Question Tubular Shapes, Different Quality Inside & Outside?

[rachael7 73]

I've had a recurring issue, that while fairly minor, is starting to affect the usability of a lot of my prints, and I'm stumped.  I'm really hoping others will have some insight. Thanks in advance for any help. I am using Cura 5.1, with a variety of printers, but mainly an Ultimaker S5.

 

What I'm seeing is on roughly tubular shapes, printed vertically or close to, that the inside and outside of the tube have very different surface finish qualities. Most of the time, it is the outside of the tube that gets the better finish - very high quality, in general - while the inside of the tube has really pronounced layer lines and substantial variability in dimension. In other words, on the inside of the tube, the walls just are not stacked very well, leading to a rough surface that can also be wavy, while the outside surface comes out smooth and uniform. I have had a few parts where the nice side and rough side are reversed, but I rarely get parts with both sides equally good. It seems to mostly manifest as excess material on the inside of the part, which I have been able to clean up in post-processing most of the time.  But I shouldn't have to do that, and on some shapes, that post-processing can be difficult and time-consuming.

 

To clarify, the parts in question are just vaguely tubular, in that they are a closed profile with a distinct inside and outside; they needn't be fully round or rectangular.  I'm seeing this effect on parts in the 25mm-100mm size range, so nothing tiny or large, just basic mid-size parts, with reasonably robust walls (on the order 3-5mm).  The roughness often seems to correspond to the presence of  other features, such as holes or bosses, even if those features are only on one side. So on a vertical tube with a horizontal hole in it for example, there might be a band of layers at the height of the hole where the rough side is noticeably proud or below the rest of the surface.

 

I'm fairly sure it is a Cura issue, as it seems to occur with any printer for which I use Cura to slice, as well as with different materials. My main printer is an Ultimaker S5, so I can also rule out compatibility issues, and I use the Cura profiles for the machine, which while not always perfect, are generally quite good; so it's not like I'm out in left field with bad profiles or anything. The inside and outside of tubes both get "outside walls", from Cura's perspective (note red lines on attached screenshot), so I do not see any obvious profile-related reason for one side of the tube to come out different than the other. I have only two ideas about the mechanism of action involved:  One possibility is that it depends on whether Cura starts the layer with the inside or outside of the tube. If there was a slight excess of material due to the flow settings, perhaps that extra material is getting pushed to one side or the other and causing the variability. The "Wall Order" setting should make it go the same way on each layer though, so that likely isn't the cause. The other thought I had was that it could be related to the variable line width - perhaps the wall position or "squish" of the wall extrusion varies a bit with width. That latter possibility really doesn't explain the difference between the inside and outside of the tube, but it could explain the change at layers where there is another feature (the other feature could change the line width calculations).

 

Any thoughts? Has anyone else observed this phenomenon? Has anyone seen it and managed to overcome it? If anyone is inclined to try printing to verify my results, I've attached a sample file which has shown the issue for me in both ABS and Polycarbonate. Thanks!

 

 

2013873475_GripPartialTall.stl

Edited September 8, 2022 by rachael7
Fixed Uploaded File

[Cuq 198]

Strange Infill Method :

Infill Density = 99%

Infill Pattern = Concentric

Connect Infill Polygons  = On

 

Is it realy necessar ?

[rachael7 73]

Good eye! Yes, I believe it is necessary in this case. This particular example is a highly stressed functional part, subject to both steady and impulse forces, so it needs a lot of strength. I've oriented the layers and infill to maximize strength in the most highly loaded directions. It is actually 100% infill though, I'm not sure why Cura renders it with gaps between the lines; I know Cura renders it that way when the flow rate is less than 100%, but I don't recall if that was the case in this particular profile or not. You do bring up a good point though, that I am running 100% infill, which is somewhat less common. Perhaps my infill flow rate is a touch high and areas with more infill, the excess material builds up and bulges out a wall. The bulge would end up on whichever side was printed last, I would think, since the material would more easily push to the warmer side. Interesting. Thanks for the feedback!

[Cuq 198]

There are always some discussions about Infill density versus print time and part strength . An interesting video :

 

 

Not sure that 100% give the best result,  Layer Height , and Wall Number are also important.  And one point I think  : Linear Infill is any way a better choice for stress distribution and adhesion between layers.

 

[Cuq 198]

It's also an interresting case where we can play with 2 new parameters on the 5.X release.

Minimum Wall Line Width  & Add Middle Line Threshold

0.34 /  75%

 

 

0.3 / 80%  

[rachael7 73]

2 hours ago, Cuq said:

There are always some discussions about Infill density versus print time and part strength. Not sure that 100% give the best result,  Layer Height , and Wall Number are also important.  And one point I think  : Linear Infill is any way a better choice for stress distribution and adhesion between layers.

100% infill nearly always yields the strongest result. It may not be the most efficient or cost-effective, but with rare exceptions, it is the strongest. More material equals more strength, just that simple. It is diminishing returns as you near 100%, but the work I’m doing is on parts that require strength which can barely be met by FDM, so I need every bit I can get. 
 

As far as layer height, my results agree with the video and I almost always print at 0.10mm-0.20mm. Minimum cooling, maximum acceptable extrusion temp, and hot chamber all also improve strength with the materials I’m using and I push those as far as I can as well. 
 

I disagree about the infill pattern though. You stated it correctly for low infill percentages - linear is indeed quite strong for infill percentages well under 100% and concentric quite weak in those cases. Grid is slightly better than lines, but both are good. As you near 100% infill though, the analysis changes and concentric just becomes more walls. So using concentric infill basically maximizes the number of walls for the geometry. 
 

