Puzzling modeling challenge

Hi guys i have a rim i am trying to model but it has me stuck a one point very tough to see the start of it. Can anyone tell me how they would go about this? 9.3dm (2.4 MB)

If you can see those three surfaces… It’s not hard to model… Unless making those corners G2 or above continuity…

I see them but the way how surface 2 transitions into surface 3 is tricky

I actually made a model like this but its easy to mark out the major shapes to be modeled but its actually how to go about it is the trick, the steps to reach to the point

This is the file with the model i made but adding smooth transitions and fillets are really difficult. have a look KevJin9.3dm (4.3 MB)

Can you send me images of this rim?
I need perspective reference and large elevation view…
I just roughly did very first part… You should start something like this at beginning…
CPs are dirty since I don’t have an exact reference image, but it shows the idea how to do certain part.

I’ve included profile curves in the file…
1.3dm (321.3 KB)

actually KevJin, this is the only reference i have for the model. I love how you made such a smooth internal curve radius, how did you accomplish this?

this is as far as i got but adding fillets are so tough9.3dm (10.1 MB) in Rhino

I don’t think fillets are specially tough in Rhino.
You have to know some basic principles and then you’ll be able to fillet anything. In filleting there’s a lot of special cases such as when surfaces are not tangent to one another, when fillets of different radius are meeting, when they meet at the intersection of several surfaces that have varying angles between each other.
You sometimes have to untrim the base surface because the fillet will be generated by the trimmed-away portion. Some fillets are built on others, the sequence is important.
In your file the tolerance was too loose. The wheel has a diameter of about 3 inches, so the fillets will be around 0.010", sometimes less. To have clean trims you need a tolerance of 0.0001".
In the file I join, the orange layer has some fillets to show how I start. The red layer has a complex corner resolved.
9-Marc.3dm (4.4 MB)

1 Like

Marc i see your point but problem is in using that method that alot of the arc surfaces that i could use as cutting surfaces do not work and do not intersect the object so i cannot even begin to cut away to do something like a blend, i need a better example, anyone? Pascal?

You have to extract the surfaces before starting the triming. In situations like that, I like to use the edges of the fillet surfaces for the triming.
I start the Trim command, without anyting pre-selected, and type “crv”. Then it’s the edge(s) of the surfaces that can be selected.

Kevin Jin

Can you please tell me more about how you made these surfaces?

Did you split the edges and use blend surface to make the others that I did not circle? I would really appreciate it.

@jazzcat81

I used sweep2

2015-05-03_19-10-46.mp4 (17.4 MB)

Awesome, thanks for the reply.

I did try using blend srf with split edges and then a srf from 3 edges for the piece in the corner and it seemed to work okay as well.

Really appreciate the reply. Thanks!

…edit…

Ohhh, after watching your video I see there is more going on there. I see you seem to be evaluating your first surface (made with blend srf) by changing the degree and mirroring the surface. Can you tell me what you are looking at/for?

So you:

  1. make the first surface with blend srf, choosing the degree you want
    2)match the surface edge to the existing poly surface
    3)make the next surface with a rial sweep

Using this process the sweeps will inherit the degrees of the original surface, yes?

Sorry for the questions, but I just want to understand the deeper logic here.

I was looking if there were enough control points to avoid jumping points or handle that bend transition. Certainly min degree 5 surface was needed in U direction. In V direction, after blendsrf, 5 degree was not enough because of .25 blending factor. In other words, there was no control over in U direction.

The best way of doing that mid part surface should be dividing them into two surfaces. Because of the model was built incorrectly, it’s easier using sweep2 (no manul adjustment on CPs.) Using sweep was to obtain G2 without splitting more surfaces.

The surface blendsrf was not good because it was from two trimmed surfaces, meaning it would create dirty surfaces with many knots or kinks?Either rebuilding like the way I did or creating the surface from rebuilt curves would work.

Okay, I think I see what you are saying.

I appreciate the reply.

I would love to see more infor on this type of surface modeling. Can you think of any particularly good resources or tutorials on this kind of “class-A” surfacing? I have yet to find a good resource that really delves into the theory of this type of modeling.

For instance I have seen others say something equivalent to “The surface blendsrf was not good because it was from two trimmed surfaces, meaning it would create dirty surfaces” but don’t really understand why this is, or what should be done to avoid it.

Thanks again.

After writing a ton of words, I decided not saying it in public, as a Rhinoceros fan.

Publishing Class-A surface modeling book or vids using Rhinoceros has been in my to-do list. By the time I become a Mcneel authorized trainer and retailer, I will do that. Don’t know when…

Please post back here if you do. Ill stay posted on this thread. I think many people would like comprehensive introduction about the theory and considerations behind Class-A instruction, and rims are a great place to start.

Some extended explanation of single span and insight into issues like “The surface blendsrf was not good because it was from two trimmed surfaces, meaning it would create dirty surfaces”, would be really great.

Ps, this is a rim I became interested in a while back. I think the interior blends are very interesting and I could not figure them out at all. Would love to see someone tackle this rim.

I tried to model this rim a while back and failed miserably.