On my website, I write extensively about my 1st attempt. It turned out pretty well and Marya and I have been able to print all sorts of things with it, including multi-color and double-sided jobs. Nevertheless, I'm dissatisfied. The oak bars are 3/4" wide, so we end up losing 3" from the usable length and width of the bed, a pretty high percentage. It's also hard to duplicate, requiring some woodworking for the oak and metalwork for the hinges.
I had the idea of using carbon-fiber pulltrusions, since I knew they were quite stiff, available in convenient sizes, and pretty easy to glue together. I buy them from a company called CST.
My first attempt was unsuccessful. I built the tympan frame using bars 1/8" thick by 3/8" wide, using simple lap joints to fasten the corners together. The area for the glue joint (3/8" x 3/8") was too small and the glue (epoxy) joints broke.
While I could use wider bits of CF to increase the glueable area, one of my main motivations is to recover usable area in the bed of the press; wider bars would defeat that goal. I discussed the problem with a friend of mine, Howard Rush, who can build anything. He came up with the idea of using L brackets to hold the corners together. Not only that, he helped me build them!
First, we needed some raw material. We built a plate of CF-epoxy, 1/8" thick and about 5" square (a similar plate can be ordered from CST, but we wanted to get to work right away).
Next, we laid out the 8 required L brackets on a sheet of paper and glued it (with 3M 77 spray glue) onto the plate.
Following the outlines, we cut out the individual bits, sanding them to neaten up the edges. We used Howard's table saw with a diamond blade, but it would be possible to use a Dremel tool with an abrasive cutoff wheel (aka Dangerous Disk!). By the way, I also used the Dremel tool to cut the bar to length.
Here are the pieces of the tympan frame, ready to go together. Before glueing, we sanded all the mating surfaces, then cleaned them with MEK to remove and dust, oil, etc. that might interfere with the glue bond.
We layed it out on a slightly oversized scrap of shelving, then worked for a while to ensure all the corners were square before glueing.
We held the bars in place with hand clamps and remembered to add some teflon (wax paper would be ok) to prevent glue from sticking to the shelving.
Next we mixed up a batch of high-zoot epoxy (Loctite Hysol 9430), weighing the 2 parts carefully, and brushed a thin coat on the mating surfaces. Finally, we clamped the L brackets in place, lightly (too tightly will squeeze out the epoxy). A slippery job, but it turned out ok.
Finally, we let it dry for a week (until I had a chance to visit again).
Next, we need to build the frisket frame. Same process, just sized a bit smaller.