The heart of the table is a support structure for the thin plywood top, and the heart of the support structure is a lattice of I-beams. Each I-beam has a 1/4" plywood core with 1” tall, 3/4" wide top and bottom rails. Five beams run the full length of the table. They have extra side caps on the ends where the legs mount and vertical spacers about a third of the way in. Those spacers also serve as attachment points for  the cross-bracing. 

The beam rails have grooves to capture the plywood core. I cut each groove in three passes on my table saw. I cut a 3/8“ deep slot , flipped the board horizontally and cut another slot (so the grooves was centered), then removed any excess from the center with a final pass. 

 

 

The visible part of the plywood core is 2” wide, so I cut 2” wide planks for the side caps and spacers.
 

Glue up was easy. I just ran a bead inside the groove, inserted the central core, and added the caps and spacers. I used a spacer strip to position the spacers. 

 

 

I used short bar clamps on the rails and spring clamps on the spacers and caps. 

 

 

Unfortunately, that clamping approach had a problem. It kept the top and bottom rails parallel, but it didn’t keep the beams perfectly straight. I should have clamped the beams to my bench to keep them straight. 

 

To correct the problem, I jointed the curved beams by clamping them to a straight board and flush trimming them with a router. Then I ripped them all to the same width. My beams ended up about 1/8” narrower than designed. 

 

I used a crosscut sled to square the ends of the beams. I also used two roller stands. One supported the beam so it didn’t dip, and the other acted as a length stop. 

 

Initially, I had not planned on using the cross bracing. After I decided to add it, I made two more  beams (without spacers or end caps) and cut them into 5” long segments (not pictured). 

 

In the original design, there are racetrack-shaped cutouts in the beams’ plywood cores for weight reduction. I didn’t add them because the weight savings didn’t seem to justify the effort. (I can’t remember exactly how much weight it would have saved.) Given that my tables ended up heavier than I expected (44 lbs.), I may have been too hasty. One day I’ll redo the calculations. There’s no way I can route out the racetrack shapes now, but I might be able to cut out a bunch of 1 1/2" diameter circles.