I am trying to loft the blade of an aircraft winglet. The tip of the
winglet has a curved leading edge, so the planform isn't a trapazoid.
When I try to loft the shape, using the rib profile at the base of the
winglet, the LE and TE as guides and then I place a point where the LE
and TE meet, the shape basically looks good right at the base, then
turns into a thin blade where it transitions to the point. I have
tried tweaking all the loft parameters but no luck.
I have tried exporting as an IGES to Rhino, then using all the guide
curve geometry I am able to get a satisfactory loft, but need to do
this using Solidworks.
Am using SW2005 SP-5.
Because you have tried everything, I'm sure this suggestion won't help,
but did you try to use the "tangent to profile" end condition option on
the point profile? Is the point sketch on a plane roughly tangent to the
end of the wing?
You may want to experiment with adding another closed profile at some
distance between the "base" and the "point" in order to better control
the transition to the pointed end.
Per O. Hoel
Good points from others about using end tangency at the point profile,
but there is something more fundemental that is, I fear, being missed.
It has been my experience after ten years of working with SWx that
lofts/sweeps fail to deliver what you are trying to get out of them
when you ask them to do too many radical changes in a single feature.
My policy is to pay attention to how a loft 'wants' to work, and break
the feature into sections that are consistent with how a loft 'wants'
I did a propeller a few weeks ago, and here is how I did it:
The main shaft, until it starts to converge to a rounded tip, I did as
a single loft. Even better, let it start to converge at the end (but
just a bit... do not completely converge), then use a trim or cut
(depending on surfaces vs solids) to cut it back to the region that is
mostly consistent in shape (there is lots of hard-won background
behind this that you will figure out with a little experimentation if
you don't look at any tutorials).
Then, I built the portion that converges at the tip. I did it as a top
convergence and a bottom convergence using surfaces (tangent to the
root body) but that is not *necessarily* required. The variance
between surface and solid techniques is a little too much to go into
now - experience, education, and comfort level will inform that.
The topic of breaking things up into regions that look 'mostly' the
same to develop the real shape you want was covered extensively in the
middel protion of my presentation at SWx World 2006 - look at
www.dimontegroup.com, under tutorials (the middle of the left side fo
the nav bar) and look at Curvy Stuff 6 tutorial. Be sure to get the
sample files and the storyboard ppt to explore, step by step, the
issues with getting greedy and doing it all in one feature and why
breaking into individual faces gets past the prblems.
I concentrated on it at SWx World because I have talked to folks who
have spent as much as a week trying to tweak a single loft into doing
too much drastic change (and failing), when, in half a day they would
have nailed it if they had just broken it up into regions that have
very similar characteristics. Concentrate on the hook and the
volkswagon steering wheel in my presentation, ESPECAILLY the hub of
the steering wheel - see how breaking into 'relaxed, similar' regions
saves the day and improves quality. Look at the sample files and
storyabaord ppt and judge for yourself.
Speaking of judging for yourself, its best not to go 100% with
potential cranks on the web without independent varification. That's
fair, and I encourage it. Just look through the ads in the design
engineering magazines you probably get, and look at how the inevitable
images with CAD 'face boundaries' lines (which look cool in ads)
Though they are ads, you can actually see what the real Pro modellers
do (because the ad agencies are just getting fed the raw CA data - I
speak from some personal experience here) - when looking at airplane,
car, and even human face images, you will notice increased face
density in regions of radical change. relative to their more relaxed
neighbors, when they break up models into multiple faces to isolate
influence - which you are NOT doing if you are trying to pull off a
complete propeller from root to tip in one feature (making one face,
longitudinally). The root and shaft is really quite simalr, while the
rounded or pointed tip does completely different stuff, screaming
'break 'em up!'.
It's a message I've tried so amny different ways to share over the
years - Let the product tell you how to model it! Think of faces, not
features, and the product tells you how to model it.
How are you controlling the twist of the wing profile (NACA?) as it
goes up the wingtip?
The wing profile needs to be restrained by means of a spine to keep
it parallel to the aircraft centreline (x axis) or wing axis.
Posted with NewsLeecher v1.0 Final
* Binary Usenet Leeching Made Easy
I had the chance to look at the file. In the end, a better result was
achieved using a construction surface at the edge and two Fill surfaces.
The loft to a point is useful in some situations, but there is sometimes
a better way to go about things.
The winglet has 1.5 degrees of twist from the root to the tip, it is
actually a washin. The root airfoil and the tip airfoils are set at
the 1.5 degree difference. The LE and TE were created on a plane, and
then projected on a twisted surface, so I could use them as guides.
On Jun 12, 7:01 am, firstname.lastname@example.org (Phil Evans) wrote:
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.