Printing Convex Overhangs

[paradiddle65]

I've had some trouble with printing objects with convex overhangs. What I mean is the lower portion of a pipe if printed horizontally, or the lower part of a sphere for example. Anytime I attempt a print like this, the lower portions comes out looking more Vish than curved. I've played around with some support structure settings, but haven't gotten it looking acceptable yet.


I'll attach some pictures when I get a chance, but does anyone have any tips to print accurate convex overhangs?

Thanks a bunch

[Roxy]

Any time you try to print out in space where there isn't already a foundation, you are going to run into this problem. The problem is you have new filament being extruded that is just barely able to attach to the previous layer. It is very hot and as it cools, it will shrink and start pulling on itself. This causes it to warp upwards. There is no good answer for this, but different people go with different solutions. Some use a fan as the material is being laid down to try to help with the shrinkage. Some change the amount and type of the infill for certain types of overhang. Others use different materials that are less impacted by shrinkage. Some use beefier support material. And others (like me) avoid it if at all possible.

If you bring up Open_SCAD and put this code in, and compile and render you will get the attached .STL file. Probably, you can't print this without curling of the edges.

As you change the value for z=??; you will eventually be able to print with no issues at all. Right now I have z set at 83. My guess is if you set z=75; you will be able to just barely print the sphere. Hopefully, this will help you understand the problem!

z=83; difference() {
sphere(r=25);
translate([-20,-25,-z]) cube([60,60,60]);
}

Incidentally, this is not a 'convex overhang'. This is the exact opposite. But it was an easy way to show you the problem. We can do a similar little piece of code to help you see the limits of your printer doing 'Convex overhangs'.

[paradiddle65]

Thanks for the response Roxy,

Btw what does the z term represent in your code? I don't have much Open_SCAD experience.

[Roxy]

In that little snippet of code, the z=?? pushes the cube that gets subtracted from the sphere down. Objects in Open_SCAD are created at the origin. So a sphere gets created with its center at (0,0,0) which means the finished shape goes into all quadrants. But 'cubes' get created with its lowest left corner at the origin (0,0,0) and extends into positive space for (x,y,z). So the cube needs to be shifted left and backwards to be able to subtract off the correct amount of the sphere. The z=?? term controls are far down the cube is shifted.

If you put this into Open_SCAD and Compile and Render it, you will see what I'm saying:

difference() { sphere(r=25);
cube([60,60,60]);
}

[RobH2]

Thanks Roxy. I wanted to test out your code/test. I decided to push the envelope and use z=95 instead of z=83 and see what would happen. I got pretty good results. The only issue I had is that the attachment point on the raft worked loose and it began to wobble so I had to stop it. But, the shape itself is not bad. I'm printing is now with a few supports around the lower portion so it doesn't break free. Here are a few shots of the first test. I'll post more when this new print finishes.

y95-01.jpg y95-02.jpg

[paradiddle65]

To be honest I've never tried printing these features without support. I think its about time I print a few tests without support to dial in the optimal settings. I think a sphere is probably the worst case scenario so it'll be a good model to test with.

Thanks for the pics RobH2. After your second test finishes, if you could post what speed/ temp you're using, it would be much appreciated.

[Roxy]

Wow... That came out much better than I would have expected! There is no way I could have done that on my printer. With that said: It doesn't push things any harder to go past 85 the way the code is written.

z=83; difference() {
sphere(r=25, $fn=100);
translate([-20,-25,-z]) cube([60,60,60]);
}

It is taking a sphere with a bottom point of z=-25 and pushing a cube with a top point of z=60 down, and subtracting it. Once you push the cube down 60+25, you are no longer subtracting the bottom of the sphere off of it. If you use the code snip in this post with the sphere(r=25,$fn=100); it will tell Open_SCAD to have a much greater resolution on the surface of the sphere. And the amount of adhesion you get when you 'push things' will (in theory) go to a single point. Probably you will have problems as you approach z=85.

Those first layers did curl up, right? They look a lot thicker than the rest of the layers.

If you want to experiment.... You can use this snip of code with one extra line in it. It will attach a 1mm panel under the sphere so you don't need support. The plate should hold adhesion to the plate and as you approach z=85 you will be bridging out at almost a 90 degree angle.

z=84.75; difference() { sphere(r=25, $fn=100);
translate([-20,-25,-z]) cube([60,60,60]);
}
translate([-25/2,-25/2,-(z-60+1)]) cube([25,25,1.1]);

[RobH2]

To be honest I've never tried printing these features without support. I think its about time I print a few tests without support to dial in the optimal settings. I think a sphere is probably the worst case scenario so it'll be a good model to test with.

Thanks for the pics RobH2. After your second test finishes, if you could post what speed/ temp you're using, it would be much appreciated.

I used ABS at 230C and 30mm/s. The bed was set at 110C. I had two fans on it, one in front and one in back the whole time.

Wow... That came out much better than I would have expected!

I'll try one later today or tomorrow with the new code you sent. I have a 12-hour print to start right now. I did do a new test last night after my first one. I created some small supports to stabilize the sphere but the ones I created were still too small and they eventually broke loose. But I did get about 75% of the sphere. The first few layers did curl and get a bit fatter than the others like you speculated. It was interesting to see how faithfully the printer captured the facets of the sphere. The sphere didn't have enough geometry to make it smooth. I want to try to print a golf ball with dimples. That would be a good test.

Here are the photos from my second test last night. I had a jam last night and was able to mostly clear it without having to pull my head but I don't think it was totally clear when the print started. You can see the first few rings of the bottom of the sphere are messy. I think those would have been nicer if I hadn't had the jam. But, still, I got very good results.


y95-03.jpg y95-04.jpg

[paradiddle65]

Those look incredible! For me, warping continues almost up to the halfway mark. I think its time to duct my cooling fan to direct the air closer to the nozzle. That's a part of my printer that hasn't gotten much attention yet.

[Roxy]

It was interesting to see how faithfully the printer captured the facets of the sphere. The sphere didn't have enough geometry to make it smooth. I want to try to print a golf ball with dimples. That would be a good test.

The code at the bottom has the fineness of the sphere turned up so you will have a better approximation of the sphere. It will take a minute or two to compile.

y95-03.jpg y95-04.jpg

And adding that second sphere(r=20) command will save you a lot of plastic and print time. It will hollow out the sphere which will give you an extra data point. At the start of the print you will be trying to bridge out at almost 90 degrees and you will see how well you can do that on outer perimeters. Once you get up halfway, you will be trying to bridge out on inner perimeters. I have more problems with outer perimeters, but inner perimeters have their own set of issues. And that is why I cut a corner into the sphere. Corners cause me a lot of trouble on both inner and outer perimeters but in different ways. I suspect you are going to see a difference in behavior once you get past the half way mark.

z=83.75; difference() { sphere(r=25, $fn=100);
sphere(r=20);
translate([0,0,-30]) cube([40,40,60]);
translate([-20,-25,-z]) cube([60,60,60]);
}
translate([-25/2,-25/2,-(z-60+1)]) cube([25,25,1.1]);


If you want to see what the changes to the code do, take a look at the attached .STL