Preparing meshes for printing a negative form for a sculpture

Hello!
Can anyone give me some tips for that?
I have sculpted a complex mesh. Now I want to make a form for it so I can 3D-print all of the parts of the form. Afterwards, I want to pour plastic in the form so that I can produce multiple copies of the sculpture.
Are there any tools in Blender that help with doing this? Or are there other programs which can do this better?
So far, I could be able to do this in Blender but it is quite a struggle.
Thanks for any help in advance. :slight_smile:

Well, I assume the cast modifier could do this, don’t take my word for it though, I haven’t ever used it :stuck_out_tongue:
Maybe shrinkwrapping a very high poly plane onto the model?

Oh, or you could just cut it in half and 3D print a 2 part mold?

Thanks for your help! I will take a look at the cast modifier!
Maybe shrinkwrapping could be used at some point, but it will not do the trick if used without any further interventions.
Until now, I tried building the parts. Later on, I hoped that the boolean modifier will cut out the volume from the parts.
The thing is, there will be many parts because the sculpture has many intersections. That’s why cutting it into 2 parts will not work. Maybe 4 parts could work if I used some kind of rubber which bends when I try getting it out of the form but when it comes to plastic, I will need many parts.

You could retopo it, making sure to create planar loops where each cut would be. Bake a displacement map, pin the unwrap. Duplicate it. Delete half the verts and extrude a cube around it. Subsurf, Displace it and apply. Repeat with other half. Am I sure this will work, god no.

or you could just try to nuke it with Boolean. Actually, what’s wrong with Boolean?

Boolean doesn’t produce good results. Maybe knife tool could work better.


Could make the initial shape by joining the objects with ctrl+J or applying (actually pressing the apply button) on the added boolean modifier.
Then cutting it with the knife tool by enabling the cut through option (Z) and maybe also angle constraint © where the mold is going to be separated. After that, rip those edges by moving the mouse cursor to give the rip direction and then press V. Then extruding and filling the gaps, W -> bridge two edge loops is handy with this.

I know the example is not as high resolution as the real one would need to be, you couldn’t use that as a mold because it wouldn’t separate and there’s no keys modeled in the mold parts, but perhaps it works as an example.suzanne_mold.blend (824 KB)

I was able to get good results using boolean as long as I subdivided suzanne several (4) times. The cube is default. Have a look, maybe there is something wrong I’m not seeing.

I also uploaded this to shapeways (as a test obviously) as an .stl and it was accepted as printable.


Boolean seems to work better than expected. I made Suzanne mesh manifold, then used 3 cubes and snapping to make a 3 part mold without having to cut one cube twice. After that, removed doubles (35-38 removed from each).

As a result, there’s only one weird face which is good news because it’s easy to fix (misaligned vertex within double merge distance).


I then added keys, boolean difference and removing doubles after applying each boolean.


Again, example is not a usable mold but making one seems to be possible with boolean, when using it carefully.

Thank you very much for your help!! :slight_smile:
Well, I guess boolean is the best way to go then. I guess I will use some further tools like the knife in order to manage the whole sculpture first so that I don’t get confused. After splitting up the model, the single parts may be easier to build up.
I don’t really know how many parts will be necessary and maybe I will overlook some intersections, but I’m pretty sure my sculpture requires at least 20 to 30, possibly far more, single parts.
Do you think sunk-in numbers will help building the mold later on?
I planned using a new number on several colliding faces so that they stick to each other and so that I don’t get confused while putting the parts together. :slight_smile: You know, I thought of this as some kind of puzzle with the pieces fitting into each other.

EDIT:
The image shows just a little detail which will probably need 3 parts. The edited plane shows the early stages of two parts.

EDIT3:
The problem at “EDIT2” is solved now.
There really has been a non-manifold problem. There was a small hole in part 1 which could not be seen easily because the model overlapped part 1.

