Ultimaker S5 strange phenomenon

[chuckmcgee 7]

Not sure if this is an Ultimaker issue, Fusion360 issue, or Cura.  I am printing a simple box.

 

This is the model in Fusion360

 

This is the body in Cura:

 

 

But this is how it is printing (and this is twice now with two different roles of filament and different infill settings (20% on the first attempt and 10% on this one - each print did the exact same thing at the same point)

 

The perimeter at the bottom of the box is narrower then the rest of the wall as it prints, creating an uneven wall up the side.  I have looked at the layer preview in Cura as well as a cross-section analysis in Fusion360 and there appears to be absolutely no reason for that to be printing that way.  I also know there was nothing odd about how it was modeled...a sketch, extrusion, and shell.  Simple box.

 

I have also attached the STL file of the box.

control box.stl

 

Would love some advice about how to fix this.  Thanks in advance!

Edited July 31, 2018 by chuckmcgee

[chuckmcgee 7]

I will say, beyond that strange indentation at the bottom, the rest of the print is going well.

 

 

Could the issue be hardware related?  

[kmanstudios 1,120]

What does it look like in your slicer? Since I have commitments until the first of the week, I could not print a test on my own to double check.

 

While you say two different rolls of filament, what type and brand?

 

I will download the model though to see what happens on my end.

 

Edit: And also, what are your infill settings?

Edited July 27, 2018 by kmanstudios

[chuckmcgee 7]

@kmanstudios Thank you for the reply.  

 

Quick answers - it looks great in the slicer...went to Layer View in Cura and scrolled the arrow to look for any anomalies in the layer preview - looks as expected - perfect.  I will try to capture an gif of that if you think it will help.

 

Re: the two filaments - both were red Pro PLA from Matterhackers.

 

Re: first attempt was an infill of 20%.  Second attempt was 10%.  Model seemed to have that same indentation around the same height.

 

A couple additional notes:

• This is printing from Extruder 2.  I will attempt to print the same model from extruder 1 and see what it looks like.
• I also have a CR-10 and will slice with Simplify3D and Cura on that printer and see what it looks like

Is there a way to export the gcode of the current print that is still going on the S5?  The print with the indentation has about another 3 hours before it completes.  I would like save the gcode if possible.  I have already closed Cura and chose the "Print to Network" option, as opposed to save to disk.

 

Thanks again.  This one is a mystery.

Edited July 27, 2018 by chuckmcgee

[kmanstudios 1,120]

OK, this is interesting. You re printing with extruder 2. But are you using an AA core in that slot?

 

You could also upload your '.curaproject.3mf' file. It will hold all the settings for your setup and allow us to see exactly what you are doing.

 

Make sure that if you do that, you use the 'save project' option. Regular 'save' will not hold the model and all pertinent information.

[chuckmcgee 7]

Yes to AA core in slot 2.  I will save the new settings as a project and post.

[chuckmcgee 7]

Here are the files (recreated in Cura) as I had already closed Cura since sending the job that is running.

 

I have attached the gcode as a save to file as well as the project file.

 

UMS5_control box.gcode

UMS5_control box.curaproject.3mf

[chuckmcgee 7]

Sliced the same model in Cura for the CR-10 and just started that print.

[kmanstudios 1,120]

I got your file. I did make a couple of changes, but I do not see how it would affect your print.

 

Changes:

1.Set profile to UM Yellow PLA (simply because I use the UM PLA profiles for my PLA printing, even when using Matterhackers Pro PLA)

2. Set brim to same extruder (2)

3. Set slicing tolerance to exclusive

4. Maximum Retracton Count to 12 (default 25 which is a bit much)

 

 

These are just basic settings I use. Like I said, I would not be able to test against anything until next week due to commitments. And even then, I would print to just above your problem area to see if I get the same thing.

 

If it is not solved by the time I can get around to printing it, I will load my MH PLA PRO (Just not yellow....do not have that) and see if there is any correlation to your experience.

 

Slicing does look good. So not sure what the issue could be (difference in S5 machines, different stock of MH PLA PRO, etc)

[chuckmcgee 7]

8 minutes ago, kmanstudios said:

I got your file. I did make a couple of changes, but I do not see how it would affect your print.

 

Changes:

1.Set profile to UM Yellow PLA (simply because I use the UM PLA profiles for my PLA printing, even when using Matterhackers Pro PLA)

Will do!

