Help with Difficult Boolean

[trinitifx]

I have a model I created of a cap badge, similar to this:

https://www.atlantakilts.com/store/scottish-crest-cap-badge.html

I printed it out with my 3D printer and it came out great!

I got some high-heat resin and want to print out a mold of the badge to make a pewter version.

I can't seem to get a good Boolean subtraction of the badge on a box to make a mold.

All of the parts are separate meshes, I'm not sure if that is the problem or not.

Can someone else look at it?

Thanks!

[leeminardi]

It looks like "Hub City..." has already been subtracted from Box001 and HCH Box is easily subtracted from Box001.  What specific object can you not subtract from what other object? 

[trinitifx]

I figured it out-

or actually, I sent it to Netfabb and had them do the Boolean.

Perfect!

[trinitifx]

Ok, I spoke too soon..

Please check the attached. I've had to extrude several parts to make sure they completely cut from the MOLD (blue box)..

All parts should be subtracted from the MOLD (blue box).

If you select the MOLD (blue box) and add a pro Boolean modifier, you can subtract the STRAP and it works fine. If you next select the purple HCH TEXT in the middle, everything disappears. If you reverse this order, the same thing happens.

I've checked all the subtraction models with the STL Check modifier, they all pass.

How to get this Boolean to succeed?

[ads_czechoj]

At one point of creating the Boolean (I am using the Boolean instead of the ProBoolean), geometry disappears.

Please look at my attached gif. Is your issue with the ProBoolean similar?

[leeminardi]

I think the problem lies with the object HCH BOX.  I deleted the side faces of HTH BOX and then recreated the polygons for the sides of one of the H's as an experiment.  In creating the new side polygons I was careful to specify the vertices in a CCW direction to ensure that their normals would point outward (not sure if this was necessary by it doesn't hurt).  I was then able to successfully use ProBoolean to subtract the new "H" from the object that was the result of subtracting STRAP from MOLD.

I suggest you rebuild the side polygons of all the geometry of HCH BOX.

[trinitifx]

I get a similar problem (see attached)

[trinitifx]

Thanks! I will check on that..:)

How could this have passed the STL check if there are problems with the mesh?

Strangely, it probooleans fine if you select it first (see attached)

[leeminardi]

I recreated the walls of the "C" and it subtracts fine. 

[trinitifx]

Thanks!

I used XView and it shows I have flipped and overlapping UVW faces on the HCH Box (obviously)

Is the easiest way to fix that just to re-create the faces?

Selecting the polygon and flipping doesn't seem to do anything

[leeminardi]

"Is the easiest way to fix that just to re-create the faces?"

I think so.  It should only take a few minutes to create new polys.  Keep the top and bottom polygons and create new side polygons.

[trinitifx]

Ok, thanks guys for looking at that!

Using XView, I checked some of the other meshes and it looks like I've got some work to do.. Arg!

[ads_czechoj]

At least one of the objects being used in the Boolean operation has some interpenetrating parts, which confuses any Boolean computation you might try.

•  The object in your model which causes the problem is the "Strap OVER HANG" (on attached image).Solution to the problem: Consider to rebuild the "Strap OVER HANG" object.

There is few places on internet where you can find the answer to the question: "Why are self-intersecting polygons bad?"

One of the explanations: 

If the line crosses over itself to form a loop, then the left and right polygons get reversed for the portion within the loop. The list still shows polygon 1 being on the left and polygon 2 being on the right, even though that is not the case within the loop.

Thus if you try to fill a polygon or determine its neighboring polygon, the algorithm gets confused when a loop is present, and it will fill or indicate the wrong polygon

[trinitifx]

Thanks for that

It looks like the root of all my problems is poor modeling.. 😞

arg