EDIT2:
Now I’ve got a new problem with the way I tried solving this.
When I try using the boolean modifier, it does not work right.
If you look at 02.jpg, you can see two parts I formed. Now I want to “delete” the volume of part 1 that is in part 2 (right now, part 2 intersects with the volume of part 1). I tried using the boolean modifier on part 2 as a Difference-operation and selected part 1 as object.
This leads to part 2 becoming the surface of the intersection of part 1 and 2. This is really weird. I tried making the normals consistent, but it does not work anyway.
When I try different operations (which do not make sense anyway, I’ve just been curious), the following will happen:
Operation Union:
Part 2 becomes part 1 without the surface of the intersection.
Operation Intersect:
Nothing seems to happen at all.
Maybe there is a non-manifold problem at some point? I don’t know. Actually, this should work just fine.

Attachments



So what is your status, have you figured it out?
Couple thoughts:

  1. It doesn’t make sense to me that any of your molds should overlap.
  2. booleans can be tricky. Often you need to hide the second object in order to see what has happened as it is covering up the boolean operation on the first. Imagine a union of a cube and sphere slightly overlapping. You use the boolean modifer on the cube, and choose the sphere as the object. you Select union, nothing appears to happen. Then you hide the sphere (eyeball in outliner) and you realize the cube is now a cube/sphere.

Other thoughts:
What material will be poured into the moulds? Specifically will it be rigid? If so you will need to take great care that no cross sectional area will be larger inside than the opening is. That’s the modeling equivalent of painting yourself into a corner.

I might suggest ordering a small two-piece test mould print at a very small size. No sense going crazy with a 40 piece set, if some how the medium fuses with the mould or has some chemical reaction.

But it sounds like a fun project. I’ve printed some jewelry in silver and steel for my girlfriend and sister (two different people :rolleyes:), and it’s great. Both pieces are pretty simple. I’m getting much better at sculpting and my next print is going to be something a bit more epic. Have a look.


Well, the problem that occurred while using the boolean modifier, has been solved. There has been a hole in part 1. causing the problem.

nothing appears to happen.

Yes. This can be tricky, if you don’t double-check everything.
I have planned using self-made plastic for the sculptures. Also, I am planning on using different pigments so that the plastic will have different colors. Maybe I will even be able to achieve a marbled effect. So far, I’m not quite sure about how I will deal with the colors. :slight_smile:
The plastic will be rather rigid. That’s why I use many parts because the finished sculpture will not bend nicely.
Actually, most people would use silicone for a mold so that it can get off the sculpture nicely.
But I don’t like that. I want to print the parts. The parts will be printed with a gypsum-like material.
As both, gypsum and plastic will barely bend, I have to be careful when I make the parts in Blender. Just like I have to be careful when removing the sculpture from the mold.
I know, I have started something rather complex for “trying out” the whole idea but on the other hand, I would really like to produce exactly that sculpture. :slight_smile:
When it comes to those intersections, I can tell you I have made some molds before manually. Those haven’t been that complex, though. :smiley:

if some how the medium fuses with the mould or has some chemical reaction.

That’s right. This could happen. Even though I don’t think it will as the gypsum will be varnished. But who knows! :slight_smile: Also, I guess, I will put some grease or foam into the mold before pouring the plastic into it.
But anyway, I am thinking about making a simple mold for trying it out. Just to make sure the complex mold doesn’t get damaged.
By the way, the jewelery looks awesome! It does look really professional! :eek:

Boolean requires that the objects are manifold - edit mode, vertex select mode (ctrl+tab), select none (A) and select menu -> non-manifold (or ctrl+alt+shift+M). If something gets selected, those are the parts you have to fix. It’s also good to remove doubles (select all, W -> remove doubles) and check the mesh after applying boolean so that the next boolean operation has some chance to work properly.


