I’m trying to make an insert for a planter, to reduce the size of a fill bin and make it easier to plant smaller quantities of seed.  The problem is that the bin is too big, and doesn’t allow seed small amounts of smaller seed to ever be picked up by the pickup wheel.  In the distant past, seeders often had a small insert that fit into the bin to reduce the space and push smaller seed better up against the pickup wheel.  I’m trying to make that kind of space filling, removable insert for this seeder.

First attachment is the bin I’m trying to partially fill.  Second attachment is the surface I started with to fill that, third attachment is a difference between the first surface and a second four units below it (I used two surface calls on two files, but probably could have done a difference on a translate of the original.)  It makes artifacts all over the place below the bottom surface.

I’m completely new to 3D printing and OpenSCAD.  I figured it would be a total pain to come up with some kind of formula that produced the shape I wanted, but surface sounded great.  But then I thought I’d be wasting a whole lot of resin making it a solid cup, so what I really want is a thick-enough-to-print-but-thinner piece, and came up with the top and bottom surface, and a difference between them.  Very close to what I’d want, but I’m concerned the artifacts mean I’m headed the wrong way.  Suggestions?

I think that’s my problem, but a lot of what you wrote doesn’t make any sense to me.  In your example, where the walls of the two cubes meet you get problems.  But these are two surfaces that don’t meet anywhere.  So I would make the bottom larger - where?  I could see your point if they overlapped, but they don’t.

I think the part I was missing is that surface creates edges for you that you didn’t actually specify.  So, if you make a surface of 10 x 10 with a height of 10, you’ve essentially created a 10, 10, 10 cube translated up a bit.  It’s clear from what actually gets rendered, but not clear from how it gets discussed in the manual.  In my example, all of the vertical edges were created by the rendering engine, even though I didn’t actually say they were there.  So when I make the bottom edge, they shared all of the vertical edges, which was your point.

Funny, but scale didn’t seem to work.  I tried scale([1.1, 1.1, 0]) and seemed to completely lose my bottom edge.  I got the whole thing  - oh, crud.  I bet I did do 1.1, 1.1, 0.  That would flatten it to a flat plane at the bottom, wouldn’t it?  That’s what I got.  Shoot.

Well, anyway, I modified my python to make my bottom surface extend one place in the X and Y dimensions farther than my bottom, and that did the same thing.  Looks much better now.

Is there a better approach though?

Very little to the code.  It’s all in the surface dat file.

//surface.scad

difference() {

surface(file = "testsurface1.dat", center = true, convexity = 5);

surface(file = "testsurface2.dat", center = true, convexity = 5);

}

“testsurface1.dat” has the base surface.  “testsurface2.dat” gets generated from the first one, now with extra values to get rid of the artifacts.