I created a spreadsheet to calculate the geometry of the truss prior to construction. The inputs to the calculation are the member sizes. The spreadsheet graphs the members of the truss and gives node positions and member angles. Through a little trial and error, the desired geometry is attained by the correct member sizes, and the next step is construction.
I decided to design the bridge deck with a little camber, which is defined as a "slight convexity, arching, or curvature". The bridges' camber can be seen best in this picture.
This was done to ensure the deck of the bridge didn't sag below the supports when under load (from self weight and the trains). Also, the tolerances of the joints stack up and result in additional deflection. I designed for the center of the bridge to be exactly 2 brick widths (about 16mm) above the supports. As it turned out the bridges didn't sag nearly that much, probably only 1/2 a brick width.
The vertical clearance under the portal frame cross bracing was about 13 brick-heights (measured from bottom of track). The tracks were spaced at 10 brick widths (center to center). This allowed for two 8-wide trains to pass with 2 studs of clearance between them. As it turned out, even this wasn't enough! Several cars crashed into the bridges at the TEXLUG meeting.
I modeled my bridges loosely after the railroad bridge over Lake Amistad, just west of Del Rio, Texas.
There are a few small differences, though! The real bridge has 6 spans and accomodates only a single set of tracks. I built only 2 spans, but with double tracks.