Forgot to mention, you should check the tolerances for interlocking and/or moving parts. In Blender, the space between two parts is virtually 0 and that doesn’t work in the real world, so you would take that into account in Blender. Otherwise, if you add keys to the mold and print it without leaving some space between two parts, they won’t fit together.
Also have some positive angle on the keys so that you’re not pulling or pushing against straight or negative edge.

http://www.pasteall.org/blend/23271
(I couldn’t include sculpted mesh and one of the steps, the file size was too big for upload, even compressed).

Edit: didn’t see your last post before posting this. I wrote something that probably is obvious to you but might not be for others so I don’t take those out. Besides, I want people to correct me if there is something to correct.

Well, if making a mold in blender is hard:
Print one (1) copy of the object and make a latex mold (or other material mold) from the actual printed versions, there’s more than one way to skin a cat you know.

CubeGodWell, if making a mold in blender is hard:
Print one (1) copy of the object and make a latex mold (or other material mold) from the actual printed versions, there’s more than one way to skin a cat you know.

Yes, I know. Personally, I’m just not a big fan of latex or silicone molds and I really like the idea of “printing molds”. :slight_smile: This may not be completely “reasonable”, especially when it comes to complex sculptures. :smiley:
Also, yesterday I have tried making a mold for a simpler sculpture. This worked out nicely. I was able to finish the two parts for that sculpture quickly. I am going to use that simple mold for experimenting with the materials.
So far, concerning my complex sculpture, I have 29 pieces for about the half of it. :smiley: I am going to post pictures from the virtual mold when it’s finished.

JA12
Forgot to mention, you should check the tolerances for interlocking and/or moving parts.

Yes, I have thought about that. But I don’t really know how much I should scale each part down. Maybe to about .999?
I don’t want the “pour lines” (I’m not sure about the terminology there) to get too big and\or too long, though. So I want to scale it down just as much as needed. :slight_smile:

JA12
Edit: didn’t see your last post before posting this. I wrote something that probably is obvious to you but might not be for others so I don’t take those out. Besides, I want people to correct me if there is something to correct.

This is cool! I’m pretty sure people can benefit from that post. The image shows the procedure nicely. :slight_smile:
How did you make one part of the mold half-transparent? Did you use material settings? I think displaying it like that can really help when working with parts of a mold.

EDIT:
I have just come up with an idea that may be very very helpful! But I’m not quite sure if it works. Especially, when using a high-poly sculpture, this may take some time to be calculated.
What I am talking about:
What about pouring virtual liquid into the finished virtual mold to check if it works fine?
Concerning the liquid, this may require some major fine tuning. But it may be very helpful so that one does not print a mold that does not work at some point.

Also, what about using the game engine and the parts of the molds as rigid bodies in order to see if one can pull out all the parts nicely? Again, I’m not sure if it works, but it could help a lot when trying to check the mold for any problems.

Check the print service you’re going to use. Shapeways has good information on their site


Yes, enabled transparency on the material and decreased alpha value. Then I enabled transparency display option in the object properties.


Maybe but fluid simulation in Blender requires tweaking and patience and I don’t think it would give much useful information. You would have to be able to simulate casting material you’re using to see how it behaves in the mold. While that might be possible, I wouldn’t try personally.

As for rigid bodies, latest versions of blender has rigid bodies available in the normal viewport, you don’t need game engine for that.
It’s easy to set up, select the static object and ctrl+shift+R to make it passive RB, select object(s) that move and ctrl+R to make it active RB, alt+A to simulate, done. You can even move the objects while it’s simulating. Rigid body settings are in physics properties and RB tools are on the tool shelf.
The difficulty comes from getting the collisions right for tight tolerances.

