Inside overhang problem

I have been trying to print a bell model and have been coming across some problems with the inside overhang of the top. For conomical use of materials and time, not to mention better finish, I have modeled this as a solid but sliced with no fill, no top or bottom. The angle of overhang on the top is 68 degrees. Although this overhang is severe it is quite printable on the outside of anything. On the inside however, it pulls away from the print while printing - this can be seen on the photo below.

Any best recommendations as to how to do this best?

Details are:-

ABS, 230 head 110 bed.
Size of bell is 54mm at mouth, ca 30mm at top shoulder
Sliced by Slic3R, no infill, no top, no bottom 3 shell layers of 0.15mm thick 0.4mm wide (0.25mm extruder)
Speed is 50mm per sec with outer at 50% speed. Fan switched off throughout

Turn on support material (will use a lot of plastic and slow things down, but will all be inside the bell), set the slicer for multiple top/bottom layers, try thinner layers up near the top (though you're already pretty thin).

An alternative would be to print it upside down with support turned on- then there would probably just be support on the outside, right under the area you're having problems.

I want to avoid supports because of the material and time. I did try upside down but on the small base and the fairly hot bed the ABS was flexible enough to have the top swinging about.

This is a pro bono design for a school and I want to give them the best recommendations for successful printing. I have a number of possible solutions including using PLA, having more shell layers, hotter or colder extruder, turn fan on for these layers etc., but is there anything important I have missed?

Yes, using support material will cost a few cents more and take longer to print, but probably not as much or as long as printing multiple failures.
Some designs are not well suited to FDM printing. If you don't/won't use support material, this is one of them.

In this case, the best recommendation for successful printing is to turn on support material.

Edited 1 time(s). Last edit at 05/15/2015 08:20AM by the_digital_dentist.

Is the inside needing to match the same profile as the outside? Instead of designing it solid with no infill, you could make it hollow and taper the inside so that the overhang isn't as sharp of an angle. You probably could get away with just having a taper the top 1/3 of the total height and not noticeable unless you flipped it over and looked inside.

Hi cdru. I have tried something like that and at least is was printable, but I still get a few problems. The surface quality of something with 0% infill is better than any defined thin shell and the speed is faster. The poorer quality seems to be because there are variable number of shells if it has a defined thickness - I am using 1.2mm with an extrusion width of 0.4mm and get from 3 to 5 shell layers and there are even gaps in some of the layers. This all results in more prominent layer artifacts. As far as the speed goes, the combination of the variable number of layers and both an inner and an outer face to be run at reduced speed. The overall effect is:-

Solid with 3 shells, 0% infill no top or bottom uses 2681mm of 1.75mm filament and prints in 56 minutes
Defined 1.2mm wall and an internal 45 degree sloping wall used 3593mm of filament and printed in 1 hour 32 minutes.

Which would probably be O.K. if it looked as good.

have you tried slowing it down on those particular layers

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

The support problem was solved as cdru suggested but the finish was marred by having a changeable number of shell layers. This was caused by me having a failure-to-think. Any slicer is only concerned by the thickness that it sees in the layer that it is slicing so that the wall thickness for Slic3r is in the XY plane, as shown below on the right, not on the left.