Hi Guys,

I thought I'd do a small how-to on designing a wing and making a mold out of it.
I hope I get the chance to mill it but even if I don't this may answer some questions on how to do wings.


So first up - import the curves into SW and draw up the planform.
I'm using the profiles and planform out of the following thread: http://www.rcgroups.com/forums/showthread.php?t=1139965. The wing is fairly complex with an elliptical planform, under-chambered airfoils and washout.

Let's get started:
First modify the polar so it uses 3 coord for every vertex. Usually polars use X and Y only so here I'm adding the Z as 0.0000.
Make sure that the polar starts at the back and loops to the front. It should also close in on itself so if it started on (1,0,0) it should close on exactly the same coord.

Go to Insert>Curve>Curve Through XYZ Points and use browse to load your curve to the table. Press OK to import your airfoil. Do this for both the root and tip.

Up next we need to draw up the planform.

Draw a sketch on the top plane with the intended planform.
I split this up into 2 sketches so I can fine tune the wingtip without disturbing the basic wing shape.


Next up we need to create a plane for the tip airfoil. Notice that on this wing I'm only using a root and tip airfoil. Different wings may need intermediate profiles.

After creating a plane for the tip draw a sketch on it and use "Convert entities" to copy the spline of the tip profile to it. This is the real trick in using airfoils in SW. Once you convert the spline to your sketch you can Scale, Move and Rotate it to fit the wing. Here I scale it to the tip's length, I move it back to so it coincides with the planform drawing and finally I rotate the tip so I get the required 1.2deg washout.

On this point we should have the wireframe of the wing ready.

Simply loft the root to tip using the planform as guide curves.

It's always a good idea to check with curve tools and zebra stripes to see you don't have a gross error on the loft.

Great, thanks for the lead-in tips. Im slowly working my way through SW & surfacing will probably be around teh corner soon.

Can you have a look at my Rhino posts #122 & #124, link below & comment on how the intermediate wing sections can be influenced by the rails. I noticed you only used 2 sections, (root & tip) for the loft & was wondering if SW resulted in a similar distortion. The solution is pretty simple, more in-between sections are required. I cant tell if your curvature graph can adequately show that aspect - maybe the intermediate loft section in straight-on plan view?

http://www.rcgroups.com/forums/showt...6#post13798116

Originally Posted by ptxman

Great, thanks for the lead-in tips. Im slowly working my way through SW & surfacing will probably be around teh corner soon.

Can you have a look at my Rhino posts #122 & #124, link below & comment on how the intermediate wing sections can be influenced by the rails. I noticed you only used 2 sections, (root & tip) for the loft & was wondering if SW resulted in a similar distortion. The solution is pretty simple, more in-between sections are required. I cant tell if your curvature graph can adequately show that aspect - maybe the intermediate loft section in straight-on plan view?

http://www.rcgroups.com/forums/showt...6#post13798116

The reason you see twisting is because you are sweeping the root chord instead of lofting between two chords. In SW there's a way to prevent the twist and align the stations in a certain direction. I'll try to put up an example tomorrow.

Erazz,

How would you address washout in this construct? Would it be as simple as drawing the trailing edge on a different plane then the airfoil would be swept between them?

Cheers,
Eric B.

Originally Posted by AirX

Erazz,

How would you address washout in this construct? Would it be as simple as drawing the trailing edge on a different plane then the airfoil would be swept between them?

Cheers,
Eric B.

I rotated the tip profile (around the LE in this case) by 1.2 deg. Lofting between the two profiles now gives the desired washout. You can also have non linear washout by making the root extrusion perpendicular to the profile and also by specifying different guide curves.

Thanks Erazz, same as newer versions of Inventor. I was curious, I have had some time with Solidworks several years back.

Cheers,
Eric B.

It seems I may have spoke too soon.

I had another quick look inside the loft and discovered that something was just wrong! If you look closely you'll see that the profiles are puffed in the middle.

This is because SW doesn't loft correctly. I'm working on a solution and I'll get back as soon as I find one.

OK, I got it sorted out.

I thickened the TE of the profiles. This I should have done in any case because I need to have a place for the excess cloth to hang. At any rate, the profiles must have the same number of segments, should look similar and be well made.