One variable I haven’t experimented much with yet is strength versus line width, for a given layer height. CNC Kitchen probably tested that too. Someone must have, so I’ll search that up next. It becomes more critical as we consider these new line width parameters!

Edited September 9, 2022 by rachael7
Proofreading

[GregValiant 1,258]

In regards to the "different quality exterior than interior" I don't see that on my prints.  What I do see when the printer is working on islands is that the travel moves can leave marks on the interior as the nozzle moves from one island to the next.  Sometimes there are the Cotton Candy type of strings stuck to the inside walls.  Changing the combing doesn't help because there must be a crossing of the gap someplace.

[rachael7 73]

1 minute ago, GregValiant said:

In regards to the "different quality exterior than interior" I don't see that on my prints.  What I do see when the printer is working on islands is that the travel moves can leave marks on the interior as the nozzle moves from one island to the next.  Sometimes there are the Cotton Candy type of strings stuck to the inside walls.  Changing the combing doesn't help because there must be a crossing of the gap someplace.

 

Thanks for the feedback. I've seen the kind of issues you describe on certain filaments as well. I think that's pretty standard. This is something different. Walls that should be dead vertical with only fine hairs or little pimples come out wavy, with some layers sticking out further than others. It typically reduces the interior dimensions of the hole by about the amount of difference between the peaks and troughs, so it suggests it varies between correct and sticking out, rather than correct and sunken in. I'm beginning to think this may be due to the fact that I'm running 100% infill and if there is any error in flow rate, that excess material has nowhere to go except to bulge the walls. Do you do much printing at 100% infill?

[GregValiant 1,258]

"Do you do much printing at 100% infill?"

Rarely.  I'm more of a 25% Grid with the Infill Line Multiplier = 3 and I use the line directions to get the infill square to the walls.  Consider that at any density this will add strength in the X Y and Z where concentric at lower densities would not touch the outer walls and so would only add strength to a compression load in the Z.

 

I didn't change anything but the infill pattern.  You can see that Infill Line Multiplier doesn't work with Concentric.

[rachael7 73]

Thanks again. Yes, concentric infill is near useless... unless you are also using 100% infill, in which case it becomes more like additional walls and becomes the strongest pattern, in my experience. For most parts, a less than 100% infill is fine, but I'm really trying to squeeze every last bit of strength out of these parts, so I'm printing them solid. That does make flow rates much more critical though, so I'll have to have another look at that.

[GregValiant 1,258]

You are putting a lot of heat into it as well.  Your layer cooling blower should be moving some of the hot air out of the center of the part but it's still a chimney that will keep the bed heat in and might effect the inside structure.

[rachael7 73]

1 minute ago, GregValiant said:

You are putting a lot of heat into it as well.  Your layer cooling blower should be moving some of the hot air out of the center of the part but it's still a chimney that will keep the bed heat in and might effect the inside structure.

 

Now THAT is a very good point. I'm not using part cooling fans at all - ABS and polycarbonate both need the fans off, or very close to, in order to get decent layer adhesion. Both print on a very hot bed as well, so the chimney effect could be quite pronounced. Thanks for the insight!

[Torgeir 243]

Hi folks,

 

Could not agree more about "strength" and 100 % infill.

Also, if the object is compressed/loaded, -I'll always orient the force direction in the horizontal (X/Y) direction.

I've been printing with nylon quite a lot and using 100 % infill and closed up printer to keep the temp on the "upper" bed temp are the way to go.. 

Z axis are always the weak point when loading our fff built objects.

 

Interesting topics.

 

Torgeir

[GregValiant 1,258]

There was a poster here who prints custom ABS body parts for motorcycles.  One of the issues he encountered was too much heat within the chamber as the part got taller.  The printer was a one-off custom and the enclosure heater and fan system was designed to insure it stayed warm enough.  There was no control for the enclosure becoming too hot.  That required a rework of the enclosure system and he ended up having to drop the bed temperature during the print.  That worked out as the upper layers of a part would solidify rather than staying mushy due to all the heat.

[rachael7 73]

34 minutes ago, GregValiant said:

There was a poster here who prints custom ABS body parts for motorcycles.  One of the issues he encountered was too much heat within the chamber as the part got taller.

 

Interesting. I've definitely gotten too hot a couple times, but mostly only when I was down near the very hot build plate. And it was easily remedied on the S5, since its passive chamber heating is not really meant to do much. I've actually gone the other way at times, starting with a slightly cooler bed to prevent the part turning into a puddle of mush, then heating up to get more chamber heat, once the print has progressed a safe distance from the build plate.

[jirodriguez72 18]

Another thing I find, especially when printing higher temperature materials without cooling fans, is you can greatly improve surface quality by turning on coasting. I print almost exclusively in Nylon+CF or PC+CF with no cooling fans and I used to get a lot of zits, blobs, and stringing. The two biggest quality improvements I made were to turn down the temp on the hot end a bit (I'm still getting really good layer adhesion), and enabling coasting. I started at the default amount for coasting and gradually increased it until I got acceptable surface finishes.

 

This is a prototype pump impeller out of NylonX, it has a 10" OD, we spin these under load in water at 3600 RPM, and the speed at the periphery of the rim is basically 105 MPH..... so I think we got strength covered 😁. That picture is straight off the printer, only thing I will do is a little bit of light surface scrapping and dissolving the supports.