D 520 Rhino Build

Discussion in 'Gallery & Designs' started by fimdan, Jan 5, 2006.

  1. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    Hello and welcome again.

    I decided to start this thread to show how I am building a model of Dewoitine D520 WWII fighter.

    This will be my second attempt to make something in Rhino. This time around, I think my life will finally allow me to slowly progress and finish this project.


    All of you are welcome here to make comments and suggestions as to how this model should be changed or improved. To those of you, who are new to Rhino, feel free to ask questions and learn from my experience. Those who already know Rhino and paper modeling, please I ask you to help me out by warning me about pitfalls and traps I may be getting myself into.

    I have chosen this aircraft for many reasons:

    1. It was easy to get a decent looking drawing online. (airwar.ru)
    2. The airplane does not have a complicated structure.
    3. There is lot of good pictures online documenting the aircraft, especially the internals.
    4. It is an aircraft that has not been published 100 times before by Halinski or other publishers. (Knowing how things work though, I bet there will be one soon)


    Here is a pic of where I am right now:

    [​IMG]

    I got about 50% of the nose section and I decided to move on and get some of the wing done. I think this is the most difficult part of building any aircraft. Being not very experienced, I spent about a day tinkering with different approaches to the problem and I finally settled with what you see in the image. This is, more or less, how the real d 520 had the wing set up.

    This time I do take cardboard and paper thickness into account. For cardboard, it is 1 mm. For paper is it .25mm.

    So far I have been enjoying this project a lot. In fact, I find doing this more interesting than building existing kits. I learn 3D software, I learn a lot about the aircraft itself, and for now, I do not have to worry about gluing my fingers together :)

    Wish me lots of perseverance though.

    Thanks for reading,

    FimDan
  2. wunwinglow

    wunwinglow Active Member

    Joined:
    Jan 18, 2004
    Messages:
    1,093
    Likes Received:
    1
    Looking good! One to watch, guys!

    Tim P
  3. sparrowhawk

    sparrowhawk Member

    Joined:
    Jan 2, 2006
    Messages:
    42
    Likes Received:
    0
    There is another reason to build it: The D520 is an aesthetically very pleasing aircraft, clearly showing it´s origins in French pre-war racing aircraft. Looking forward to see this project progress! Will you do a float conversion, too?

    Greetings, Martin
  4. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
  5. John Griffin

    John Griffin Member

    Joined:
    Sep 30, 2005
    Messages:
    175
    Likes Received:
    0
    Nice paint job

    Wow! Check out the great paint job on the belly of the giant D-520. He must have fuel-proofed the entire thing as it looks like it's glow powered. Corrugation is a great idea for strength on large scale AC. Thanks for the link.
  6. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    More pics

    Work with the wing continues..

    Couple of images about my design of the area where the wing and the main body of the aircarft intersect.

    [​IMG]

    Below you can see what needs to be done to reflect the thickness of the material. Gets pretty complicated.

    [​IMG]

    Time to get some rest.

    Hope you liked this one.



    FimDan

    Ps. I wish you could show us some of your design work John. That Tempest of yours is very nice. Great job.
  7. wunwinglow

    wunwinglow Active Member

    Joined:
    Jan 18, 2004
    Messages:
    1,093
    Likes Received:
    1
    Some great modelling going on here; and Pink Floyds 'The Wall' too! Great thread..

    Tim P
  8. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    Thanks for all the kind words.

    I think I bumped into a problem. The patch command seems to be doing something strange.

    Here is more or less what happens:

    I create cross sections by constructing them from two curves as shown below. One curve is a straight line; the other is the actual outline of the cross section. I use the end snapping tool to make sure that both of the curves are connected but I do not join them with the Join command.

    [​IMG]

    Now when I select my two curves and I run surface->patch command. I get a surface that does not precisely much original curves (I tried different parameters for this option as well, still does not help)

    [​IMG]

    The only way to get this done correctly involves using the Edge Curves (or Surface from 2 , 3 or 4 edge curves button ). This does it.

    [​IMG]

    Could some of you try to recreate this? Does this happen on your machines as well? Is this a bug, a feature, or I simply don't understand something?

