C-17 Globemaster Problems...

I’ve been working on-and-off for the last few weeks on a C-17 Globemaster cargo plane but I just can’t quite get my head around how to model certain parts of it. I’m quite new to Blender, and so far I’ve only done a few guns with reasonable success, but a plane is a new type of challenge that presents some very different shapes I haven’t encountered before.

Specifically, my major problem is with the wing root and the surrounding area.
The following images show what I mean:



I’m not sure how to go about effectively creating that “hump”, the region connecting the wings and just below them.

Should it be done by keeping it as a separate piece to the fuselage, or should I extrude from the cylinder shape of the fuselage somehow?
Does it look like it would involve a lot of gradual adjustment by hand and eyeballing the shape, or should it be a bit more straightforward to achieve without a lot of small-scale tweaking until it looks right?

This image from a model seems to show a very useful view of the fuselage without its actual wings blocking the geometry:


It shows quite well the trailing part behind the wing that follows on from the U-shape connecting the wings in the middle on top.

I have these good images but I’m still not able to figure out a way of creating all of the basic geometry that approximates the shape accurately. It’s quite a complex series of shapes compared to the relatively flat guns that I’m used to.

So far I have the following:


Seems like it could be the start of a semi-decent approximation of the hump region towards the front, but I’m not sure about how to transition it into the wing roots and the shape of the region below the wings, as well the trailing part that extends aft. I’ve played around with it but I really don’t have any good ideas for how to continue.
Another reference image that may be helpful is this set of diagrams:


Any major advice for how I should approach this?

My suggestion is to model out shapes to be used as templates for the major forms and use snapping to model the desired mesh over them. (And for much of it, it’s even not all that necessary to boolean those.) The workflow for that is going to be something like this:

As you see it makes it easier to get many blended forms, and then you just adjust loop placement to control the surface tension between shapes if subdivision is used.

Interesting… so it’s like shoving the basic shapes together and then weaving a mesh over it. I’ll keep that in mind for later, though I think I’m having trouble getting close enough to the basic shape as it is…

One thing in particular is how the wings attach in front:


We can see the wing goes towards the fuselage and sorta sweeps upward to join with the “hump”. I’ve been approximating the shape roughly but I can’t figure out how to use SubSurf (which is clearly necessary to get a decent result) and still get the beginning of the wing to join at exactly that point. If the vertices are lined-up exactly there, then obviously SubSurf screws that whole thing up and it has to be moved back somehow without messing up the geometry…

Going to try and see if I can find another approach…

It is a hard thing to figure out when just starting modeling. Topology is one of the biggest downfalls along with poor reference images or lack of. Whenever you are working with multiple mesh objects that are to be joined together seamlessly then you will want to plan ahead for those areas and watch where you are putting in your edge loops and such and plan for where they are going to join the other mesh at and keep those two join areas close to the same geometry. Also try to align your quads to flow along the direction they are supposed to go along the way instead of having to tweak everything once you got the basic shape done.

For a simple and easy example if you take a cylinder with 16 verticies and add a sphere with 16 you will now be able to seamlessly match that sphere to the top of the cylinder without any issues.

Speaking of which, this is something that’s bothered me for some time with modeling: how important is it (or when is it important etc.) that adjacent geometrical features are actually perfectly aligned with each other, rather than just one being embedded into the other (i.e. overlapping/extending some distance inside) so that it looks like one is perfectly aligned with the other?

When I first started modeling I wanted perfectionism with such cases and got frustrated, but then when I started modeling some guns recently I relaxed myself more and achieved some good results. Now with the interwoven shapes on the C-17 I’m not sure again.

Here’s a pic of my latest attempt:


I positioned the upper section by snapping to the fuselage’s vertices, but then I added a few more loop cuts. Either way, they wouldn’t have perfectly fit together because of the SubSurf modifier. Does this case seem like it makes sense to not worry about aligning the two objects perfectly, or is there a better way? Sometimes it does feel like adjustments might be made easier if the objects were aligned, but maybe it isn’t worth the effort trying to do that and I should just make an approximation?

Hi there… so wellcome to the blender unity… but you are new to this. well.

As far as I see from your attachments, yaou are missing some real essential basics about modelling. So I would suggest to get some tut’s where some things are explained.

Especially conc. your “task”, there is a two-part tut on youtube (made by one of the real “gurus” of Blender)

… which is the first one and the secc. you’ll find in the list on youtube.

The task is quiet the same, but you might find some real “hints” on how to start and how to proceed.

Keep your faith and happy learning/blending

I could give a shot at recording an ad-hoc approach to this aircraft later on similar in fashion to that video I linked. But I’d make the form for the fuselage from one shape, elongated spheres as another shape for the wing-root, then the entire wing V as another shape. Then overlap those and then snap the final mesh over all of those. Or at least do that for the area where they join and the blending of the shapes is needed. Also keep in mind that you can join or separate meshes and take advantage of modifiers as well. And sometimes adjusting settings of geometric primitives is handy because it’s possible to match things up in terms of mesh density and available geometry. (Still takes a while to get a knack for it, but it’s a lot easier once you get the concepts down.)

Ok, well my latest attempt is below. Getting closer I think.


Does it look like I’ve made any obvious errors? Or should I be able to continue from that point?

