Hi niko, here's an example of how to use that technique with a curved rail, this is how to build the second example I posted entirely in MoI.

I started by drawing a polyline snapping its points on to the rail like this:



Now sweep along the polyline, it will build a mitered result like this:



Select just the edges around the original juncture point, do a Ctrl+C and then Ctrl+V to duplicate the edges as regular curve objects, and then delete the segmented sweep, giving you this:




Now use the Transform > Orient command to copy the starting rectangle to the very end, or actually you could do it by duplicating those edges from the sweep as well:




Now it is ready to sweep, use Edit > Separate to break the rail into individual pieces, and do the sweep in sections, first one sweep here:



Then a second one here:




That's how I produced the attached 3DM file created only in MoI.

- Michael

Hi Michael,

>>That's how I produced the attached 3DM file created only in MoI.<<

I presume that result is just for fun, as it is a very bad output.
You will not produce a good result attempting a single angular bisector/trim at the 3d intersection

Hi Steve, sorry I don't have any idea what you are referring to - that's not a bad output at all, it's a full solid and seems to check out fine, no self intersections, etc...

Also I don't understand why you are talking about an angular trim - the sweeps in this case touch each other exactly because they come to a common angled profile.

That's the whole point of this method - to make a joined touching piece without using any trimming.

You have to use some method like this to make pieces join in this kind of situation because 2 regular shaped tubes will not really come together in a simple miter like that when the paths are all curving around in 3D space, that's what I've been describing in several previous posts in this thread with several examples.


Using an angled profile shape that is common to both sweeps, does make a result that touches, at the expense of it being a slightly slanted tube and not completely even thickness throughout.

If you think it is a bad result, you will need to explain what your complaint about it is - as far as I can tell it's all fine, for example here I have done a boolean in that area with a sphere and as you can see it gives a clean result:





- Michael

Hi Michael, just because a solid is produced does not make it correct lol.

Take a section of the solid to show profile.(which is far from correct)

Hi steve, actually that is totally correct for this case - the whole point of that result there is to generate pieces that touch each other at the sharp corner in the rail.

The purpose of that particular model is not to make a tube that has a uniform cross section.

Read back over some of the posts in this thread, I have some explanations in there.

Geometrically, you can either have a totally regular shaped tubes with even cross sections in them that do not come to a simple miter-like joint, _OR_ you can have a non-uniform tube that does come to a common simple joint which it does by making a somewhat slanted shift to the sweep.

Just because something has non-uniform thickness does not mean that it is not "correct" if it is solving some other problem that someone has put a higher priority on, like in this case coming to a common angled profile.


If you think it is possible to have both a regular shaped cross-section, _and_ at the same time have the pieces meet in a simple common miter-like joint, then please show an example.


- Michael

Also you could probably reduce (but not eliminate) the distortion by putting in a non-slanted profile like the starting one at some middle point or something like 3/4 point along the sweep so that it would hold more of a regular cross-section through more of its length and only start to shift closer towards the end.

But I think you've missed the point which is to make something that looks like a miter joint at the sharp juncture in the sweep path.

- Michael

I thought the point was to get the PROFILE swept around the curve with a mitered corner. There does need to be compromise but at what expense? Having such distortions in the profile IMHO is not good and I would follow more of a path of what NX(and other CAD) do by averaging the intersections.

This is a very(very) quick example that is more to what I would expect, although of course there is still compromise.

Hi Steve, well that approach of shifting control points just at the end as NX does produces bad geometry errors in other situations, see this previous post in the thread where I describe it in detail:

http://moi3d.com/forum/index.php?webtag=MOI&msg=3575.11

Is that kind of self-intersecting geometry what you would consider a proper result?

Not only that, but taking a section through the result you posted also shows distortion as well:




Right now it seems like that is not the best way to go, it seems to be a more risky approach with it making bad results on shorter pieces, although it does help with making the distortion located more at the end on longer pieces.