    Thanks,

    FimDan

    Ps I just tired joining both curve. Makes no difference.
  9. Matthias

    Matthias New Member

    Joined:
    Dec 13, 2005
    Messages:
    24
    Likes Received:
    0
    From the (Surface-Patch) help:
    • This command can be very useful, but it also can produce some unexpected results. Use Sweep2 instead if possible.
    • For a trimmed patch, select curves that form a closed shape, and select them in order so each additional curve touches one already selected.
    • You can select additional curves that will influence the shape of the patch (such as dips or peaks in the middle of the patch). These do not have to be connected.
    • The patch is a fit surface and may not pass exactly through all of the input curves, but with proper configuration. it will be very close to reasonable input. Points are sampled from the input curves and the surface is fit to the points. Input curves that are surface edges are also sampled for normals to use in orienting the surface.
    I don't know any of the tricks in Rhino- can you explain the advantage to using a patch surface in this application? Using either 'loft' or 'edge' curves gives you a "perfect" surf picking up the needed control points from your curves. 'Planar' or a trim from a surf give you perfect edges while allowing the meshing grid to stay "square".

    Matthias
  10. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    I learned two lessons from this:

    1) RTFM
    2) Do not use the patch command for creating patches.


    "Patch" seemd the right name for what I was doing. That's why I used it. No tricks here. Strangly enough though, loft is not very exact in this case either.

    Thanks,

    FimDan
  11. Maurice

    Maurice Member

    Joined:
    Jan 16, 2004
    Messages:
    509
    Likes Received:
    2
    I've found sweep 2 rails seems to be the best process to use here as it leaves the prog in no doubt as to what is wanted. Elsewise it can get confused and will sometimes muddle starting and end points as well as direction.
    Remember Rhino's are small brained and shortsighted.

    Cheers
    Maurice
  12. Matthias

    Matthias New Member

    Joined:
    Dec 13, 2005
    Messages:
    24
    Likes Received:
    0
    If ONLY the manual told you what to use as well as how to use it... The process of figuring out the tricks is a big part of learning a new package, and my old ones (Autoship, mostly) don't apply anymore!

    Loft should give you perfect results if your curves are properly defined _and_ you remember to turn off simplify.

    Matthias
  13. Willja67

    Willja67 Member

    Joined:
    Feb 12, 2005
    Messages:
    907
    Likes Received:
    0
    In most cases I extrude a line so it creates a flat plane and then I can project the curve onto it then I trim the excess off and I have my former.
  14. Matthias

    Matthias New Member

    Joined:
    Dec 13, 2005
    Messages:
    24
    Likes Received:
    0
    Muddling end points usually is due to imperfectly defined splines or clicking in the wrong order. Setting up your guides, curves etc carefully in the first place makes a big difference further down the road, but for simple modeling (we are only brushing the surface of Rhino's capabilities) we can probably get away with workarounds. Surface/curve direction can wreak havoc with volume calculations, for instance, which in marine design can have you end up with (say) a buoyancy tank that when mirrored for the port side of the hull ends up with a magical "negative volume" that acts as if it were filled with double density water! Fine if you catch it in time :-?

    [​IMG]

    Here's a bulkhead surfed five ways. The first three all result in "perfect" matches. The first two have surface edges that are editable using control points matching the original spline. Planar and trim fits your surface into a virtual one representing the boundary of your pre-trim surface (extents in planar) and allow you to edit only the control points of _that_ surface (four corners, usually) while patch depends heavily on your settings.

    You can see that the geometry Rhino uses is quite different. The UV directions curve with the edge in the left hand examples, while the other ones are "punched" out a quad. This has a dramatic impact on how these surfaces would take a rendering texture or bump map, if you tried to work on the nurbs within rhino rather than exporting a polymesh to a more suitable prog.

    [​IMG]

    On export, you can see clearly what Rhino is doing here. The first two are considerably "cleaner" meshes, and are less likely to cause trouble. They are also more efficient polygon-wise to achieve the same edge resolution.