SW has a "centerline" option. This is great for wings because it makes every profile section normal to this centerline. However, this doesn't seem to work very well with guide curves. Unigraphics NX does this very very well. I had to make a lot of "support sketches" (look at the tree at the left) to make it work right.

Oh well, at least I have a "plug" which I can now turn into a mold.

Just out of curiosity, did the lofting surface behave any better if you inserted some 'in-between' (root & tip) rib sections? Somebody, somewhere explained this to me in terms of minimum number of defining control curves in terms of how it fits & extrapolates, but that may have been specific to that particular 3d modeler (rhino).

Also, I assume you are making these wing panels as 'solids'? Ill be interested to see any subsequent filleting operations on fuselages etc.

Thanks for posting, following with interest!

Originally Posted by ptxman

Just out of curiosity, did the lofting surface behave any better if you inserted some 'in-between' (root & tip) rib sections? Somebody, somewhere explained this to me in terms of minimum number of defining control curves in terms of how it fits & extrapolates, but that may have been specific to that particular 3d modeler (rhino).

Also, I assume you are making these wing panels as 'solids'? Ill be interested to see any subsequent filleting operations on fuselages etc.

Thanks for posting, following with interest!

It certainly would behave better. The problem would be interpolating the airfoil in the "in-between" sections. The ultimate solution is to use a centerline that forces the sections to be parallel and scales them according to the guide curves. This is a snap to do in NX but a pain in SW

Yup, these are solids. The subsequent operations are to produce a mold (I hope to do this tomorrow). The fuselage is somewhat in the future but I promise to post it as well

Thanks!


p.s. Just ordered some aluminium plates. Should have molds within a week

Originally Posted by erazz

It certainly would behave better. The problem would be interpolating the airfoil in the "in-between" sections.

If the root & tip sections are the same airfoil, that would just require a simple 2D scale command to match the in-between chord length at any particular panel span station.

If the root & tip are different foils, I typically let Profili do the work (which is usually where I obtain & modify the base airfoils anyway). I cant recall the command off hand, but its a button click. You can create a hybrid blend based on any defined amount, like 90% of root & 10% tip for example. A mid-span blend would be exactly 50/50 etc. I think its just a numerical interpolation between the two curves, but anyway you have a new airfoil to import. You could probably acheive the same result in a graphical/cad program starting with the native x,y 100,0 type coordinates.

Re SW, Im just getting into it now but Im really impressed with the shelling capabilities. I can see a lot of applications for modeling composite wings & internal structure components like spars & supporting ribs working from the outer surface in. Im not too sure how it performs on the curvy swervy world of wings though vs simpler solid bodies Im playing with. Complex surfacing is usually what highlights any limitations in the cad program.

Originally Posted by ptxman

If the root & tip are different foils, I typically let Profili do the work (which is usually where I obtain & modify the base airfoils anyway). I cant recall the command off hand, but its a button click. You can create a hybrid blend based on any defined amount, like 90% of root & 10% tip for example. A mid-span blend would be exactly 50/50 etc. I think its just a numerical interpolation between the two curves, but anyway you have a new airfoil to import. You could probably acheive the same result in a graphical/cad program starting with the native x,y 100,0 type coordinates.

Doh!

Your right. I totally forgot about Profili! It's not easy to do with CAD. But Profili is great for that. Thanks!

Originally Posted by ptxman

Re SW, Im just getting into it now but Im really impressed with the shelling capabilities. I can see a lot of applications for modeling composite wings & internal structure components like spars & supporting ribs working from the outer surface in. Im not too sure how it performs on the curvy swervy world of wings though vs simpler solid bodies Im playing with. Complex surfacing is usually what highlights any limitations in the cad program.

Shelling usually works great for smooth things. It has problems with thin areas. I work as a plastics engineer and we use shelling in almost every product. I really like SW for most things because you can design things really quickly. But for designing wings and fuselages I still wish I had NX. I used to works with Unigraphics 15 on a Unix station. It was a total pain to design with but lofting it did splendidly!

OK, Back to the mold.

First up I want to split the faces where the parting line will be. You can have the parting line split the faces by itself but I find that I get more control on the parting line if you split it by yourself.

Since the mold is going to be open in the root section I don't bother with splitting the root.