| ~ COMPOSITES|
This week I tried something I've never tried before: composite layup. It was definitely an experiment but I learned a lot.
I decided I wanted to try to make an arm that would be used as part of the digitizing probe in my final project.
I designed the part in Solidworks and prepared them to be milled on the Shopbot out of 2" blue insulation foam. This went smoothly and the parts came out really nice looking.
Some tricks I found useful for doing 3D parts on the shopbot:
Of course, I made the stupid mistake of not mirroring one of the sides.
So I made a third...
Once I had both parts, I glued them together with wood glue. This was harder than I anticipated because the foam wasn't incredibly stiff and the clamps were rather heavy (so it was very easy to add an unwanted curvature). I ended up rigging up a clamping fixture that didn't seem to strain it too much. Unfortunately... I forgot to take a picture.
To prepare the part for laying-up I just went over it briefly with the hot-air gun an some sandpaper. I also cut out specific shapes from the linen in the hopes of having the composite match the foam-core as closely as possible. What I thought would be a few quick snips turned into an hour or so time-commitment. I retrospect I think I should've taken faces from the 3D model, tweaked them (to add flanges and such), and laser cut the linen.
Now it was time for laying-up and vacuum bagging... I followed the steps that Kenny demonstrated:
Laying-up was really hard with my part. Even with my specifically cut linen pieces it was difficult to get the pieces wanting to lie in the right place. The pockets I added on the side of the part made everything much more difficult.
I forgot to add the breather bridge from the tube to the breather layer under the part so when I first hooked it up to the vacuum pump only part of the bag was fully evacuated. To fix the issue, I cut into the bag, added the breather, and resealed it.
I waited 5 hours to take the part out. At that point it was almost fully cured but I certainly could've waited another hour or two for it to harden fully. It came out alright but has a ton of wrinkles. In the future I might try wrapping thin strips of linen across the entire part rather than having a large piece for each face.
Since my part was a little complex and didn't come out quite the way I wanted, I tried laying-up another object. I used a hydrostone cast I designed this summer at Haystack (using Neil's fab_server). I did all the same steps as above but used mold-release before laying-up.
Kenny suggested I try to cut and sew pieces of linen together to better approximate the curvature but I found that since each corner is radially symetric and is just a decaying sinusoid, a simple quarter circle worked perfectly.