Need a book? Engineering books recommendations...

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

RE: 1991 NDS Values for Lumber

[Subject Prev][Subject Next][Thread Prev][Thread Next]
About 20 years ago, I investigated the failure of the bottom chord of one
bowstring truss of 11 spanning about 110 feet over a bowling alley.  During
the design of the repairs (took approximately 3 days) an additional 7
trusses failed in the same manner.  The roof sagged but did not collapse.
Three of the 11 trusses had glulam bottom chords (over an addition to the
bowling alley constructed about five years after the original construction)
while the remainder had sawn bottom chords.  Two of the three glulam chords
failed. Six of the sawn chords failed.  

When checked under the stresses in effect at the time of the original
designs using the actual roof dead loads estimated from an inspection of the
actual layers of roofing on the structure at failure, both the sawn and
glulam chords were just under their allowable tension under DL + LL.  The
trusses had been in place about 25 (sawn) and 20 (glulam) years respectively
at the time of the failures.  The contractor making the repairs had done the
original construction.  They were a locally prominent long span roof firm
with a good reputation for quality work.  Their general superintendent told
me at the time (late '70's) that they were extremely distressed because a
number of the long span roofs they had constructed in the 1950's had
recently failed after about 20 to 25 years of service.  Considering this
anecdotal information, it appeared to me at the time that the allowable
tension stresses for long term loadings were too high by a factor of almost
3  (30 psf DL + LL (design) vs 10 psf DL on the structure at failure).  

As with the structure Roger reported, the roof collapse appeared to be
prevented due to the roof acting as a barrel shell (albeit with some rather
sizable deflections.).  After the failed roof trusses were shored
(essentially closing every third lane), the bowling alley resumed twenty
four hours a day operations during the repairs.  The General Manager of the
bowling alley said the loss of every third lane during repairs was
inconvenient but tolerable.  However, he said it would have been
catastrophic if the failed trusses had been over the bar!  [ROTFL]  That
apparently was where they made most of their money with the bowling
operation being a "loss leader" to get the drinkers into the bar.  

Repairs were effected by attaching a shoe to either side of the truss at the
heels and installing Stressteel bars along each side of each truss, then
replacing the failed members.  Original roof shape was restored by jacking
the shores until the roof was in position.  The bars were then stressed to
just pick up the DL.  We installed a pressure cell under the center shore
which went to zero load when the bars had the full DL.    

I had only been out of school for a short time.  The construction foreman
had about 45 years in the business.  When I told him the load on the center
shore would drop to zero when we had stressed the bars sufficiently to take
up the DL, he told me that no "college boy" would ever convince him that
jacking horizontally on the bars could possibly pick up any vertical load.
I really enjoyed the look on his face when the load dropped to zero on
center shore of the first truss to be jacked.  He immediately retired to the
bar muttering something about the world just not being like it used to be.
:<)

Regards,
Bill Cain, SE
Oakland, CA


	-----Original Message-----
	From:	Roger Turk [SMTP:73527.1356(--nospam--at)compuserve.com]
	Sent:	Monday, September 13, 1999 16:38 PM
	To:	seaint(--nospam--at)seaint.org
	Subject:	Re: 1991 NDS Values for Lumber

	Dennis Wish wrote:

	>>The problem is that the 91 standards is based upon the results of
a twenty 
	year study on the quality of wood. Although the quality of the wood
may be 
	the same after the '91 NDS was published as it was in 1980, the old
stress 
	values were allowed by code prior to adoption of the '91 NDS.
	In my opinion, there is sufficient room in the factor of safety (as
well as 
	the fact that this is a statistical average for wood) to accept the
prior 
	code stress when recalculating the member for revised loading or
remodeling 
	occuring after the adoption of the '91 NDS. <<

	Dennis,

	I would like to take exception to that statement.  I have witnessed
a 
	bowstring truss (101' span) that had (2) 3 X 12's as the bottom
chord.  Out 
	of 9 trusses, 7 bottom chords had failed in tension (from memory).
When 
	analyzed based on allowable tension stresses in effect when
constructed, the 
	trusses were completely adequate under dead and live loads.  *But*
when 
	analyzed based on the allowable tension stress after 1980, the
bottom chord 
	was seriously overstressed under dead load alone.  That the roof did
not 
	collapse was, in my opinion, due to the roof sheathing (diagonal
boards) 
	causing the roof to act as a barrel shell rather than as a truss
supported 
	roof.

	I would recommend that *all* analysis of older wood structures be
based on 
	the allowable stresses that resulted from the full size, in-grade
tests, 
	i.e., the values in the 1991 NDS and later.

	As for equivalent grades, Construction Grade was the commonly used
grade 
	prior to 1970 and is equivalent to today's No. 1 Grade.

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