what sort of things do I need to keep an eye out for in models?

From this forum, I've learnt to break solids up into their surfaces and look for edges that are folded over.
but I still have some strange things happen like parts of models disappearing when preforming Boolean-unions.

What things have effect on Boolean function? (they seem to give me the most trouble.)
is there anything to gain from separating a solid into it's surfaces and running the "shrinktrimmedsrf" function over them? do the untrimmed points effect anything?

is there a "tips" post I should be looking at?

Cheers, Steve

Hi Steve,

> What things have effect on Boolean function? (they
> seem to give me the most trouble.)

Stuff like self intersecting surfaces (surfaces that fold back over top of themselves) will definitely cause problems with booleans.

Can you post the 3DM model file of the new situation where you're having boolean problems again?


> is there anything to gain from separating a solid into it's
> surfaces and running the "shrinktrimmedsrf" function over them?

Not usually if the surfaces are not mangled in some way. It could be possible that some mangled piece of a surface could be discarded after doing ShrinkTrimmedSrf but unless you have badly formed surfaces in your model to begin with it won't really have any effect. It's better to avoid making those kinds of self intersecting surfaces in the first place.

Also if you do ShinkTrimmedSrf, you can do it to a full joined object, you don't actually need to separate into surfaces first.


> do the untrimmed points effect anything?

Not usually - it can be useful to have the original surfaces since you can then recover the full surface if you want to restore it later on. But if the full surface is quite large in size compared to the trimmed region that is actually active it can reduce the size of the model data if you do a ShrinkTrimmedSrf on something like that.


> is there a "tips" post I should be looking at?

Well, the forum is basically full of tips and answered questions for all sorts of different models that people have worked on before. But a lot of times problems are pretty specific to something in your particular model.

- Michael

Hi Steve, also another general tip for making things more friendly to booleans is to avoid positioning 2 objects that are being booleaned in such a way that they have coincident surfaces between them that are barely skimming over the same surface area.

That makes the intersection calculations a lot more difficult for the booleans to figure out.

Instead of having barely skimming coincident surface areas, try to position objects so that they actually punch through each other by a little distance.

- Michael

Hi Steve,

I have struggled with this since trying out MoI. I love MoI but it takes a lot of patience and trial and error. I have found what drives the Boolean and fillet failures/variable results is the number of points that define the lines/arcs/curves/circles etc... Also, how those are broken up and joined together. Using show points and the rebuild command is the key. Then do all your ops and see how it works out. Repeating this over and over you will see how widely the results vary. Unfortunately it is almost impossible to make the same model twice unless you can remember every single setting along the way. Moreover, you have to have the basis for the booleans setup exactly the same way every time. This obviously isn't really feasible unless you have very few curves you are working with.

Right now, and I am still experimenting, I have found changing any arc to at least 9 points and circles to 36 points alievates a lot of problems. Curves sometimes have to have more points added same with elliptical segments. Lines are usually fine with the default 2 points however making them 3 points changes how all the boolean ops work. Also it is helpful to rotate solids to get the seam out of the way on occasion.

It is a really good idea to zoom in and look at all your small features. With lighting set to metallic you can usually see issues from a farther distance out. The pain of it is, you may not discover a problem until the last op, usually a fillet, and then have to delete everything, rebuild the curves and so forth and do it all over.

Rebuilding to a tolerance and rebuilding joined curves are both bad ideas. You can experiment to see why I say that. Sometimes reducing the default mesh angle from 10 degrees to 1 degrees can help correct graphical display glitches.

All of the above is my own experiences with MoI over the last few months.

Anthony

Hi Michael
here is a file, the object on left is my master, I used Transform-> copy to make the two copies on the right, when I try to boolean-union them , the bottom petals on the right most fleur disappear.


