each time)
Please also zero the M665 XYZ ABC correction and leave just the radius and the right height (calculate it with all zeroed stuffs and a single G30 and update the override).
Please control the belt tension!!! Hope you don't have the metal core ones because are not good for small turn and are too rigid.
Motors dampeners flex a little , if you have them and can avoid them is better for rigidity.

The pcb support for the metal balls attached on the carriages ? The adapter are all 3 identical? Have you measured with the digit caliper? (deep , parallelism , etc)

Please try also to print cube on the far extreme of the glass. If you have geometrical issue the more long are the movements the worst the effect should be. And you should see it even at the first layer , maybe more squished in some points respect others.
In the center , for almost tiny part the first layer can come good enough but on the extreme if there are problems should be more visible.
Your printer is small so could be a little tricky found the issue.

Yep , also the gcode , Alan will help on that side too.

ReVeRb

@giostark

Ok, I did some G32 (4 time), here are the consol result

I homed, then send a G32 from DWC.
(mean, deviation)
before (-0.737 , 0.437) after (-0.000 , 0.014)
(-0.001 , 0.019) after (0.000 , 0.015)
(-0.009 , 0.012) after ( -0.000 , 0.012)
I homed again, then send an other G32.
(-0.018 , 0.011) after (-0.000 , 0.009)

Then I send a G29 (61 points mesh)
min error -0.052
max error 0.043
mean -0.010
deviation 0.022

Also, in config-override:
M665 L(I set manually) R... H... B... X-0.077 Y0.451 Z0.000
M666 X0.355 Y0.115 Z-0.470 A0.00 B0.00

ReVeRb

Ok, I'm actually printing 3 cube of 1 wall, side by side with 20mm distance in between. The cubes are 25mm x 25mm x 25mm.

While slicing, I had to "uncheck" the spiral mode (a.k.a vase mode), otherwise it would never stop extruding in between the cubes. NOTE: My big tower was printed with "vase mode" activated.

Now, it print the 1 wall cubes, all around, then go back to every corner to dip. Anyway, as for now, the walls seem to be perfect, without distorsion.

I suspect that the "wavy" walls were caused by the "vase mode". What is your though about that ?

engikeneer

@carl I reckon it's still thermal related. Could be that your cooling fan isn't strong enough, or wasn't running at full power when you tried that? On my delta I have to run spiral vase prints very slowly so that they cool enough and don't distort (though that's with PLA).
0.3mm single wall is quite thin so it won't take much shrinkage to distort the walls. Quite likely that one side bending in will twist the other sides as well

giostark

@carl
those calibration are pretty good. Its true that the results on a small glass come better than on bigger ones BUT those are still really good.
So 2 cent of deviation is not a number that can produce such distortion. I have the same deviation on 30cm of glass and no distortion , and my system is not that accurate , so the geometrical problem can be excluded. In the past I had similar distortion for geometrical issue , so now for your case I wanted not exclude such waste of time from the variable. What make me think about it is the expansion and contraction of the picture you posted. The heat problems I faced till now have just torsions.

For the side of heat... this is what is explained for the vase option...
https://all3dp.com/2/cura-vase-mode-all-you-need-to-know/
Like is stated maybe 40mm/s with a not proper cooling could be problematic.
You can try to print slower and with a stronger cooling.
Apart understand the printer the other tricky variable is understand the material we are using.

Wall Line Width:
The width of a single wall is preset in the slicer and corresponds to the nozzle size that you’re using. For a 0.4-mm nozzle, the width of a single line is 0.4 mm. A higher line width will create a much stronger print, and a larger nozzle will produce a higher line width. But swapping out the nozzle isn’t the only way: A 0.4-mm nozzle can also print a wall thickness of up to 0.6 mm if you get the following two settings right.
Temperature:
Vase mode or not, getting your temperature settings dialed in will go a long way. A larger nozzle size will be pushing out much more material, so a 0.6-mm nozzle will need a slightly higher temperature than your 0.4-mm nozzle. And it’s the same story when you’re printing with a higher line width. Try increasing the temperature in steps of five degrees at a time.
Speed:
Vase mode needs to be slow, so that the material has time to cool before the next pass of the printhead. Plus, if the line width is higher, you need to go slower still. A 25–30-mm/s wall printing speed is a good starting point.

mrehorstdmd

I would chalk this problem up to a combo of the material used, the shape of the object, and the nature of FDM printing. A straight line single wall is very easy to deflect from its position- more so than a curved wall (that one reason why your car is full of curved sheet metal instead of flat panels). The longer the straight wall, the easier it is to deflect. So easy that a little shrinkage of the plastic as it cools will cause the wall to deflect away from the target position. That means the next line will be slightly off to one side of the wall. This will start happening vey quickly as each line of plastic starts shrinking as soon as it is laid down.

The answer? print more than a single wall as others have mentioned, or maybe try a much bigger nozzle/wider line width. You might get better results with a material that shrinks less. Like poor overhang performance, this is one of the inherent limitations of FDM printing. Fortunately, we rarely need to print large, straight, single walls.

ReVeRb

Ok, thank you everyone. I'll have to make more test with different speed and temperature. But this morning, when I turned ON my printer, a new inexplicable problem appear from thin air.

When 1 start up, in DWC, I see the tool position coordinate, as usual: X0 Y0 Z292

As soon as I HOME the printer, all the coordinate turn n/a.

When I HOME, the effector move up until the endstop are activated, then it lower a bit, raise again until endstops activate and it stay there... it does not go down a bit as usual.

After the HOMING, I can't move the head all around.

