Metal Deck Diaphragm and a New Interior Masonry Wall
A single-story 240 ft × 120 ft warehouse has 8 in reinforced CMU perimeter walls 20 ft tall and a roof diaphragm of 1½ in, 22 gage galvanized metal deck, screwed to open-web steel joists at 5 ft on center. There is no concrete topping; the deck is the only diaphragm. The base shear is 135 kip, and the two long walls each take roughly half. The owner wants to add a 40 ft long, 8 in thick, full-height CMU wall inside the building to enclose new offices, framed tight to the underside of the roof joists and doweled to a new continuous footing. The contractor calls it 'just a partition.'
1. Rigid or flexible diaphragm? How does that assumption change the load that reaches the new wall?
Read the worked answer
Treat the roof as a flexible diaphragm. Two checks support it:
- ASCE 7's deflection criterion: a diaphragm is flexible when its midspan deflection exceeds twice the average story drift of the vertical elements. The CMU walls here are extremely stiff - 135 kip spread over 480 ft of wall gives a unit shear well under 0.5 klf and an in-plane drift of a small fraction of an inch. A screwed, untopped 22 gage deck spanning 120 ft at that shear deforms on the order of an inch. The ratio is far past 2, and it stays past 2 even if the deck's published shear stiffness is doubled.
- The span-to-depth shortcut does not pre-classify this roof (120 ft span over 240 ft depth), so the deflection comparison is the governing test and must actually be run.
Consequence: distribute the 135 kip by tributary area, not by wall stiffness. A wall that is stiff relative to the perimeter walls does not 'collect' extra load - the deck cannot deliver it. So the new wall is designed for the shear from roughly half the roof span on each side plus its own tributary, and the perimeter walls stay designed for their original tributary until the new wall is proven to take load.
2. If the owner wants the new wall counted as part of the lateral-force-resisting system, what has to be provided?
Read the worked answer
Counting it means designing it, and the list is longer than the wall itself:
- A continuous load path from deck to footing. The screwed deck must be able to deliver the tributary shear to the wall through a continuous collector/drag strut along the wall line, with fasteners and edge distances checked at the actual shear, not just at the deck's catalog capacity.
- Anchorage and chord. Positive connection of the deck to a wall bond beam, and a chord at the wall ends if the wall is acting as the boundary of a diaphragm segment.
- Overturning. The wall must resist out-of-plane and in-plane overturning with its own footing - dowels, footing weight and soil bearing, and uplift at the tension end. A 40 ft × 20 ft wall generates real overturning even at modest shear.
- Detailing consistent with the system. If the perimeter system is ordinary reinforced masonry, the new wall must be detailed to the same R factor, and it must be reinforced and grouted to match - not a nominal partition.
- Redistribution. Once the wall takes load, the tributary of the adjacent perimeter walls shrinks. That is only valid if the wall is actually connected and detailed to deliver; otherwise the perimeter walls remain sized for their original demand.
- Accidental torsion. The wall is off-center; check the resulting torsion on the flexible-diaphragm distribution.
3. The contractor proposes a ½ in gap at the top and a flexible closure so the wall is 'non-structural.' Evaluate that.
Read the worked answer
It can work, but ½ in is a number pulled out of the air and the wall does not become free just because it has a gap.
The gap must be sized, not guessed. It must accommodate the diaphragm deflection at that location plus the wall's own in-plane drift, and it must be maintained - a gap that gets packed with mortar or blocked with a rigid closure during construction is not a gap. For this roof, a flexible diaphragm moving on the order of an inch at midspan makes ½ in clearly inadequate.
Isolation cuts both ways. If the wall is isolated from the roof, it is no longer braced by the diaphragm and must be stable out-of-plane as a 20 ft tall cantilever from its own footing: reinforced, grouted, with a footing designed for the out-of-plane seismic and wind moments. An unreinforced 8 in wall 20 ft tall is not stable under any reasonable out-of-plane demand.
Document it. The drawings and the structural notes must state that the wall is isolated, show the gap, and forbid rigid closure; otherwise a future tenant will grout it solid and silently change the lateral system.
My recommendation: either design it as a shear wall (cleanest for the owner, more work now) or isolate it properly with a sized, detailed joint and a self-supporting reinforced wall.