- Michael

Hi Michael,

>>Is that kind of self-intersecting geometry what you would consider a proper result?<<

I put forward a possible way to work, not that intersecting geometry produced from that are correct. You will have problems with this type of construction, but I still say that attempting such miter based on one intersecting bisector will not work well.

Here is that model I posted(before) again, but with the accuracy upped a little to attempt to produce a more straight profile.

- Steve

Thanks Michael

hi Pilou,

This models is composed mainly of boolean. It is very easy to make a curve, an extruded, and remove what is too much. Moi3d has very strong boolean I think. I use a lot of intermediate lines to produce the volumes, but it's surely the principle of this type of modeler. I have other modelers, however, it is very flexible and allows more creativity I think.
I'm french, I usually keep the interface in English, I make an exception this time ;-)

Temporary Curves:

@ ppj: L'interface française fait gagner 11 pixels de large sur l'espace de travail par rapport à l'anglaise! :D

French interface saves 11 pixels wide workspace compared to the English! :D
:)

Hi Steve,

> but I still say that attempting such miter based on one
> intersecting bisector will not work well.

I don't quite follow - it seems to be pretty clear that sweeping between angled profile curves does work better on the situation that was originally posted.

That's the one that was attached to the first post of this thread here:
http://moi3d.com/forum/index.php?webtag=MOI&msg=3575.1

That's actually a more simple situation than the one with the long curves.

I'd say that it is more important to generate a tolerable result (without any badly mangled or self-intersecting geometry) for the original case than focusing on generating a more uniform thickness result for the 2nd case which is more of an exaggerated test case.


I've attached the result of using this technique on the original posted model, it creates a result that has no self-intersecting areas, so it can be used for things like a boolean with the main hull.

The technique that you are saying is better makes for badly formed self-intersecting geometry here, which would likely cause boolean calculations to fail. I would have a hard time convincing people that is a better result than the attached version.

- Michael

Hi Michael,

>>I don't quite follow<<

That is obvious.

You are still working from a single bisector, although you can get a result of a solid, that does not make it correct. It may be OK for hobby stuff, but the model you posted, it would be rejected in my work, as there is error of between 0.025/0.05 in the position of the angled profile (I only bothered to check the first.)

As you now have it in your head, due to the bad geometry posted from the output of NX by another member, and you appear to believe that calculations can only be made from a single bisector, I will not waste my/your time further.


- Steve

Hi Steve, I'm sorry but you're not making very much sense here...


> It may be OK for hobby stuff,

Wow, you mean anybody else other than what you work on must be doing "hobby stuff" ? That's a quite patronizing attitude.

There are all kinds of people producing professional work that would like to simply get a good looking result for that case. Just because they want something quickly that looks good does not mean that they are doing "hobby stuff".


> but the model you posted, it would be rejected in my
> work, as there is error of between 0.025/0.05 in the
> position of the angled profile (I only bothered to check
> the first.)

So would the result posted from NX with the self intersecting geometry be accepted as a proper result for your work instead?

Or is there no CAD program that can automatically generate a result for that which you would consider acceptable? If so then MoI is no worse off than any other CAD program since none of them can do what you seem to want there.

You will probably need to do some custom modeling instead of using an automated tool to get what you consider to be correct.


> As you now have it in your head, due to the bad geometry
> posted from the output of NX by another member, and you
> appear to believe that calculations can only be made from a
> single bisector, I will not waste my/your time further.

Well, it's not just NX in particular, but rather that whole technique of morphing just the end control points of the sweep to the miter that is prone to that kind of self-overlapping error.

It's something that would be quite difficult to avoid with an automated algorithm without spending a really long time with heurstics or almost artificial intelligence type of analysis of the spacing and shaping, etc...


It's really curious that you say the method I used is not good, yet here is a simple case where it produces a result that the original poster would have most likely considered to be a great result...


- Michael

Great model ppj, it's cool to see an other modo user enjoying MoI booleans power :)

Hello Michael,

>>Wow, you mean anybody else other than what you work on must be doing "hobby stuff" ? That's a quite patronizing attitude.<<

I did not mention anyone else, please show me where I did!
I think overall that MoI is no good for my work, MoI output is not accurate enough and when the models are taken into the other packages I use there are many errors, yes, the errors are (usually) within the tolerance of MoI, but not for being acceptable for my work (I do not make the rules of what is and what is not acceptable). That is not patronizing, just a fact.

I find it amusing that you jump between examples of this mitering and then just post an acceptable result from the first posted by the OP

Why did you post the example (http://moi3d.com/forum/index.php?webtag=MOI&msg=3575.4) I was replying to (and I though we where discussing), and then not post an example showing a good result with you method?

The first rail/profile posted can easily be done with the various CAD systems I use, even Rhino can give a good\usable result with 5 mouse clicks.

Anyway, have fun.

- Steve

Time to close this thread I suggest. It's just going around in circles.

Nigel / 3DKiwi

Hi Steve,

> I did not mention anyone else, please show me where I did!

You said that the result generated "may be OK for hobby stuff" - however, the result that I posted would be ok for a lot of different purposes, like someone more focused on just quickly getting a model for rendering, or for some design like a piece of jewelry that does not have to be exactly tubular to a tight tolerance, or a ton of different things.

Your statement is basically placing all these other kinds of people's work as just "hobby stuff", even though what they are producing is actually a part of their real job, it's just different from your job.


Of course, I can certainly understand if MoI is not the right tool for your job, if you have specialized requirements then use specialized tools.



> Why did you post the example
> (http://moi3d.com/forum/index.php?webtag=MOI&msg=3575.4)
> I was replying to (and I though we where discussing), and then
> not post an example showing a good result with you method?

Well, there were various different parts to the discussion earlier in this thread, that particular one I was giving an example of how you can't actually have a completely even thickness tube that has a common miter-like joint between 2 pieces that are curving around all over in 3D.


> The first rail/profile posted can easily be done with the various
> CAD systems I use, even Rhino can give a good\usable result
> with 5 mouse clicks.

One that matches your definition of within tolerance and not the "oh-so-terrible" result from sweeping between angled profiles?

If so then by all means show it.

If you do a regular 1-rail sweep in Rhino it will produce this result:



If you enable the "Untrimmed miters" option, then it will produce this result:




Note from the slanted isoparms there, that this is a sweep between angled profile curves, which is the kind of result that you had earlier said was totally unacceptable.

Is it suddenly acceptable now?

- Michael

Hi Steve,

> I think overall that MoI is no good for my work,
> MoI output is not accurate enough and when the
> models are taken into the other packages I use there are many errors,

Just curious, what other packages do you use ?

I use NX6 at work and I have done various models in MoI with quite good results when importing into NX, we work to an accuracy of 0.005mm(0.0002") and 99% of the time the models from MoI pass the tolerance check amongst the other surface checks which do well most of the time, if a surface does fail it doesn't take much to fix.

I don't know how complex the models are that you are bringing in from MoI to your other packages but this model I made 2 years ago in V1 passed the tolerance check at 0.005mm, there were a couple tiny objects and a couple of self intersections which where caused by the fillets which took me no more than 5 minutes to fix.

I believe working below an accuracy of 0.005mm is impractical in the real world unless you're into nanotechnology which is an entirely different world altogether.

If your getting a lot of errors with your models maybe try another modeling technique, like, I found I get better results if I use basic curves and arcs where possible instead of a freeform curve where you don't know what degree it is or how many segments it has.

I actually purchased MoI for home use and truthfully, I didn't expect a super accurate solid modeler but in the end it has preformed better than my expectations, I can even say it does better in a lot of areas than some mainstream CAD modelers that cost 10 to 20 times more.

Cheers