Also, If I auto-calibrate, it do it perfectly... then I send M500. Still the tool position are n/a.

if I try to launch a print, it heat the bed and nozzle sequentially, but it does not home. Instead it goes to X0 Y0 Z0 and prime the nozzle and try to extrude...

I never had that problem before... is it related to the new duet effector or the mesh grid I now can generate since I have the duet effector ?

Which file I should check ?

Phaedrux

At this point please post your config.g, bed.g, homedelta.g, and the results of M122 and M98 P"config.g"

ReVeRb

Ok, I've wrote down how my printer behave now... I joined the files at the end. (I don't know how to show the content of the file directly like other do).

I turn ON the printer and the tool coordinate are shown (Z0 Y0 Z292)
If I HOME, the printer, all the coordinate turn n/a.

When I HOME, the effector move up until the endstop are activated, then it lower a bit, raise again until endstops activate and it stay there... it does not go down a bit as usual.

After the HOMING, I can't move the head all around.
3) If I turn ON the printer and directly load a print and launch it, the tool coordinate turn n/a, the sequence of heating is normal, but it HOME 1 stime, then it go down CRAZY fast to X0 Y0 Z0 and start extruding (extruder crackling)

I paused and still can't move the effector around
I canceled and still the same

I can only HOME (or maybe launch a G32, but didn't tried it).

I notice that in the "resurrect.g" there is a file I started yesterday (my last print) that I did not finished. Normally, I pause and cancel a print, but I may have hit the "emergency stop" red button at top right corner by inadvertence.

[4_1629312457066_bed.g](Uploading 100%) [3_1629312457066_config.g](Uploading 100%) [2_1629312457066_homedelta.g](Uploading 100%) [1_1629312457066_M98_M122](Uploading 100%) [0_1629312457066_resurrect.g](Uploading 100%)

ReVeRb

the files
resurrect.g
homedelta.g
config.g
bed.g

8/18/2021, 2:33:49 PM M98 P"config.g"

HTTP is enabled on port 80
FTP is disabled
TELNET is disabled
Warning: Heater 1 appears to be over-powered. If left on at full power, its temperature is predicted to reach 882C

8/18/2021, 2:33:13 PM M122

=== Diagnostics ===
RepRapFirmware for Duet 2 WiFi/Ethernet version 3.2.2 running on Duet Ethernet 1.02 or later
Board ID: 0JD0M-9P6B2-NJ4S8-6J1FJ-3SJ6L-T84LK
Used output buffers: 3 of 24 (11 max)
=== RTOS ===
Static ram: 23460
Dynamic ram: 68116 of which 0 recycled
Never used RAM 20504, free system stack 200 words
Tasks: NETWORK(ready,186) HEAT(blocked,308) MAIN(running,457) IDLE(ready,20)
Owned mutexes:
=== Platform ===
Last reset 00:00:37 ago, cause: power up
Last software reset at 2021-08-18 14:21, reason: User, GCodes spinning, available RAM 20468, slot 2
Software reset code 0x0003 HFSR 0x00000000 CFSR 0x00000000 ICSR 0x0041f000 BFAR 0xe000ed38 SP 0x00000000 Task MAIN Freestk 0 n/a
Error status: 0x00
Aux0 errors 0,0,0
MCU temperature: min 33.2, current 36.3, max 36.6
Supply voltage: min 12.1, current 12.1, max 12.2, under voltage events: 0, over voltage events: 0, power good: yes
Driver 0: position 55012, standstill, SG min/max not available
Driver 1: position 55012, standstill, SG min/max not available
Driver 2: position 55012, standstill, SG min/max not available
Driver 3: position 0, standstill, SG min/max not available
Driver 4: position 0, standstill, SG min/max not available
Driver 5: position 0
Driver 6: position 0
Driver 7: position 0
Driver 8: position 0
Driver 9: position 0
Driver 10: position 0
Driver 11: position 0
Date/time: 2021-08-18 14:33:12
Cache data hit count 52825024
Slowest loop: 10.77ms; fastest: 0.21ms
I2C nak errors 0, send timeouts 0, receive timeouts 0, finishTimeouts 0, resets 0
=== Storage ===
Free file entries: 10
SD card 0 detected, interface speed: 20.0MBytes/sec
SD card longest read time 4.9ms, write time 0.0ms, max retries 0
=== Move ===
DMs created 83, maxWait 0ms, bed compensation in use: none, comp offset 0.000
=== MainDDARing ===
Scheduled moves 0, completed moves 0, hiccups 0, stepErrors 0, LaErrors 0, Underruns [0, 0, 0], CDDA state -1
=== AuxDDARing ===
Scheduled moves 0, completed moves 0, hiccups 0, stepErrors 0, LaErrors 0, Underruns [0, 0, 0], CDDA state -1
=== Heat ===
Bed heaters = 0 -1 -1 -1, chamberHeaters = -1 -1 -1 -1
Heater 1 is on, I-accum = 0.0
=== GCodes ===
Segments left: 0
Movement lock held by null
HTTP is idle in state(s) 0
Telnet is idle in state(s) 0
File is idle in state(s) 0
USB is idle in state(s) 0
Aux is idle in state(s) 0
Trigger is idle in state(s) 0
Queue is idle in state(s) 0
LCD is idle in state(s) 0
Daemon is idle in state(s) 0
Autopause is idle in state(s) 0
Code queue is empty.
=== Network ===
Slowest loop: 9.94ms; fastest: 0.02ms
Responder states: HTTP(0) HTTP(0) HTTP(0) HTTP(0) FTP(0) Telnet(0), 0 sessions
HTTP sessions: 1 of 8
Interface state active, link 100Mbps full duplex