Need a book? Engineering books recommendations...

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

Wood connector allowable values

[Subject Prev][Subject Next][Thread Prev][Thread Next]

The results from the NDS are not the only non-intuitive requirements of 
codes.  Although I do not use it much, ANSI's Cold-Form Steel Design suffers 
from the same problem, namely, in a bolted tension member, if you increase 
the width of the section, the permitted tension stress goes down. (Eqns 3.2-2 
and 3.2-3, 1986 Edition)

Another illogical requirement of the NDS is the beam stability factor, C(L).  
In calculating the maximum effective length for a 2 X 12 that the formula can 
be applied to, you come up with an absurd effective length of more than *42 
feet!*  (Even an actual unbraced length of half that amount, 21 feet, is an 
absurd length to consider a 2 X 12 capable of carrying any kind of load.)  
In researching the original paper in which this formula was derived, I noted 
that the tabulated values of the shear modulus were of the same order of 
magnitude as the modulus of elasticity, instead of 1/19th the modulus of 
elasticity.  This was brought to the attention of AF&PA several years ago, 
yet nothing has changed.

I think that the reason for these "errors" in the code is that the basic 
principle of scientific research, that the results of all experimentation 
*must* be independently verified for it to be accepted, has been ignored in 
the interest of co$t.  Why is it that the same institutions repeatedly are 
the sole source for code information without independent verification from 
other institutions?

Why is it that the results of only *5* tests can get included in the 
connections section of the AISC manual and *assumed* to be valid?  I know 
that the narrative at the beginning of the tables states that the values are 
limited to *only* the configurations described because of the limited tests, 
but who reads this, and for those who do, how long do they remember it?

Research institutions need to return to the scientific principle and conduct 
more independent research, rather than rely on grants from trade 
associations.  The results of research from one institution should not be 
included in codes until it has been independently verified by other 

A. Roger Turk, P.E.(Structural)
Tucson, Arizona

Nels Roselund wrote:

. > I often calculate wood connection allowable values using the formulas of
. > Division III of Chapter 23 of the 1994 UBC; I've entered many of the 
. > formuals for bolts, lags and wood screws into Mathcad worksheets.
. > 
. > I try to make sense of the results before applying them to a project, and 
. > find that some of the Code formuals defy my attempts to give them 
. > intuitive meaning. A function of a variable raised to the power of 
. > another variable in the denominator gets me completely boggled. So I hope 
. > the folks who developed the formula were smarter than me and I just try 
. > to at least make some sense of the results.  
. > 
. > One example is the group action factor Cg of Section 2335.5.6.  One of my
. > often-used details is a bolted wood drag strut splice.  Intuitively, I say
. > that the greater the spacing for the bolts, the less likely is splitting 
. > of a drag-strut member when highly loaded in tension.  On that basis, I'm 
. > used to using long splices and 6" to 8", or even 12" bolt spacings rather 
. > than the minimum 4 diameter spacings.  The formula for Cg does not 
. > confirm my intuition. For example, for 8 3/4" diameter bolts in a line, 
. > for example, the bolts at 3" spacings have allowable values of 14% and 
. > 22% greater than bolts at 6" and 8" spacing, respectively. Would I really 
. > be doing a better design job with shorter splices? Or how about clusters 
. > of bolts at 3" spacing at each end of a long splice?
. > 
. > I have read in the NDS commentary, that the formula is taking into account
. > non-uniform loading of connectors in a row, but I'm still skeptical.  I'm
. > visualizing, for another example, a mudsill bolted to a foundation with 
. > bolts at 32", or 48", or whatever spacing is appropriate for the load to 
. > be transfered into the foundation at a shear wall.  Calculated with Cg, 
. > 3/4" diameter bolts at 32" have about 60% of the capacity that the Code 
. > allows if computed as individual bolts. Does anyone use Cg for 
. > calculating mudsill anchor bolts?  Should we?  
. > 
. > What can someone say to relieve my skepticism?  Buddy Showalter, are you
. > there?
. > 
. > Another mudsill matter: several years ago, the SEAOSC Existing Building
. > Committee had a subcommittee working on techniques to retrofit mudsill 
. > anchors in existing weak concrete. We found that pretty crumby concrete 
. > (concrete that you can't tighten a wedge-anchor into) can develop the 
. > strength of a 1/2" diameter anchor bolt in a 2x4 mudsill. We also found 
. > that at high loads, an anchor bolt being bent over by a displacing mudsill
. > splits the mudsill by cutting into it on the side opposite the load. We 
. > found that a square plate washer was very effecting in preventing that 
. > kind of damage. Bill Walker of Simpson participated in the tests, and 
. > soon after, Simpson began selling their BP series of plates. Now, I 
. > always require a BP-type plate on all of the bolted wood connections that 
. > I design for seismic forces. I'm sure, based on what I saw, that they 
. > will have greater capacities than bolts with a standard cut washer, but I 
. > use the Code values, figuring we're getting better seismic response at 
. > small extra cost. But, It would be good to know how to use the BP-type 
. > plates efficiently. Are there any values, or any testing going on, for 
. > bolted-wood connections using steel plates instead of cut washers between 
. > the head or nut and the wood?
. > 
. > Nels Roselund
. > Structural Engineer
. >