Framing a Cold-Formed Steel Partition

Spacing is doing two jobs at once

Stud spacing settles a layout question and a load question in the same figure, and the two pull in opposite directions. On the layout side, 400 mm and 600 mm centres (16 in and 24 in) exist because a 1200 mm or 4 ft board lands on a stud either way, and anything else means cutting boards to find a bearing edge. On the load side, spacing is the tributary width of the stud — widen it and every stud takes proportionally more of the pressure on the wall, which shows up directly in the deflection check as a larger line load. Going from 400 to 600 mm removes about a third of the studs and adds half again to what each survivor has to resist.

So the count is a consequence of two decisions, and it is worth taking off properly rather than dividing length by spacing in your head. The field count for a straight run is the spaces plus one, which is where the extra stud at the far end comes from. What that arithmetic does not contain is every stud placed for a reason other than rhythm: corners, T-junctions into other partitions, abutments to columns and mullions, door jambs — usually boxed or back-to-back and often a heavier member than the field — and backing studs behind anything wall-hung. On a cellular fit-out those extras have been counted at a quarter of the field count before now.

Start the layout from the end that makes the boards work, not from the end nearest the door. A run set out from the wrong corner puts the last stud a few millimetres past the last board edge, and the fix is either a cut board on the most visible wall in the room or an extra stud that nobody priced. Where the wall is not a straight run — a curve, a segmented bay — spacing stops being a constant and becomes a chord calculation, which is a different problem with a different answer at each end.

Putting the section against the limit

The check itself is ordinary elastic beam theory: a stud spanning floor to head, pinned at both ends, uniformly loaded across its height, deflecting 5wL⁴/384EI at midspan. What makes it a cold-formed problem is the I. A cold-formed section buckles locally in its own flat elements long before the material yields, so the stiffness that survives is an effective moment of inertia derived through AISI S100 — not a value you can compute from the web depth and flange width on the requisition. It is published, per size and per mil thickness, in the manufacturer's tables: SSMA's product technical information, or a fabricator's own literature such as ClarkDietrich's. Check whether the figure you have taken is for the punched or the unpunched section, because the standard punchout most manufacturers roll is a hole through the web and the two values are not interchangeable.

The load side needs converting before it goes anywhere near the formula, and this is where most people lose a factor. Codes give a pressure; the deflection check wants a line load along one stud. The International Building Code puts a minimum lateral pressure of 5 psf (0.24 kN/m²) on interior walls and partitions above about 1.8 m (6 ft) in height, and where the wall faces real wind — an open atrium, a facade set-back, a pressurised shaft — the pressure comes from ASCE 7 instead and is very much larger. Multiply that pressure by the stud spacing in metres and you have the N/m the checker asks for: 5 psf across 16 in (406 mm) centres is about 97 N/m, and across 24 in (610 mm) centres about 146 N/m.

The limit is a project decision more often than a code one. The building code's deflection limits table sets the floor and is written around whether the finish is brittle or flexible; the tighter ratios come from the finish itself. Ordinary taped and jointed board is generally run at L/240. Plaster, a rigid render or a tiled face pushes it to L/360, and stone, a masonry face, or glazing framed into the partition head commonly brings L/600 — and that figure will be in the finish or glazing manufacturer's literature, not in the code. Because the limit denominator enters linearly, the arithmetic is unforgiving: on this 6.4 m wall at 97 N/m, L/240 asks the section for roughly 0.40 × 10⁶ mm⁴ (0.96 in⁴) of effective inertia, and L/600 asks for just under 1.0 × 10⁶ mm⁴ (2.4 in⁴). Two and a half times the stiffness, for a decision that lives in the finishes schedule.

Height is worse than that, because span enters as the cube once the limit is expressed as a ratio. Take the same wall from 4.9 m to 6.4 m — a third taller — and the required effective inertia goes up by a factor of 2.2. This is why a partition schedule that works on a typical floor stops working across an atrium void or a double-height reception, and why "same wall, just taller" is the most expensive sentence on a fit-out. It is also why widening the spacing to save studs is so often a false economy: the stud count drops by a third and the section has to grow to take fifty percent more load, which usually costs more in steel than it saves in labour.

1. Take the design pressure from the project documents — the code minimum for an interior partition, or the ASCE 7 figure where the wall sees real wind.
2. Multiply it by the stud spacing to get a line load per stud, in N/m.
3. Look up the effective moment of inertia for the exact designation and mil thickness in the manufacturer's table, noting whether it is punched or unpunched.
4. Take the clear span as slab to head, not floor to ceiling, unless a bracing line is genuinely anchored back to structure.
5. Run the check at the limit the finish demands, not the code minimum, and if it fails change the member or the spacing rather than the limit.

Two runs of track, and only one of them may be rigid

Track is the quantity that surprises estimators because it does not follow the studs. Whatever the spacing, a wall run needs one continuous run at the floor and one at the head — twice the wall length, plus waste for laps, corners and boxing round every opening. On the 30 m atrium wall that is 63 m (207 ft) at a five percent allowance, and the allowance is the part people leave out: every door and every borrowed light is framed with track, and on a cellular floorplate that quantity is well past a rounding error.

The floor track is the straightforward one. It is set to the layout line, fixed down at the spacing the design or the installation standard calls for — ASTM C754 for nonstructural framing receiving screw-attached board, ASTM C1007 where the studs are load-bearing — and the anchor into the slab is only as good as its evaluation report. Powder-actuated pins and screw anchors both have published conditions covering base material, edge distance and embedment, and a post-tensioned slab is scanned before anything is fired or drilled into it, without exception.

The head is the one that goes wrong, and it goes wrong in a way that is invisible until the structure moves. A partition under a slab, a beam or a joist has to let that structure deflect without picking up load, which means a deep-leg deflection track, a slotted track, or a proprietary clip system — never a standard track with the studs screwed hard to its legs. The movement allowance is a number on the structural drawings, not the 19 mm somebody remembers from the last job, and the whole detail has to survive being boarded: the board is held short of the structure at the head, the fasteners stay out of the movement zone, and where the partition is rated the joint is a tested head-of-wall system to ASTM E1966 or UL 2079 rather than a bead of whatever was on the trolley. A studded-through deflection head is not a slow failure. It cracks the first time the floor above is loaded.

Counting screws you will actually buy

Framing fasteners are the last item priced and the first item to run out, mostly because "four per stud" gets treated as a fact rather than as a shape. Two screws at the top connection and two at the bottom is the ordinary arrangement for a stud seated in plain track, and on the atrium wall that is 76 studs at 400 mm centres — seventy-five spaces plus one — for 304 screws in the track connection alone. Take the same wall to 600 mm centres and it is 51 studs for 204 screws, which sounds like a saving until you remember what the previous section did to the required section.

But the top connection is exactly the one that may take none. At a deflection head the whole point is that the stud is free to move within the track, so the fastener count there is whatever the head system specifies: nothing through the legs on a deep-leg detail, screws through slots on a slotted track, or the clip manufacturer's schedule on a clip system. Structural studs, jamb assemblies and anything in a tested fire-resistance design carry their own connection schedule and it is usually more than four. Treat the screws-per-stud figure as the thing you looked up, not the thing you assumed, and the count comes out right.

The screws themselves are specified, not generic. Self-drilling tapping screws for cold-formed framing connections are covered by ASTM C1513, and the point number is chosen against the total thickness being drilled — that range is in the fastener manufacturer's literature and nowhere else, so a point that drills 43 mil track and stud together is not the point for a boxed jamb of two 68 mil members. The head matters too: a low-profile wafer or pan head at the frame so the board sits flat, a bugle head into the board, and never the reverse. A screw that has spun rather than drilled has cut its own hole oversize and holds nothing, which is worth catching on the first bay rather than the last. AISI S100 sets minimum spacing and edge distances for screwed connections in multiples of the fastener diameter; check them rather than eyeballing the flange.

The number that catches people out is not the track connection at all. Two rows of bridging on a channel-and-clip arrangement, at four screws per stud per row, is another 608 fasteners on that same wall — twice the entire stud-to-track total, and it appears in none of the counts above. Add the stitching in boxed jambs, the strapping at openings, and the fact that board screws are a separate quantity on a separate schedule entirely, and the honest answer is that the frame on a wall this size consumes fasteners in thousands, bought in boxes, in more than one specification.