Need a book? Engineering books recommendations...

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

Re: Determining Rho with a mixed system

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Dennis-
    Tie down deflection is critical to shearwall performance--both strength and
stiffness.  The tie down "protects" the nails in the sill adjacent to the
uplift corner.  Whatever post/sill separation the tie down allows, is the
deflection  the local nails have to accommodate--that's where most wall
failures start in the lab tests I've seen.
    A narrow shearwall amplifies the tie down deflection by the h/l ratio.
That's what Bob Kazanjy's tests show (although he doesn't seem to be
comfortable releasing the tests he did for Simpson).  In the calculations, the
only two terms that produce much deflection are the tie down and the nail
slip.  With narrow walls or "loose" tie downs, da is very significant to the
total wall deflection.  But if  you're just looking at relative stiffness, the
actual da is probably not so critical.  So far I'm considering da values
between .1" and .25"  (it's not clear to me whether the sill compression was
folded into the en calculations, but I'm sure that ordinary contruction slop
isn't).
    As to rho, check out the example on page 50 of vol. I of ICBO's Seismic
Design Examples.  It isn't just the highly loaded walls you need to check, 10/l
"helps" the long walls and penalizes the short ones, so you should check them
all as part of your spreadsheet.
    In the only "mixed" house I've done with moment frames so far, I applied
the same logic to both kinds of wall (which is the way I think the code
reads).  I calculate Ri for frames as the percentage of the story shear and
then calculate rho, just the same as the shearwalls are checked.  For moment
frames there doesn't seem to be anything comparable to 10/l, but rho must not
exceed 1.25.  Basically, the idea seems to be that you don't want too large a
percentage of the load in any one wall, especially short walls or moment
frames.  I don't know if that's correct, but it seems to be a reasonable
reading of the code--and the results have looked reasonable so far.
Chuck Utzman