Return to index: [Subject] [Thread] [Date] [Author]

Re: Curved Stair

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Brad, I did an analysis on a similar but narrower curved stair a
professional welder had already built on his own and installed in his new
home. The stringers were strip steel without flanges, and the treads open,
without risers. The method Fountain Conner suggests is a good starting
point. Because of the curve in your stringers, their flanges want to move
sideways, and the treads have to hold the stringers from rotating about
their lengthwise axis. The whole stair as a free-body has calculable end
reactions to the stringers, and not in the direction to be assumed from
habit. Any portion of the stair as a free-body can have external loads,
shears and moments calculated, and the analysis  assembled like a dinosaur's
vertebra in a museum. A tip on the flanges is to temporarily model them as
straight between tread intersections. The welder built his stair that way in
fact, and it calc'ed out as a very efficient design as well as being
geometrically precise.

Roger Turk mentioned a curved wooden stair. I did one for a U.C. math major/
woodworker for an elegant home where large gatherings were expected. It was
a box beam and torsion box, glued up with boatbuilder's epoxy off-site on a
jig, from hardwood lumber and Baltic Birch plywood (Marine grade DF plywood
was inadequate in rolling shear.) End reaction connections at each floor
were challenging because of the 90 deg curve and resulting axial torque
simultaneous with vertical shear. We intended the stair to be fit to support
a finite elephant.  

Charles Greenlaw

At 04:13 PM 06/01/2000 -0700, you wrote:
>The stair is 6'-6" wide with outside radius of 13'-6" and inside radius
>7'-0".  The stringers are made from C12 and the risers and treads are 12ga
>closed system welded to the stringers. The arc is 90 deg.
>Are there design charts or tables that have already been prepared for stairs
>of this type?  If not, are there shortcuts to design other than doing a
>finite element anaylsis.
>Brad Friederichs
>VE Solutions, Inc.
>(916) 505-0519