Need a book? Engineering books recommendations...

# Re: Plywood rigid diaphragms

• To: seaint(--nospam--at)seaint.org
• Subject: Re: Plywood rigid diaphragms
• From: Mlcse(--nospam--at)aol.com
• Date: Mon, 31 Aug 1998 03:07:52 EDT

```In a message dated 98-08-29 10:44:11 EDT, you write:

<<<
Michael Cochran says:

"Los Angeles has adopted a new voluntary ordinance called Division 93:
Voluntary Earthquake Hazard Reduction In Existing Wood Frame Residential
Buildings with Soft, Weak or Open Front Walls.   In this ordinance, you are
required to calculate shear wall deflections to show that the deflection is
less than .005H.   The shear wall deflection equation is from the UBC Volume
3."

I think this is good but they could have gone further by requiring that, for
walls in the same direction, if the relative rigidity is such that one wall
line is 50%(or any number thay want) or more rigid than any other wall line,
the distribution of lateral loads to the shear wall line should be by
relative
rigidity assuming a rigid diaphragm. But they should require that the
distribution at each wall line should not be less than the value obtained
using the traditional tributary width-flexible diaphragm analysis.>>>

I agree that the design load to wall line should not be less than what you
get by tributary area, and that you can not reduce the load because of rigid
diaphragm redistribution.  I don't think that the use of the term rigid is
completely appropriate.  Lets say you have 5 parallel wall lines separated by
30 feet between each wall line, as you go from left to right, with initial
calculated deflections of  0.25", 0.3", 0.35", 0.2" and 0.4" respectively.  I
don't imagine the deflection of the far right wall line with 0.4" defleciton
is going to transfer any (very little) load to the left wall which initially
had a deflection of 0.25".  The wall with 0.2" will see more load, and its
deflection will increase as the far right wall with 0.4"calculated deflection
initially moves.  The stiffer wall lines will pick up load as the adjacent
wall lines deflect, but not like a rigid diaphragm which has a center of
rigidity.

<<< You said "Using this equation, the biggest part of the calculated shear
wall
deflection
is due to holdown slip (minimum 1/16" oversized hole in post for holdown
bolts,.............................."

How about requiring that all holdowns be tightened such that we finger
tighten
the nuts on the bolts to the posts or studs first, then, finger tighten the
bolt on the anchor and then turn it three(or less/more) revolutions with a
wrench before finally tightening the post bolts. This should reduce the
holddown's contribution to the shear wall deflection. >>>

This is a good idea, but if you still have some wood shrinkage at the sill
plate or the floor joist, now the bolts are no longer tight in the post holes
and you may have the 1/16" oversized hole effect again.  In evaluating
existing buildings, you may find that the actual post hole is greater than
1/16" and that the combined shrinkage of the rim joists and wall sill plate
and double top plate may be 1/2" or more.  This shrinkage can play havic with
the shear wall deflection

<<< You said "This becomes a nightmare when you try to distribute the load by
rigidity
though, since you do not know the rigidity of the plywood shear walls,
........."

Yes we do not have an exact formula for plywood shear wall deflection
considering the contribution of the holdown (and don't forget the slip in the
sill bolt on the sill plate hole). But if we consider the "relative" rigidity
using the inexact but approximate deflection formula to distribute the
lateral
forces, since they will all be subjected to approximately the same
unpredictable and unquantifiable conditions(assuming we are analysing all
plywood shear walls), then this should be good enough, at least better than
not considering relative rigidity at all.  That is, until we have more test
and results converted to formulas which the engineering community will agree
upon and accept. >>>

Remember that some walls along a given wall line may have significant gravity
since the joist are running parallel to the wall.  The gravity loads greatly
influence the wall overturning resistance and holdown size.  It is hard to
determine the relative rigidity when the holdown movement could account for
50% or more of the shear wall deflection for some walls, but significantly
less for other walls.  The closest I think you can come to using your
suggested approximation method would be to set all holdown deflections to the
same amount, lets say uplift is always 1/4" which then the holdown assembly
component of deflection would be (1/4)(h/d).  This would take into affect wall
length effects.

<<< You said "I think the concerns are that if you design strictly by
tributary
area, you
may end up underdesigning certain walls.  In a simple box shaped building
with
perimeter shear walls only, the short wall on one side is designed for 50
percent of the tributary area, the long wall on the other side is designed
for
50% of the tributary area.  The short wall deflects (never sees the full 50%)
and the long wall now sees 70-100% of the load when it was actually just
designed for 50% of the load.  Now the question is how to come up with a
reasonable solution, which can be implemented within the design fee structure
that as engineers we typically get for designing wood structures.

Any suggestions?......."

See my first suggestion about checking the relative rigidity of shear walls
in
the same direction. Let's do some trial calculations of relative rigidity of
different combinations of full length plywood shear walls at the back and a
series of short walls at the "soft" front, using the only deflection formula
we have so far. Then pick a combination that we feel warrants a distribution
other than our standard "50-50"distribution, convert it to a percentage of
relative rigidity difference, round it off to a "nice" number(which I notice
a
lot on our code) and use it as a criteria when relative rigidity calculations
should be required. Again, use the greater of the values obtained by using
both the relative rigidity distribution and the traditional distribution.

These are my suggestions. And I'm intersted in finding out what SEAOSC or the
presenters in the November seminar on this topic are going to say.

>>

I like the idea of making parallel shear wall lines have similar deflections,
lets say within 10% of the adjacent wall line.  You then balance the load
between wall lines until they all have the same deflection within 10%.  When
redistributing loads to adjacent wall lines, you only check that wall line for