 

 

8 minutes ago, kmanstudios said:

2. Set brim to same extruder (2)

This was already on extruder 2.  Because I re-created the job in Cura to send you the project file, I forgot to change that.  It was changed on the 2 models that printed

 

8 minutes ago, kmanstudios said:

3. Set slicing tolerance to exclusive

Will do

 

8 minutes ago, kmanstudios said:

4. Maximum Retracton Count to 12 (default 25 which is a bit much)

Will do.  I think this was the default profile

 

8 minutes ago, kmanstudios said:

 

 

These are just basic settings I use. Like I said, I would not be able to test against anything until next week due to commitments. And even then, I would print to just above your problem area to see if I get the same thing.

 

If it is not solved by the time I can get around to printing it, I will load my MH PLA PRO (Just not yellow....do not have that) and see if there is any correlation to your experience.

 

Slicing does look good. So not sure what the issue could be (difference in S5 machines, different stock of MH PLA PRO, etc)

 

[chuckmcgee 7]

So...just checking in for an update.  I am still having the issue, even after the great settings shared by kman.  But I also have some theories I want to share and solicit input about.

 

I decided to try my model again (not the same box as before, but designed the same exact way) and let it run (3 day print).  This is targeted to finish on Thursday at 2am, but just noticed that the same phenomenon has plagued this print.  I am still printing this on the 2nd extruder with the AA core (0.4)

• Ripples on the walls of those bottom levels
• Small indentation as if the lower levels are contracted beyond the normal perimeter beyond that point

I have made a couple changes to the print:

• I added a 45-degree chamfer to the bottom of the model
• I changed the filament to eSun Pro (a filament type I have used extensively on my Ultimaker 3 Extended)
• Changed layer height to 0.15 (versus 0.2 on the boxes I contacted you about)
• Added a brim

I have a theory that there is something wrong with my model as the walls may be too thick (requiring too much infill) for which I am only doing 10-20% triangle infill.  Is it possible that the heated bed is so precise on the Ultimaker S5 that it is melting those lower levels?  I wondered what would happen if I changed the outer shell from the default 3 layers to maybe 5.  

 

After giving this some more thought - I am starting to wonder if the issue isn’t the model and slicing.  The most profound “shift and ripple” came from the box I printed with 10% infill.  Remember that the walls of this model at 5mm thick.  As a result, that requires infill since there are only 3 layers in the inner and outer shell.  A 10% infill isn’t much for holding up the structure if the printer temps are such that it could be warping the print.  Things settle down as the model gets further away from the bed an you get to the actual inside floor of the box.

The two models which have the far less pronounced “shift and ripple” were done at 20% infill, including the 3-day print that is going right now.

 

I printed the same exact box on my CR-10 (with eSun filament and all the same temp settings) except the box printed perfectly.  This had an infill of 50%.

 

For the record, I have yet to try a higher than 20% infill on the Ultimaker S5 since the prints would take longer than I wanted to wait.  I may change my mind and attempt another print at a higher infill to test my theory - even if it means just printing the bottom portion of the model at a higher infill

 

Things I still want to do to troubleshoot:

• Print the box again but increase the inner and outer walls to 5 (from the default 3)
• Print the same model on extruder 1
• Print the box again on extruder 2, but increase the infill to 50%

[cjs 34]

In my opinion the errors/strange phenomenon is related to pressure changes in the nozzle. Some people use this effect quite artistically to paint pictures on objects. In the community it's known as velocity painting. ( https://www.youtube.com/watch?v=HdKaKKeyz7g&amp;t=435s  <- recent video explaining it. Different filament do show this effect more some less. Ultimaker uses quite a lot of different speed changes and velocity to reduce effects like ringing. I think on the S5 even more than on the UM3 as it is bigger. 

I would cross out the 2nd extruder as the cause of this problem. 

Instead I would open all speed settings of Ultimaker Cura (Walls,infill, skin, top layers) and set them to about the same speed. 

If this doesn't help I would turn of all the different acceleration and observe the result. 

You might want to create a small model which prints in about a few hours to test all the changes. 

 

I also read that some people do put a dummy model like a box beside the print to reduce the effect. 

 

[SandervG 1,521]

On 7/27/2018 at 2:06 PM, chuckmcgee said:

Not sure if this is an Ultimaker issue, Fusion360 issue, or Cura.  I am printing a simple box.

 

This is the model in Fusion360

 

This is the body in Cura:

 

But this is how it is printing (and this is twice now with two different roles of filament and different infill settings (20% on the first attempt and 10% on this one - each print did the exact same thing at the same point)

 

The perimeter at the bottom of the box is narrower then the rest of the wall as it prints, creating an uneven wall up the side.  I have looked at the layer preview in Cura as well as a cross-section analysis in Fusion360 and there appears to be absolutely no reason for that to be printing that way.  I also know there was nothing odd about how it was modeled...a sketch, extrusion, and shell.  Simple box.