Alright, first of all, thank you very much for helping me and sharing your thoughts on this, JA12.
Unfortunately, I did not know a mold would be so expensive when printing it. As there are so many little parts which have to be massive, the mold uses a large amount of volume which leads to a higher cost.
Therefore, I did not finish the work on the mold. I can share a picture of the work so far, though.
So after doing some further research, I have come up with the following idea:
(This may not be suitable for anyone else who just wants to have the finished product. This procedure is just something technical I find interesting.)
1.: Printing a positive with a barrier around the object
2.: Printing an approximate two part mold
3.: Putting the positive in the mold and pouring alginic acid in the gaps so that the approximate mold and the alginic acid become a precise mold together.
4.: Taking out the positive
5.: Pouring liquid plastic into the volume the positive has been at

Has anyone got some thoughts on this? It may not work if the liquid plastic is hotter than the temperatures the alginic acid can take.

This is an image that shows the mold I have originally worked on:
(Some parts of the mesh are not shown so that it is easier to work on it.)


How would you take the positive (plug) out of the alginic? Did you mean like detail layer of the mold (that is cut) and reinforcing part of the mold? Not sure how big the model in the image is, which parts you were going to make as a mold (if any at this point), but it looks like two part mold is not enough.
As you say, the more volume your print has, more expensive it is. Making a solid mold or even a sturdy reinforcement would be costly if it’s big.
The only way I can think of keeping the cost reasonable is to model the mold parts so that they have:

  • enough thickness for possible rework/finishing if the print detail is not good enough,
  • ridges with bolt holes,
  • room for washers, and that the interlocking parts can be slided a bit and locked with the bolts.
    Then reinforce those individual parts so that the whole mold is rigid. The final mold would also need to accommodate the mold outside shape so that the mold inside surfaces can be easily reworked/finished if needed, have attaching points if it’s going to be rotomolded etc.

That might not work if the model is too big. It would be ridiculous (and expensive all in all) with print services that can print just small parts, or too expensive with services that can print large scale parts. In that case, it would be hard and time consuming but relatively cheap to just make the whole mold by yourself.

Edit: perhaps it would be better to contact those print services with an email and present your ideas (shortly), ask if the print materials withstand many castings, can be reinforced (and with what), and/or order a sample piece you can test with.
Also, google search brings up results for “3d printing a mold”. Could see what others have done.

EDIT:
Sorry, the next post has been posted twice for whatever reason.

Thank you for your reply, JA12!

JA12
Did you mean like detail layer of the mold (that is cut) and reinforcing part of the mold?

Exactly. As the alginic can bend, this sculpture could possibly made out of 2 parts.
I would have one part on the left side and one part on the right side.

Even after thinking a lot about this, there is still a fundamental problem. When I try the method I was thinking about, I need a hollowed positive shape that I can print. Blender doesn’t seem to be willing to do this, though.
I can hollow out the shape by using the boolean modifier. This is extremely time consuming as I want to print the sculpture very thin. And even when I really try, I am not able to achieve a nice uniform thickness on the entire mesh.
You know, if I use some rough shapes in order to hollow out bigger spaces within the mesh, printing it would still cost me about 1000 euro. I would really need to create a mesh which is build out of a thin layer that is just stable enough to pour anything in there, because after printing the hollowed shape, I can still pour wax or plastic in there so that it becomes more solid. The material I would fill this with are not that expensive. :slight_smile:
Oh, and about the size: It is about this big:
X: 24,5cm; Y: 47,5cm; Z: 29cm

JA12

  • ridges with bolt holes,?

What do you mean with that? putting holes in the mesh so that the dust can get out of the hollowed parts of the mesh?
Or are you refering to the parts of the mesh that hold the pieces together? If so, do you mean the parts between the two printed pieces of the mold or between each printed part of the mold and the alginic layer?
Also, I’m not quite sure what you mean with that:

JA12

  • room for washers, and that the interlocking parts can be slided a bit and locked with the bolts.

I planned using duct tape or sticky tape so that the fluid does not run out of the mold if that is what you are talking about.

JA12
That might not work if the model is too big. It would be ridiculous (and expensive all in all) with print services that can print just small parts, or too expensive with services that can print large scale parts…

As far as I know, both, small and larger parts can be printed at the place I let this print.

