Stair Stringer Fully Braced 9

[jechols]

I am designing a simple stair stringer that spans from landing to landing. The stringer is 18'-6" long (horizontal). The tread width is 3'-6". I am using LL=100 psf and DL=50 psf. Also I am using "Metal Stairs Manual" NAAMM STANDARD AMP 510-92. The manual states: "Stringers are considered to be laterally braced by attached treads and risers." The question is do you consider the stringers laterally braced by the treads and risers, and if so would you still consider the stringers braced if there were no risers just treads?

Thanks!

J

 

[Lion06]

I am very curious to hear thoughts on this. I have never designed stairs, but have often wondered about the bracing. If you assume a uniform load on the stairs and one beam gets loaded to the point of buckling (that the tread would help resist by virtue of it bracing the stringer), isn't the other stringer just as loaded and susceptible to buckling? What is to prevent the two stringers from buckling together? Unless of course one stringer is attached to a wall (but this is not always the case).

 

[JKW05]

I wrestled with this question for some time, and I finally convinced myself that the stringers are essentially braced. While I haven't done any real analysis on this, it seems to me that if the treads have any kind of mechanical connection to the stringers, that the force required to resist the buckling is probably small enough that it can be considered to be fully braced, or at least braced at the distance between the nosings. Additionally, the National Association of Architectural Metals Manufacturers (NAAMM) publish a "Metal Stairs Manual" in which all of their tables for stringers are based on being "considered laterally braced by attached treads and risers".

I'm not aware any of problems with stair stringer failures within our, so I have been comfortable with my conclusion. I do avoid using plain plates for the stringers for which the Metal Stairs Manual provide load tables for; at least even a minimal channel flange improves the buckling resistance. If anyone does know of stringer failures, please share the conditions in which that occurred!

Regarding StructuralEIT's question, if it is assumed that the tread/stringer connection has sufficient capacity, then could the buckling capacity for the entire stair run be based on the entire assembly with the "width" of the member equal to the width of the stair + the width of the flanges? This of course assumes that the single stringers have the buckling capacity for the length (distance) between nosings.

 

[csd72]

The bracing is easy to justify when you have a constinuous plate for treads as this basically acts like a (bent) diaphragm to take any lateral loads to the ends.

With discrete treads it is not as simple, even though the treads do act as a torsional restraint it may be hard to justify a fully restrained condition.

I would try the beam as unrestrained for the full length, I usually find that it makes no difference.

csd

 

[matrixeric]

I agree with JKW05 ...

- LTB of the individual stringers is local to the stringer itself.
- If one wants to consider buckling of the stringer+thread/riser assembly+stringer then the geometric properties of the entire stair assembly should be taken into consideration.

 

[Lion06]

I am not saying to consider buckling of the entire assembly. Well, maybe it appears that way when you visualize it, but if you have two discrete stringers with a tread between them, the tread can not possibly brace both of them without being connected back to something else. You consider one stringer braced by the tread because it is attached back to the other stringer, but that stringer is relying on that same tread being tied back to the other stringer, which we just said was being braced by the same tread. There is nothing to stop them from buckling together. The tread will not help with this situation. If both stringers are loaded equally and simultaneously (as would be the case with any uniformly loaded stair), I don't think you can use the tread as bracing unles one of the stringers is attached to a wall (or something similar).

 

[Bagman2524]

agree 100% with matrixeric. disagree with structuraleit. a brace spanning the compression flange of two beams braces both beams from LTB. The stringer does not need to be considered as an individual beam because the stringers and treads are all a connected unit. If the stringer is considered an individual beam, then the treads brace the stringers against LTB.

 

[Lion06]

then what is holding the tread in place so that it can brace the stringer? Best I can tell, the other stringer.
If these two stringers reach their buckling capacity together, how can one brace the other?
I would agree if you were getting some kind of composite action out of the treads, but I don't think you are.