    What does all this mean? No clue! But I learned something today (although I was really hoping FimDan had a trick... I'm still stumbling) :)

    Matthias
  15. wunwinglow

    wunwinglow Active Member

    Joined:
    Jan 18, 2004
    Messages:
    1,093
    Likes Received:
    1
    The menu listings in rhino are generally arranged so the simpler operations are first (top of the list). Try the most appropriate high items first, and only go on to the more complex commands if your situation demands it; in other words, if you can loft the shape you want, do that, rather than a 1 rail sweep, but if a 1 rail sweep won't give you the right surface, use a 2 rail sweep.

    And remember, for paper models all the surfaces you want will need to end up flat, so straight lofts will be perfect. Just unrollsrf them to get the pattern you want. You can export these shapes as vector lines (ai file) and avoid all those nasty little polygons completely. However, if you really MUST export as a mesh, Rhino has far and away the most comprehensive mesh controls of any CAD package (not me saying that, but an article in a top CAD magazine a couple of months ago. And we are talking in comparison with top end CAD packages like ProEngineer etc!) Matthias, your mesh former examples could be massively altered by the mesh control settings, no matter how the original shape was defined. And I would suggest the three right-hand meshes are far better as the curved edges will be much better resolved than the two left-hand ones! I know which I would use if I was building these as SLA models (my daytime job!)

    Patch and Curve Network commands will always have some degree of approximation going on, unless the curves used are very accurately placed, in which case you might as well use on of the simpler commands. Treat them as fixes when all else fails. Better still, clean up your construction accuracy so you don't need to use them.

    Tim P
  16. Maurice

    Maurice Member

    Joined:
    Jan 16, 2004
    Messages:
    509
    Likes Received:
    2
    I shall repeat my plug for sweep2rails as the operation of choice when creating the surface between two bulkheads (not talking about within a bulkhead). This simply because it not only eliminates the muddle factor but also provides the most even distribution of polygons within the surface.
    The rails are swept at a more uniform rate, and this irrespective of the severity of the change in section between the 2 bulkheads.

    Cheers
    Maurice
  17. wunwinglow

    wunwinglow Active Member

    Joined:
    Jan 18, 2004
    Messages:
    1,093
    Likes Received:
    1
    Maurice, sweeping (or lofting, or any rhino surface command for that matter) doesn't create any polygons! They are NURBS surfaces! you only get polygons when you generate a mesh from the surfaces, and that is a completely seperate, and totally controllable process, and in many ways completely independant of how the original NURBS surfaces are created. Rhino is NOT a mesh modelling package.

    I agree that sweeping between two rails is unambiguous though. But if Rhino gets confused, it is usually because the operator has asked it to do something ambigouous, or inappropriate. You can't blame Rhino for that!

    Tim P
  18. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    Hey

    Thanks a lot for explaining all that. I am glad that we all learned something here. I could not possibly imagine that my innocent request to clarify something would evolve into a storm of ideas. Good, those who are learning along with me will get a lof of interesting readings. That's why we love this forum, right :)

    Thanks for all the constructive criticism.

    FimDan
  19. B1Z4RR0

    B1Z4RR0 New Member

    Joined:
    Jul 10, 2005
    Messages:
    1
    Likes Received:
    0
    Ok!! great Rhino design, great work....but...do you really think how you can expand (unfold) this model to transfer it to carboard??...Rhino can not unfold all surfaces....only the expander rhino plug-in can do it. (and it is expensive)
  20. fimdan

    fimdan Member

    Joined:
    Jun 9, 2004
    Messages:
    77
    Likes Received:
    0
    I sens a category 5 coming. Better get your sand bags ready:)

    Yes I have a lot of faith in rhino as far as unrolling. In fact, I make everything so it unrolls. If you use the loft command, everything should unroll. If you happen to get something that rhino doesn't like, than you have to check all the curves you use to construct the surface. With a bit of experience you can get Rhino to unroll a lot of things. If you need some examples of that, feel free to explor cardmodels.net.

    I have one way of unrolling difficult surfaces. I will show it here soon.

    Expander. I do know it exists but I have never tried it. Plus is costs a lot of money. At this point I am in no need for such tools. Maybe in the future, who knows :)


    FimDan