Also just remembered something else I wanted to ask, if I line up 5 copies of the master and flow them around the curve, and have to use "stretch" to complete the ring, a small bulge forms where the ends meet, if I design things so the ends come together has close has possible, will that reduce the bugles that form? or is it just a side effect of using stretch? ( I have started make my masters, with out the top and bottom rods and add them last , so after the flow there is only 1 join, but I still get a bulge. ( most likely, the bulge would not be visible if I actually had the ring made up, but the wife see it in the renders which is bad enough :) )

Hi steve,

> here is a file, the object on left is my master, I used
> Transform-> copy to make the two copies on the right,
> when I try to boolean-union them , the bottom petals
> on the right most fleur disappear.

Something appears to be damaged in the trimming boundary of that right-most fleur in your original master object.





See that kind of odd missing triangle there, something is messed up in the trimming boundary there, so that fleur will need to be replaced, maybe something like delete the surface and mirror the other good side over.

I'm not sure exactly how it got messed up in the first place though, your surfaces seem to be all ok in that area and it looks like you may have run into some kind of bug in the booleans there. I'll investigate it some more.


> a small bulge forms where the ends meet, if I design things so the
> ends come together has close has possible, will that reduce the
> bugles that form? or is it just a side effect of using stretch?

I'll need to see the actual bulge to really be able to give any advice on this part. But some types of bulges are a normal side effect of warping things to go around a curve - basically it works similar to an offset and when you offset something from a curve, in areas where the curve bends inwards it makes for a bunching type of effect.

- Michael

Hi Steve, one thing that is a bit odd about the object in that general area is when I delete these 3 faces here:



There remains this little tiny slivery surface here that was part of the model:




That's like a little tiny micro shelf, and those kinds of little slivery surface bits can have a tendency to mess up the object structure in various ways, those can get formed if the pieces that are being booleaned together are just slightly misaligned or skewed from one another where different planar pieces come together.

- Michael

thanks, michael

Once again,

I can send you a the file where I get the bugling, if you want to take a look. I think my biggest problem was trying to take short cuts, I get a better result, if I use flow on the fleurs and then add the top and bottom rails with either sweep or revolve. I guess I should only use flow when it's really needed :).

You have also made me realize that I need to go back to my library models and make sure they are clean of defects :)


Side note: I've been playing with MoI on a Dell tablet, windows 7 sucks has a tablet OS , but MoI is quite usable (the Dell comes with a Stylus, that make a real difference) I did try Windows 8 review on it, but Moi had some problems on windows 8, which is has one would expect, I think the problems where with webkit , second level menu selection could not be selected.

Hi Steve, it may take me a while to get to the bottom of why your original fleur had that damaged area in it.

In the meantime, try the attached updated version.

For this version I mirrored over the fleur from the other side and trimmed it into place, also deleting a couple of those little tiny micro shelf areas as well, and this version looks like it should now boolean union ok.

- Michael

Thanks Michael


I went back to my original template, luckily, I had all the solids used to make the fleur saved has separated objects, after playing around with it a bit, I found the micro shelf areas are formed when I boolean Union them together. If I deleted the micro surfaces, it would not join back into a solid, left me with some edges that did not match up, since I was not sure how to fix this, I just cut away the edges and made some patches, then it would reform into a solid. I hope no one will notice a .03mm patch :) I think I will just replace it if the one you created. I'm sure it will cause me less pain, long term :)

P.S, I've just had my first piece cast in Sterling Silver. nothing fancy, but it's something my wife wanted. is there a live post, where users post photos of real world object created from MoI designs?

Hi Steve,

> P.S, I've just had my first piece cast in Sterling Silver.
> nothing fancy, but it's something my wife wanted. is
> there a live post, where users post photos of real world
> object created from MoI designs?

Sounds great, I'd love to see it!

If you'd like you can post it as a gallery item on the MoI gallery here:
http://moi3d.com/gallery/

Or you can also just post it here in the forum, either as an attachment on this thread or you can start a new discussion thread for it if you want.

- Michael