Convair 240 – T-Splines Nose/Cockpit Modeling

One of the hardest parts of modeling an old airliner like this is the front – there’s a lot going on surface wise.  Furthermore, when making a surface like this with traditional NURBS modeling packages, there’s a lot of small transitional surfaces to make it all work and flow nicely, which means that it’s difficult and time consuming to edit if you want to tweak it.  So, I thought a lot about whether I could do the whole front of the airplane as one T-Spline.  At first I thought “no way……not with the current version of T-Splines and the limitations of the tsCrease function.”  Then I thought, “well…..okay….maybe…..just maybe.”  Let me explain why this is difficult.  tsCrease, the command that is needed to make nice hard edges, like those at the border between the window and fuselage, has three limitations.  They are:

1. You cannot pass a crease through a star point, or else it will radiate out in all directions from the star point.  This will look awful.

2. A crease fade can only take place over two edges – there is no way to set a different “falloff” for tsCrease.

3. You cannot have a crease fade out within two faces of a star point, or it will continue on to the edge of that two face region, and THEN fade out.

If none of this means anything to you, then just take my word for it that creases are very hard to manage and integrate into T-Splines models successfully.  For those of you who are using T-Splines, let me show you the solution I cooked up.  I call this the “Creased Column” and I’m quite proud of it.  Let’s say you’ve got a 3X3 T-Spline surface like this:

If you wanted to make the middle square of this face a window, you’d be stymied by the fact that the crease will pass beyond the four edges that form the internal face – this would make the area outside the window creased as well.  S0, that won’t work.  Okay, so let’s try extruding the center face:

Well….that won’t really work either, because if we crease the edges that form the top of that face, the crease will then pass through the star points, and we’ll be no better off than before.  Ahhh….now here’s an idea:

Well this is interesting, I’ve inserted an edge loop between the top and bottom of the extruded face.  Do you see the edge loop I’ve highlighted?  Well, since it’s in the middle of that extrusion, it forms a complete loop unto itself.  So, if we crease that edge loop we’ve got two things going for us.  One, the loop doesn’t pass through any star points.  Two, you don’t have to contend with crease fades, as it’s a closed loop.  So let’s crease that loop and toggle it smooth:

That may not look great, but we’re actually very nearly there.  Let’s set the z level of the creased edge loop, and the edge loop inside of it at the same z value – you can use SetPt for this:

Ahhh!  Now we’ve got it.  We have a completely faceted face, inside of a nice smooth T-Spline.  Now, that’s a very round window – in this case, that’s not what we want.  So, what to do?  Insert some edges like this:

They only need to go in one direction.  Now when we toggle smooth, we get this:

That’s very nearly what we want – notice the corners are nearly square and the sides nearly straight.  How to finish this off?  Manually collapse the points on the edges you added where they intersect the creased loop:

In smooth mode we get:

We’re exploiting the degree 3 nature of T-Splines surfaces here – by collapsing three points onto each other, we are forcing the corners square and the edges straight.  So, voila, behold the creased column!  Now, armed with this idea, I was finally able to tackle the nose of the Convair – this is pretty much the creased column run amok.  Here’s what I got:

You can see how I added an edge down the middle of the front windshield to add that bit of arch to the upper and lower edges.  I was also able to inset the windows just a little from the surrounding geometry, which makes it look very nice.  Basically, you can tweak these windows into any shape your heart desires at this point, it’s just a matter of adding geometry and fiddling with it until it looks right.  Furthermore, any 6 window nose like this can be easily modeled by using this T-Spline as a base.  Pretty cool!

PS – One thing I realized after I posted this is that the windows act almost like a “topological garbage disposal.”  Since the inner edge of the windows is far enough away from the star point to accommodate a t-point without any problems, you can start or terminate edges very easily, and with absolutely no effect on the surface.  If you look at the first two screen shots of the nose, you’ll see there is an edge that goes from the top middle of the side windows and then over the top of the cockpit.  There is no need to carry that edge down across the bottom of the nose, as you can terminate it so easily inside of the window.   Very, very useful.

7 Responses to “Convair 240 – T-Splines Nose/Cockpit Modeling”
  1. Frank Jarratt says:

    Amazing. Thank you for posting this.

    Frank Jarratt

  2. Dana Terrill says:

    You Crazy Man……

    operative spline…

  3. Chris Holtorf says:

    Amazing work. It would be nice to see an article on how you made the Convair’s nose – from begining to end.

    Chris Holtorf

    • Chris – That’s certainly something I could do! In broad strokes, I first figured out how to make the windows, then I started with half of a t-splines sphere for the nose. I extruded back and up to form the windows, then added geometry to crease the windows as shown. After that, it was just a matter of extruding the fuselage back and using tsMatch to make it round where it hit the fuselage. I’ll see if I can cook up something a bit more detailed in the coming weeks.


  4. Naresh says:

    Very cool sky! Thanks for sharing.

  5. Sangwa says:

    Hi! Always appreciate your contribution to modeling in T splines! If it weren’t for your excellent webinars and works I wouldn’t have tried the software at all.
    Im having a lot of trouble modeling a ‘pod’ for my ultralight trike and had little luck with rhino loft and curves functions. What approach should I take when doing it within t splines? So far I’ve tried starting from primitives boxes and quad ball; neither seem to have adequate flow and control.
    Should I take a half quad ball and subdivide it a few times?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: