by sgreenawalt » Thu Mar 10, 2011 10:44 am
Adam -
Here's a bit of info on how I dealt with the wingtip lenses. Let me start by saying that it would be much smarter/better/easier to just design the lenses in TSplines, and then machine molds, but in this case the customer had invested quite a bit of time and money in making the wingtip lens molds (and many sets of production parts) before I came along, and so it was important to him that we keep that existing work. Since the customer is always right, I had to find a way to make the existing parts work on the "new" wingtips.
I started by digitizing the lenses with my FARO arm. There is a definite art to this, and I feel I'm getting there, but this is a whole other skill set that takes time and practice to develop. I've found that digitizing evenly spaced points on the surface results in the best and most accurate surface, so I actually plan out every point that I digitize ahead of time. Digitizing is about 75% prep and 25% actual work with the arm. I had to find a way to fixture the lenses without breaking them, so I used some chevant clay to hold them on a piece of shelf board. Masking tape was applied to the board so that the clay would stick better to it - I thought I had it all worked out the first time, and then the lenses just popped right off the board with the clay! In this particular case, there will be a flange around the edge of the lens that will attach it to the tip, and so that's the most crucial part. The actual lens surface itself isn't nearly as important as the flange around the perimeter. So, I laid out two concentric rings of points - one along the perimeter, and one about .75" inside of that - each with the same number of points. For the rest of the surface, I just made three cross sections in one direction, and one in the other. Then I rebuilt all my curves, which helps to take out some of the waviness, and I made all my cross section curves planar. Then you have to compensate for the radius of the ball tip on the digitizer - the digitizer captures the point at the center of the ball probe, and so you have to compensate for that - in this case since I used a .25" ball probe, everything had to be offset by .125" After that, I imported the lenses into my main file, and then placed them on the tips manually, just by moving them around until I felt they were as close to where the main surface was, and then tweaking my wingtip surface.
This was where the power of the new tsSymmetry command really shined - I was able to keep working on the back 75% of my model (since I was still making some final tweaks to the winglets) while matching both the left and right tips to their respective lenses. There were significant enough difference between the left and right tip to justify tweaking each side independently. The way I matched the tip to the lenses might sound like insanity to some, but what I did was simply manually tweak my tip surface until the meshes of the tip and the lens got all "mottled" with each other. I wanted to have the lenses exert the least amount of influence on my surface, and so this method, while seemingly backwards was a good work around for the situation. Like I said before, these lenses have been the hardest part of this project, but T-Splines made it much easier. I left myself with one last option for adjustment when I go to fabricate and assemble these - since the parting line between the top and bottom shell goes right through the lens, I can slide the top and bottom shells around a little before I join them together, using the actual lenses to ensure a perfect fit.
-Sky
- Attachments
-
- After manually "matching" the tip to the lens, the meshes get all fuzzy and mottled when they match up.
-
- Blue inner objects are the offset version, which is the corrected shape.
-
- Lofted outer ring to make surface
-
- Points were digitized, and used as the basis for curves
-
- Lenses affixed to shelf board with clay
-
- Faro arm on stand, 200 lbs of salt keeps it nice and stable.
I took some measurements, made some modifications and now I'm tantalizingly close to cutting molds on this one. The silver surface in the current rendering is for the tip light lenses. Those have been the absolute bane of my existence - since the molds for those were hand carved, getting the "ideal" tip surface to nicely mesh with what is a very "analog" surface requires quite a bit of work. 
. That sounds like a pain -- I guess you'd just load in some scanned curves and try to fit to them?
.