JA12
Edit: perhaps it would be better to contact those print services with an email and present your ideas (shortly), ask if the print materials withstand many castings, can be reinforced (and with what), and/or order a sample piece you can test with…

Well, I have asked about the things I try to print and I was told that no one has ever printed a mold there so it’s not sure if it will work. So that’s why I will test this using a cheaper mold first. (I am currently working an a less complicated and smaller mold I want to print.)
I am going to do some testing on this with materials then, before printing anything of the taller sculpture. Also, when I am ready to print the taller mold, I will ask someone who is specialized when talking about molds. This is less risky than trying something big and failing because of some intersection or something.
Hopefully, it will work so that I can show you some pictures of the finished mold and the finished sculptures. :slight_smile:

EDIT: I have added a picture of how I think this might work. The green area is supposed to be filled with the alginic and the white lines show the printed mold which is hollowed out. The white plane shows where the two parts of the mold are supposed to be separated.


EDIT2:
Of course, the plane is supposed to be straight. :slight_smile:

You could try solidify modifier. You have to have consistent face normals though, since it needs the direction information.


I was talking about detailed mold that consist of multiple parts, thinking out loud how to keep the print cost low (less volume) and still end up with a usable mold that could be used for casting. Low volume print would not be strong enough on its own, but if you could add support material and hold the separate pieces together with screws, it might.

Here’s an example of what I was thinking:



(Of course it would be good to model in some support structure. The detail parts in each mold piece are thicker than the rest though. Perhaps even a container for each mold piece that could be filled to make them more rigid).

I have hard time justifying the print cost for both the plug and the mold, because if the only function for the mold is to support the detail layer, you could do that yourself using cheap materials.

Thank you for your reply, JA12!
I have tried using the solidify modifier. There will be a problem when I use this, though.
I have added a picture of the problem that occurs when I try using the solidify modifier for this complex mesh. (A)
Also, I have tried duplicating the mesh, scale it down and place the second mesh in the original one. This doesn’t work either with this mesh, though.
When I use them on simple meshes, both ways work just fine. When trying that on complex meshes, it will not work. This has probably something to do with the faces become “negative” (smaller than scaled down to zero, then becoming bigger the smaller you try to make them because the magnitude rises again).
I have also taken a screenshot of scaling (als+s) down the mesh. (B)

Oh, then I guess, what you are thinking about is not all that usable for the sculpture I want to print a mold of. This may very well work for a mesh like you posted because of limited intersections. Then, this may even be the best method when the material can be reinforced.
When it comes to my mesh, I would have to print about 60 parts so that it would actually work. Most of the parts would have to be rather small. So small that I can barely take volume out of it because if I do, it may break easily and I guess reinforcing small parts is very difficult. Besides, I would have to hollow out about 60 shapes. :smiley: If possible, I try to avoid this.
If you mean the mold I was talking about before, this is something completely different. I just don’t know because you wrote:

I was talking about detailed mold that consist of multiple parts

On the other hand, the pictures seem to show what I was talking about before: a mold that is more abstract which could be used to pour alginic between this mold and the positive print of the sculpture.

I have hard time justifying the print cost for both the plug and the mold, because if the only function for the mold is to support the detail layer, you could do that yourself using cheap materials.

Well, I would have to print the positive mesh anyway. And if I can hollow out the more astract parts of the mold nicely, I guess the positive print would be much more expensive than building the mold. So when comparing to the basic cost, this doesn’t add too much.
But I know what you’re talking about. The mold could be done with cheaper materials, too. Besides from just being interested in the technical procedure, I cannot justify this either. :slight_smile:
This may be a nice way to getting molds if 3D printing gets cheaper, though. :slight_smile:

Hopefully, this can be done as I plan it. :smiley: Anyway, I will tell you - even if I get a big mess. :smiley: