I cannot for the life of me figure out why the rest of this blend cannot be completed.
I know it's not MOI's problem and I'm not reporting this as a technical issue, I just want to know why it happens and what can I do to fix it(meaning what's another way to achieve this result).
I split the edges to help with the blends but that apparently did not help either.
Please, please do not tell me that this is easier done in Sub'ds...I already know that and I prefer the faster workflow of nurbs.

Thanks in guys advance

Hi niko, if I turn on surface control points of that middle piece, I can see that the area that you're having problems with is a collapsed down "singularity" of the surface:




It's quite easy for such collapsed down areas to have pinches in the actual surface shape, sometimes the pinching can be in a pretty small area too. That kind of irrregularity or bumps or little lumps in the surface shape will tend to have a noticeable affect on blending, since blend is trying to adapt itself to the surface shape.

If you look at the isoparms in Rhino, the middle isoparm seems to bend off towards one side right near its end, there is probably some undesirable sharp change in surface curvature right in that area.

- Michael

This is a screenshot of a high density mesh export from MoI (angle = 1), you can kind of see a bit of the change in shape right near that collapsed area:

I guess that's because I swept a profile with two converging curves.
Would the optimal way to create that surface be a network if I split the converging curves into 3 (for a total of 4 curves for the network)?

Hi niko, network can actually have a pretty similar shaping problem in areas where things are collapsing down to a single point.

It's kind of a general geometry problem in a certain sense - when shapes are shrinking down if there is any irregular spacing like different distances being traveled on either side (which is hard to avoid), the differences tend to create irregularities in the collapsed down area. There tends to be too much difference between the area immediately away from the collapsed point, creating twisting or torquing pretty easily.

To get the best quality surfaces you kind of want to avoid collapsing if you can - collapsing and pinching kind of just tend to go together.

If you can make an extended surface that does not have collapsing in it and instead punches right through the other surface, that can be more robust. You kind of give up some amount of "direct control" with this method though because you rely more on how a surface intersects with another one to give you your outline rather than having drawn it directly.

I have an example of how this would work in your situation which I'll post shortly.

- Michael

Hi niko, also you wrote:

> Would the optimal way to create that surface be a
> network if I split the converging curves into 3 (for a
> total of 4 curves for the network)?

I think you're talking here about having a layout where the "rails" have shared tangents with the "profiles" ?

That tends to have another kind of semi-related problem - it's more like uneven "stretching" as things approach the shared tangency corner.

Also having a single NURBS surface that has tangent corners instead of distinct corners tends to also degrade the mathematics of the surface evaluation in that area as well - the definition of the surface normal is usually being the shared perpendicular (called the "cross product") between the u and v directions at any point on the surface. When the u and v directions are tangent with one another it is a degenerate situation where there are a whole lot of possible directions perpendicular to both instead of just one direction perpendicular to both.

So what you are talking about there can tend to have its own kinds of problems as well.

- Michael

Hi niko, here are some steps to show you how this alternate "extended surface" technique can work.

I've attached the result 3DM model as niko_protrusion.3dm .


To start with, I projected your smaller outline curve on to the surface, and used Trim to hack off the ends about where they started to curve in:



Then I went to the other side of the surface (the "inside" part of it), and drew 2 lines in at the ends of those curves so those lines are going inward towards the center of the shape:



Then I used the Move command on each of those lines, picking the line's own endpoint as the base point, and the line's own midpoint as the target point, so the lines now had some spacing off of the surface. I also trimmed back a little more off of the surface curves as well:



Then I put in curve blends and joined those pieces together to make rails that start off hugging the surface but then curve inwards towards the center of the shape:



Now a sweep can be done to make a surface that looks like this:



Or as seen from the outside, it looks like this:



So notice there that the rounded "tip" shape is a result of the sweep being submerged as it passes through the outer surface, instead of trying to build a surface directly to that outline curve.