 

I have also attached the STL file of the box.

control box.stl

 

Would love some advice about how to fix this.  Thanks in advance!

 

Hi ChuckMcgee, it seems like I can't view your photo's. Perhaps something went wrong during the upload. Could you try to upload them again? I would love to help and document your issue, but without the visual aid that is rather difficult. 

[chuckmcgee 7]

2 minutes ago, SandervG said:

 

Hi ChuckMcgee, it seems like I can't view your photo's. Perhaps something went wrong during the upload. Could you try to upload them again? I would love to help and document your issue, but without the visual aid that is rather difficult. 

Very strange...just fixed the images.

[chuckmcgee 7]

Here are some additional pictures of various prints:

 

First attempt on Ultimaker S5 with 20% infill - using MatterHacker Pro PLA (roll 1)

 

 

Second attempt on Ultimaker S5 with 10% infill - using Matterhackers Pro PLA (roll 2):

 

Third attempt printing on CR-10 @ 50% infill with eSun Pro+ PLA (this one was perfect):

 

Current print @ 20% infill, but with a 45 degree chamfer built into the design and a brand new roll of eSun filament:

Edited July 31, 2018 by chuckmcgee

[chuckmcgee 7]

Here is the new STL with chamfered bottom - no other changes

control box.stl

[SandervG 1,521]

Hi Chuck, thank you for the file and uploading your photo's again.

If I understand it correctly, the photo's in your original post was suppose to have completely straight walls or did the photo of the print also already contained a chamfer?

 

This shift and rimple you see is equally represented on all 3 sides of your controller box (not the back, since it is flat?) 

 

Did you try any of the speed changes cjs recommended?

[chuckmcgee 7]

9 minutes ago, SandervG said:

Hi Chuck, thank you for the file and uploading your photo's again.

If I understand it correctly, the photo's in your original post was suppose to have completely straight walls or did the photo of the print also already contained a chamfer?

 

All of the images in the original print have no chamfer at the bottom.  Just straight walls to the bed.  

 

9 minutes ago, SandervG said:

 

This shift and rimple you see is equally represented on all 3 sides of your controller box (not the back, since it is flat?) 

 

The shift and ripple was seen around all sides....except in the current print going on my bed (chamfered bottom edge @ 20% infill), the chamfered sides have the shift.  The front flat wall does not.  

 

9 minutes ago, SandervG said:

 

Did you try any of the speed changes cjs recommended?

 

I haven’t tried anything new beyond adding making the setting changes suggested by @kmanstudios, adding the chamfer, changing to a brand new roll and brand of filament, and setting infill to 20%. 

 

When the current print is done, I plan to try whatever the best recommended changes are. My inclination is to change speed and set infill to 50%...and to print only the bottom portion of the current model (for the sake of faster turnarounds between tests)

[kmanstudios 1,120]

2 minutes ago, chuckmcgee said:

My inclination is to change speed and set infill to 50%

?

 

 

[chuckmcgee 7]

and, if successful @ 50% infill, it begs the question about other changes I can make.  The model is pretty involved in it's final state.  Not easy for me to change wall thickness from 5mm (which is not intentionally 5mm...lesson learned).  At 50% infill for as large as the model is that is printing, it will take forever.  I also plan to try 20% infill, slower speed, and 5 inner and outer shell walls (from 3) to see if that helps to keep the surface integrity.

[chuckmcgee 7]

24 minutes ago, chuckmcgee said:

The shift and ripple was seen around all sides....except in the current print going on my bed (chamfered bottom edge @ 20% infill), the chamfered sides have the shift.  The front flat wall does not. 

 

Correction - even the flat wall shows the effect slightly.  You can kind of see it at this angle.  The wall making up most of the camera frame should be perfectly flat.

 

 

[kmanstudios 1,120]

Tonight I will get a chance to just do a partial print. I am using your files posted before. I will be using version 2 project file and will be printing from Extruder 1 as it is set up for the main plastics while my extruder 2 is always set for PVA.

 

Though it will not get fully printed by morning when I have to slap another job on the machine, it should get past your problem area.

 

So, even though I am using extruder 1, it will not be totally the apples/oranges comparison....more like apples to crabapples.... :P 

[chuckmcgee 7]

Love it!  Thanks a ton @kmanstudios!

[kmanstudios 1,120]

Ye be welcome me hearty :)

[jakepoz 2]

I'm curious: is that an aluminum bed in those pictures, or the standard glass one. Could that be contributing to more heat transfer to the part itself and therefore the folding of those layers?

 

@model_dude I saw your thread here: 

Is this what you mean by lower-wall cave in?