When designing the floor system, because of the large variation in needed support height (1” near the “shallow” end, ~15” near the “deep end”) and the concrete floor’s compound slope, I had decided against using tapered PT joists (2×12 / 2×10 / 2×8 / 2×6…) anchored directly to the concrete floor. Instead, the floor would be supported on 12’ 2×4s whose ends attach to the perimeter using Simpson joist hangers, with 2 additional supports at the 1/3 points.
My web research had turned up other high-elevated-subfloor builds using basically this approach, and the pros considered the results solid enough for their clients. Interpolating from building code span-vs.-deflection tables gave me confidence that 2×4 joists would have acceptably small deflection if supported every 48”.
For these interior supports I Tapcon’d short segments of PT 2×4 to the concrete. I pre-drilled the holes using the Bosch Bulldog roto-hammer that I bought for the project—it went through the PT and into the concrete like a knife through butter, with way less vibration than my old hammer drill.
Each of these plates is about 18” long—spanning one joist pair outside-to-outside, each supporting 2 vertical “studs” (each individually beveled and mitered at the bottom to match the compound slope of the plate). These uprights in turn support the underside of the joists. For stability along the joist axis I gusseted each joist/stud junction on one face with 1/2” ply. For stability perpendicular to the joists I added 2×2 “blocking” between the joists near each support point.
The resulting structure passed the “stand between the supports and and bounce” test with flying colors—the skeleton was solid! Now for the skin…
-- Accomplish the great task by a series of small acts. (Tao Te Ching / Lao Tzu)