Now with these 2 surfaces you can select each of them and use Fillet to build a surface/surface fillet to make a smooth juncture between them, you can use the G2 blend option for the fillet shape if you want a that kind of blend-like result:



That's an alternate approach to this problem which builds an extended surface that then gets trimmed back by the fillet, rather than trying to build it directly to a curve layout that does not have a regular non-pinched quad type structure.

You give up some of the "direct drawing" control over the shape of the tip with this technique, you have to manipulate that tip shape kind of indirectly by controlling how fast the surface submerges through the main one. But you're going to actually have much more regular surface structures with this way and you also gain some simplicity with a smaller number of surfaces involved, like just one fillet/blend surface to bridge between the pieces rather than fragments.

Also you can adjust the radius of the fillet which is kind of like adjusting the spacing between the 2 offset profile curves that you were using.


I hope this alternate approach may be useful. Let me know if you try it and get stuck on something.

- Michael

I have to say again, this is the kind of stuff I feel needs to be in a Tips & Techniques section somewhere in the manual.

It's never about how each tool works - that's easy, it's about why, when and where to use each tool - I learn from these kinds of threads every time!

-Will

Hi Will, well this is a good example of a technique that is tied to a particular shape that someone is asking about.

Those kinds of tips that involve a particular model are are kind of easier to do when the person is able to present their particular shape that they are working with.

The manual isn't interactive like the forum is, so these kinds of tips just generally happen on the forum more than in pre-canned written material.

- Michael

Thanks Michael.

Yes I'm aware of this method and it works, but as my example shows it calls for situations where an adjustable fillet would be necessary (otherwise the filleting in this case is almost non existent).
Hopefully this is planned for V3 along with the continuity options for network and sweep commands.
I also have included the version where the fillet is done in Rhino using the adjustable fillet via add handles.

Hi niko, in the attached model I have your Rhino created one on the left, and a regular constant distance fillet one created in MoI on the right.

Here is what it looks like shaded, I guess it is hard for me to see much of a difference?



I do want to add a variable radius fillet into MoI though. I'm not exactly sure when that will happen yet.

- Michael

Michael
I don't know how you did it but my version of MOI cannot produce the fillet you just did.
How did you do it??????????

Hi niko, I untrimmed everything, and moved the small one a little tiny bit to the left (so it wasn't just barely grazing the edge of the main piece) and used surface/surface fillet.

- Michael

Hi niko, here is the file before filleting.

You don't need to try and trim and join the pieces beforehand for surface/surface fillet, just have the 2 surfaces as separate objects and select them and run Fillet.

- Michael

Wow thanks

I was under the impression that surfaces had to be joined and you then fillet the edge. that is much easier...

Which brings me to the next question


I guess the rules of filetting is a dark cloud to me.
How come I can't fillet a cube intersecting a sphere the same way you did the last two surfaces??
Sorry for being a dumbass man I just want to learn everything I can about MOI..

Hi niko, Michael did mention 'surface to surface', so if have a sphere intersecting a cube both are solids and that method of filleting won't work for that situation, but if you separate the cube into individual surfaces you can then use surface to surface filleting and then join everything back to a single solid body.
Usually in that situation you would Boolean union the solid bodies then apply the fillet using the intersecting edge, either way works.

Cheers
Danny

Hi niko,

> How come I can't fillet a cube intersecting a sphere the
> same way you did the last two surfaces??

Well, you could, but like Danny mentions you would need to break the cube up into individual surfaces.

That other kind of fillet is a surface modeling technique that builds a fillet between just 2 surfaces.

If something is made up of a lot of surfaces and also has sharp corners where those surfaces meet, then the regular edge-based filleting is usually better since the edge-based filleter handles creating multiple fillets surfaces and also deals (or tries to anyway) with trimming back fillets that collide into one another and putting in corner patches in those areas.

But if you are already working with surfaces, the surface/surface fillet can be handy, and also it is somewhat less complex than the edge-based one, so it can sometimes succeed in building you a fillet surface to work with where the edge-based one can't figure stuff out.

- Michael

Thanks guys