Shop Products
Houzz Logo Print
kaseki

Reinforcing wood joists with steel strapping - reprise

kaseki
15 years ago

In the thread on this forum "Reinforceing floor ..." started by "Galroc,"

http://ths.gardenweb.com/forums/load/repair/msg0915531723856.html

respondent "brickeyee" provided considerable information on the approach of adding steel strapping to the underneath of the joist. The steel strapping approach argument that the steel will resist stretching is unarguable, but joist stiffness depends on whether composite action succeeds, or not.

I was enthused about this approach for a problem with my own house until I came across a paper by the Corps of Engineers titled "Evaluation of Expedient Techniques for Strengthening Floor Joist Systems in Residential Dwellings." (Available from NTIS as ADA013987.)

In this paper, experiments with various reinforced joist sections were performed, directed toward strengthening floors to resist nuclear blast effects by testing reinforcement approaches to failure. That isn't my goal, but some aspects of the steel strapping results caught my attention.

Three of the 15 test samples were steel strapped underneath. The steel straps were 1.5-inch by 1/8-inch, and were nailed to the joists with 16p nails 6-inches on center. The straps extended the entire length of the 16-foot joists, which were supported so that there was 14 feet of span.

The moment of inertia of the three 2 x 10 southern yellow pine joists used for a test section was calculated to be 296.8 inches^4. That for the steel reinforced joists was calculated to be 512.5 inches^4. However, while the measured moment of inertia for the unreinforced joists was 300.4 inches^4, very close to the calculated value, that for the steel strap reinforced joists was only 324.4 inches^4, a very modest increase for a lot of work. The inferred conclusion was that the performace was well below that expected from composite action. This conclusion is supported by the deflection vs. load curves that were published.

I am unable to see any obvious material difference between the experimental configuration and the description in answer to Galroc, although nail hole diameter in the strapping wasn't specified and may be key. So, has anyone on this forum tried the steel strap method and observed, even qualitatively, a significant reduction in sag or an apparent significant increase in floor stiffness?

kas

Comments (13)

  • brickeyee
    15 years ago

    "The steel straps were 1.5-inch by 1/8-inch, and were nailed to the joists with 16p nails 6-inches on center."

    Screws work better, but need to be in tightly dimensioned holes.
    Even 16d nails are a little small for the compressive strength of most wood types in this type of application.

    If the holes are large enough to drive nails it is unlikely you will get much benefit since the joist can still move until the hole tolerance is removed from the movement.

    If you want to use nails they should probably be power driven to produce a very tight hole fit.

    This method has been used for flitch beams to increase load rating also.

    I would like to see these guys try and make a flitch beam.

    Composite structures put a premium on correctly dimensioned holes and compressive loading of the wood.

  • kaseki
    Original Author
    15 years ago

    Thanks brickeyee.

    Certainly for the first part of the deflection vs. pressure curve oversized holes would let the strap system appear to be disconnected. As the pressure increased, one would expect the slop to be taken up and the slope of the curve to become sharper. That wasn't evident in the results.

    I have to assume that either the nails started to shear or pull out just from tension from inception of the applied pressure (which was their observed condition at joist failure) or the joist moved in some three-dimensional way (out of plane) as it sagged, putting the steel strip in torsion (and helping pull the nails).

    Certainly screws as you described in one of the other threads on this subject would help overcome this. I wonder if using angle (attached only at the bottom edge of the joist) instead of strip would help keep the screws in pure shear.

    kas

  • tessy78
    15 years ago

    im curious to hear brickeyee's response to this comment!!!!!!!

  • brickeyee
    15 years ago

    Steel strapping is a way to stiffen an already installed joist.
    A test to failure is not meaningful here.

    The numbers they calculated also appear way off.
    Steel has a modulus of elasticity mush higher than wood.
    The normal calculation is to convert the cross section of the steel to an equivalent wood section based on the ratio of modulus.

    I wonder what they measure for flitch beams?
    They will fail with oversize holes also.

  • lucas_gowesty_com
    14 years ago

    The problem is with the approach. Adding steel at the BOTTOM of the joist THEORETICALLY increases the area moment, ONLY if the steel is jointed intimately to the wood, thereby putting x-section of the steel in pure tension. The problem is that, in practice, joining the steel intimately to the wood is not really doable, no matter how many nails you used. The nails are put in a shear loading, the wood is not stiff enough, and metal just slides on the surface of the wood and never gets put into a tensile load. The better approach would be to install steel on the SIDES of the joist, oriented vertically. The taller the piece of steel, the greater the area moment, and the more it will resist bending ONLY if it is kept from buckling. This is where the wood comes in. Fastening the steel securely to the sides of the wood will keep the steel from buckling, and thereby increase the stiffness of the joist by an amount equivalent toe the area moment of the steel multiplied by the modulus of elasticity of steel, divided by the same for wood. The fastening system works here because the fasteners are put in a tensile load, and very little at that since the steel does not buckle. For example, a wood 2x4 (actually 1.5 x 3.5) has a modulus of around 12.5 and a area moment of around 5.36 inches cubed. Steel strap 3.5 inches wide and only 0.095 inches thick has an area moment of only 0.338 inches cubed, BUT steel has a modulus of around 200. So, if you use the formula for simple deflection, it turns out that steel that is only 0.095 inches thick and 3.5 inches wide is just as stiff as a wood that is the same height, but 2 inches thick. The trick is to keep the steel from buckling, which can be avoided by fastening every few inches with nails, screws, or bolts. The fasteners are under practically no load as long as the steel does not buckle, and of course it will not under such small deflections.