Ya, I’ve watched a few of his videos including those; they were quite useful in getting my head around a few things like forming wings etc., but I’ve still found it hard to apply the ideas to the rather complex shape at the center of the fuselage.

Think I’m getting there now, and it looks a bit better with SubSurf than in the image above. You can see in the pic tho how the corner in the bottom-right is overlapping and going into the fuselage; I wonder would I pretty much have to create a loop cut in the fuselage with a vertex in the spot where I want the wing root to snap onto, or is there a quicker way? Face snapping never seems to work like I’d expect, i.e. you’d think you could move a vertex along an axis until it exactly intersected a face but it doesn’t seem to work that way… or at least it’s quite fiddly and hard to use.

@WireOfEvil I just did a quick reference line up and quick model. If this is what you are looking to achieve I can share my blend file with you so you can go through it and see what I did and look at the geometry. But here is a screenshot of what I did and it actually didn’t involve any additional objects or joining of parts. Just started with a 16 segment cylinder and pulled the edge loops I created at the position of the geometry change and slid them into place. After that just a simple extrusion on the z allowed for a nice position for the flow of the wings to be extruded outwards and put into position as well.


Here is one as well with draw all edges and wire turned on and subdivision turned off (so all you see is the loops that were needed)


And yes you have made progress from your first attempt, however I still think you have a little too much geometry to start with around the wing section. Edge loops are a help after you get your basic geometry set up so you can tighten up areas and shape them a little more vs. a crease.

That looks about right, the blend file would probably be helpful thanks. Somehow I just wasn’t able to get something that looked close without using a separate object.

Yes, mine looks more complicated but with a less accurate result and the annoyance of two separate objects to deal with. I suppose the extrusion gets easier the less geometry you can get away with.

Is there a rule of thumb for when creases should be used? It did feel kind of sloppy to use a crease when I did; the real reason was to avoid the SubSurf causing the bottom of the wing root section to come away from the fuselage. I suppose in this case the answer was just to not use a separate object.

Here is the blend file if you want to look through it, as you can see I got a little carried away modeling :slight_smile: but for the wing part it was two loops added to adjust the strength by sliding them closer or further from the extruded geometry. And using creases are a nice ‘cheat’ and work well when you do not want to add additional geometry, kind of helpful for hard surface modeling but for organic modeling it isn’t really a preferred method as well as with animation. Best advice is keeping the model simple at first then adding the detail once you got a base mesh laid out.

ExampleC17.blend (829 KB)

Cool, thanks for that. I’ll give it a good look and see if I can apply the ideas to my own model…

Ok, I took a break from Blender and then took a while to gradually understand exactly how to reproduce the geometry in the example file. I wanted to figure it out properly by myself if I could, and sometimes that took a surprising amount of time before realizing it was simpler than I’d thought at first. I now think I’ve managed to go through how it was done and assume I understand a much better way of approaching these kinds of shapes. Posting my results below…

I began by working backwards from the example and trying to recreate what I assume the fuselage looked like before any extrusion, in terms of the edge loops and vertex positions it had. At least I managed to discover the Grid Fill tool during all of this, which I didn’t previously know existed but I’d been looking for a way of doing exactly what that does so that’s useful.
You can see the result below; I’ve tried to smooth it all to a roughly consistent cylindrical shape but obviously it’s a bit iffy. Still, I assume this is roughly what the stage before any extrusion should look like as far as number of edge loops and vertex positions are concerned?


For the wing root, I selected the top faces, extruded them out and then pulled them inward a bit to create the gradual curve. Then I moved the frontward-most top vertex nearest to the side forward a bit (this took me far longer than it should have to realize what had happened in the example... sheer inexperience). Then I selected the appropriate 9 faces on the side and extruded along the X axis (again, I at first figured out how to approximate this by deleting 4 vertices and extruding the surrounding edges but it didn't at first strike me how to do it by simple extrusion...). I then cleaned it up by deleting the end faces and aligning the vertices. I noticed I had to dissolve 2 edges running along the top of the wing then to achieve the exact same topology as the example; I assume it's just good to remove all edges that seem unnecessary when you're trying to get the very basic shape, and that they would have also been manually removed for the example? Otherwise the topology must have been quite different before the extrusion, but I can't picture that since it's just 2 edges running width-wise across a single length-wise edge loop... probably not important either way.


I extruded the wing stub out again to the full length of the wing and scaled it to taper like a wing, then just cleaned it up by moving a few edges and vertices around by hand until it looked more like the example. (In the below pic, my attempt is on top and the example is on the bottom)


For the bulge, the inner loop especially confused me at first because I wasn't sure how to create that with extruding alone.




Then I just extruded twice and scaled inward each time, counted the faces and they were the same amount as the example so I assume that’s the same topology and all I’d need to do is pull the vertices around until they look about right?


All in all, does it seem like I have a good grasp of how it should be done? I think I've learned a few things either way.

Oh and something else I noticed but couldn’t find the answer to: I tried changing the render engine to Blender Render in the example file (to see what it would look like in low-res without any AA like I was experimenting with some other models) but the render always turns out completely grey. I messed around with the camera a bit and did some searching but I wasn’t able to get anywhere; it only works in Cycles. Why is that? I could always render my own models in both Cycles and BR, so I assume there’s some setting that’s been changed somehow that prevents it from working with BR?