 

[csd72]

StructuralEIT,

In this case it is not about direct connection to a solid bracing element, it is more like a horizontal vierendeel truss spanning to the ends and bringing the bracing loads out to the supports.

As I said above the treads do provide some bracing, but it is difficult to justify full restraint.

csd

 

[Lion06]

don't you need the members connected in triangles to form this horizontal truss to take the brace loads out to the connections?

 

[jechols]

A vierendeel truss by definition has no diagonal members. The web members are considered moment connected to the chords.

Thanks to everyone for your valuable comments!!

 

[JAE]

Most treads in stairs are pretty rigidly attached to the stringers with clip angles that are adequate for torsional restraint - which is what you need to resist LTB.

Visualize the stringer if you laid it flat (instead of upward sloping) - the treads now are very similar to vertical rigid diaphragms that you see on bridge girders.

But instead of being vertical, they are at an angle - perhaps as much as 57 degrees from the vertical(7" on 11"). But that is still adequate for bracing and there is one every 11" or so.

 

[UcfSE]

I agree with JAE.

 

[mfrad]

I have seen steel stairs with horizontal lengths longer than NAAMM tables not perform well even with the deepest channel sections. The stairs were designed as discussed in the above posts. When the stairs were isolated away from a parallel sheetrock or masonry wall, the effect of being isolated permitted uncomfortable vibration to occur from foot traffic. the calculate deflections of 100 psf live load were analyzed and found to be with code limits. Whereas even a 150 pound person walking down next to the handrail area made the stair appear to be a galloping gerty. A repair for this condition is either a vertical post to cut down span, or as we had done added a welded plate to convert the channel to a tube section.

I find that stairs which just touch side walls perform better and do not bounce using customary size of stringers found in NAAMM. this may be due to the fact that when the stair lurches, it leans on the side wall and provides some bracing laterally to the stair system.

 

[civilperson]

If the the treads are attached near the top of the stringer then the stringer is braced. Near the top would be within 20% of its depth measured from the top.

 

[COEngineeer]

If you have no risers, it will still be braced. Each thread is not on the same axis so the thread acts almost like vertical joists on a beam. Sorry I didnt read other comments. Probably someone already said the same thing. Good luck.

http://www.swijetty.com

Sea Water Intake and Jetty Construction

 

[jechols]

Thanks again for the comments!!!

J

 

[JAE]

civilperson,

I don't think the position of the treads relative to the top matters that much. Bracing against lateral torsional buckling can be in the form of either resistance to lateral translation of the compression flange or by resistance to torsional twist of the section.

In the case of the treads, they are really working as elements resisting twist (a rotational restraint) and not a compression flange translational restraint.

Almost any tread connection that has some depth to it would work (there are so many treads).

 

[apsix]

I would be wary of assuming that "Bracing against lateral torsional buckling can be in the form of either resistance to lateral translation of the compression flange or by resistance to torsional twist of the section" is always valid.

Quoting from the commentary to the Australian code AS4100;

C5.4.2 Restraints at a cross-section
The case of a cross-section that is restrained against twist rotation but free to deflect laterally is not included because of the difficulty of providing simple guidance.
Some very stiff torsional restraints may induce buckling modes in which the cross-section acts as if fully restrained (Clause 5.4.2.1), while in other cases torsional restraints may be comparatively ineffective, and the restrained cross-section may deflect laterally.

My view is that stairs are restrained by a combination of Vierendeel truss action and torsional restraint (but I've seen nothing to support this view).
Where these assumptions really affect design is if using flat plate as stringers. In my experience some practices are comfortable with using flat as stringers, while others insist on using channels. It would be good to see something authoritive on the subject.

 

[csd72]

apsix,

I think the something authoritive is the fact that these flat plate stringers are not collapsing. It doesnt take much to buckle an unrestrained vertical plate so if they are working as stringers then the treads must be providing restraint to some degree.

I have specified plate stringers but only with continuous folded plate treads.

csd