    Yours truly, S. Lucas Valdes, PEME, Cal Poly, SLO, 1984

  • mightyanvil
    14 years ago

    If the fasteners are sized and placed properly the steel strap works quite well. It's not necessary to transfer the entire strength of the steel strap to the wood joist; it only needs to be stiffened enough to reduce deflection and "bounce" to a more comfortable level.

    Here is a link that might be useful: Here is a more complete discussion

  • fixizin
    14 years ago

    Sorry to bring my engineering curiosity to the tail end of this discussion, but... wouldn't ANY scheme of cladding (for lack of better term) steel onto existing wood joists have to include FIRST UNloading (i.e. "jacking up") the joist to a neutral--or even negative--deflection, THEN securely/intimately attaching the steel?

    This would *pre-stress* the steel, such that it was carrying a significant % of the dead load, right from the go. It would also induce/force one to use a NON-slip fastening method. Intuitively, pre-stressing *seems* like the way to get the most "balanced" strength out of both materials... yes/no?

    Side Note: Being in a High-Velocity Hurricane Zone (HVHZ), and having a roofing system designed to hold concrete tiles which weigh 940-1080 lbs./square(!), I probably have the most serious type of mass-produced roof trusses, and the steel components are all attached to the SIDES (BOTH SIDES) of the 2x6 and 2x4 lumber components, none to the bottoms/edges... FWIW.

  • mightyanvil
    14 years ago

    Jacking up the beam would not prestress the assembly unless the steel is installed straight and level while the center of the beam is jacked up above level and shims are installed to keep them that way.

    An easier way of doing this is to forget the jacking and attach a steel strap only at the beam ends and then drive a opposed wedges between the steel and the joist at mid span. This is a very old method of reinforcing floor joists. Of course, the strength of this design is limited by the amount of anchorage of the steel strap but if reducing deflection by half is the goal, it works quite well.

  • fixizin
    14 years ago

    That's pretty ingenious and SIMPLE with the wedges... even better if the strap can be anchored not on the bottom, but "around the corner" on the ends.

    PS: DO they drive the wedges with a hydraulic jack, or just a BFHammer? ;')

  • b_b_com
    14 years ago

    We have a winner! The key is the steel will not help if the load is already being carried by the wood. You need to transfer a portion of the load to the steel and fixizin describes a good way to do it. You can see it recommended in this article http://www.taunton.com/finehomebuilding/PDF/Free/021184090.pdf

    * Posted by fixizin (My Page) on
    Wed, Apr 8, 09 at 18:30

    Sorry to bring my engineering curiosity to the tail end of this discussion, but... wouldn't ANY scheme of cladding (for lack of better term) steel onto existing wood joists have to include FIRST UNloading (i.e. "jacking up") the joist to a neutral--or even negative--deflection, THEN securely/intimately attaching the steel?

    Here is a link that might be useful: Stiffen Floors

  • brickeyee
    14 years ago

    "The key is the steel will not help if the load is already being carried by the wood."

    If this was true even sistering would have no effect, buts it is well known to reduce bounce in floors (though it can be very hard to do if mechanicals run through the joist).

    Bounce occurs when additional transient loads (like walking) cause even more deflection.

    Adding metal to the height (side) of the joist creates a quasi flitch beam, the problem that occurs is preventing buckling of the metal.

    In a full flitch beam the metal is trapped between two pieces of wood, preventing buckling under load.
    Flitch beams often fail to deliver their rating since the holes in the metal are over-sized.

    This problem can be reduced by drilling the holes with the metal between the wood members and the stack clamped together. Minimum holes size still needs to be used with multiple smaller fasteners being less likely to exceed the compressive limit of the wood under load transfer.

    I saw a report once that showed good results using powder actuated tools to drive nails through the flitch and clinched over.
    Like any other use of powder actuated tools getting sufficient impact force to drive, without over-driving, the fastener is an issue.

  • PRO
    CLA Engineers
    2 years ago

    mightyanvil,

    That is not the case here. Certainly you will get more tension if you insall straight and level with end attachment only and then drop the beam, but even jacking up the beam and installing from a central nail in firm contact with the joist will prestress by the difference in elongation between the compressed state of the joist when deformed into an upward deflection, and the tensed state when it drops. The difference in bottom chord length is very much measurable. Alternately, you can just tense the element directly with a cinch, come-along, or turnbuckle. Strapping tighteners absolutely exist. They use them in shipping. Be careful not to break the tensile strap though... that can get hairy.

  • millworkman
    2 years ago

    12 year old post and mightyanvil has not been active in many years.