Need a book? Engineering books recommendations...

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

RE: Sawtooth Roof

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

When I refer to the roof as a sawtooth, I mean the roof slopes up on a 5 on
12 slope, then there is a vertical face (originally with sash) as the roof
drops straight down to a valley, then it slopes back up over the next 30'
bay and so on.  In other words it is a series of monosloped surfaces.  Hence
the purlins do not tie together at the ridges.  What we would have done if
we had designed this building is to run sag rods diagonally from the
midpoint of one of 12" beams to the 1/4 point of the next, and from the 1/4
points back to the column lines (where the main girders would act as
compression struts).  This would make a Pratt truss in the plane of the roof
with the purlins being the verticals, the 12" beams the chords, and the sag
rods the diagonals.  That is why, without sag rods, there is nothing but the
weak axis stiffness of the 12" beams and, more significantly the diaphragm
stiffness of the 2' x 6' precast panels to resist the downhill force.  It
seem to me that the most critical time was during the original construction
when the adhered roofing was not yet in place to help tie all the panels
together into a diaphragm.


> -----Original Message-----
> From:	Jim Kestner [SMTP:jkestner(--nospam--at)]
> Sent:	Wednesday, December 09, 1998 5:48 PM
> To:	seaint(--nospam--at)
> Subject:	Re: Sawtooth Roof
> Ed:
> If I understand you correctly, aren't the 6" purlins acting as sag rods?
> How do
> they tie together at the ridge? Are the ridge beams strong enough to
> resist this
> accumulated downhill force from the 6" purlins? Wouldn't this keep
> everything
> from sagging downhill? The 6" purlins provide lateral support to the 12"
> beams
> as well as take out the lateral force @ 6' O.C. How does this calc. out?
> I agree that the angles must be acting as keepers to prevent sliding of
> the
> panels.
> Jim K.
> Ed Marshall wrote:
> Jim,
> Thanks for the comments.  A detail I didn't include in my original
> description is the presence of L2x1 1/2 angle iron "ledgers" which are
> attached to the top of 6" wide flange purlins (which run up and down the
> slope at 6' o.c) after every 4th precast panel (8' clear between these
> angles).  Their purpose appears to have been an erection aid in that they
> keep the precast panels from sliding down the slope.  12" wide flange
> beams
> spanning 30' running parallel to the ridge (set with their webs
> perpendicular to the roof surface) support the purlins.  These 12" beams
> are
> located on 10' centers.  So, in addition to friction and the butterfly
> clips, the angle iron ledgers tie the purlins to each other, and the
> purlins
> tie the 12" beams to each other.  However, to keep everything square (and
> not sagging downhill) there is only the weak axis stiffness of the 12"
> beams
> and, more significantly the diaphragm stiffness of the 2' x 6' precast
> panels along with many layers of adhered roofing felts and shingles.
> Ed
> -----Original Message-----
> From: Jim Kestner [SMTP:jkestner(--nospam--at)]
> Sent: Wednesday, December 09, 1998 3:29 PM
> To:   seaint(--nospam--at)
> Subject:      Re: Sawtooth Roof
> An old rule of thumb is to use sag rods for roof slopes greater than 3 on
> 12. I
> prefer to locate the sag rods in the upper third of the beam since I
> believe
> they are more effective at that location. This helps provide lateral
> restraint
> as well as takes out the lateral force.
> If the deck is not positively attached to the beams to resist lateral
> forces,
> than you cannot count on it to provide lateral bracing or resist the
> lateral
> force component. I do not recommend counting on friction to provide
> lateral
> support!
> Many buildings, even those 60 years old, may not have ever seen their full
> design loads. That may be the case here, especially since you have a
> sloped
> roof. Signs of distress aren't always evident or even occur since these
> structures have to be loaded to approx. 1.7 or more times their design
> load
> before a failure may occur. Since there is a precast roof, the overload
> has to
> be even higher to cause a failure since you have to also overcome the DL
> safety
> factor. Just because there is no distress or failure does not mean there
> is an
> adequate safety factor provided.
> The best you can do is advise the owner about this situation and recommend
> what
> you consider to be the best construction practices to provide an adequate
> safety
> factor.
> Jim Kestner, P.E.
> Green Bay, Wi.
> Ed Marshall wrote:
> We are working on a project that, among other things, requires that we
> evaluate the safe load carrying capacity of an existing sawtooth roof of
> an
> industrial building.  The roof was designed for 30 psf live load.  The
> current building code (SBC) requires between 12 and 16 psf depending on
> the
> tributary area for the particular supporting member.  There may now be as
> much as 20 psf of equipment, utilities, or catwalks suspended from some
> roof
> purlins (in some cases the load has been present for many years).  The 
> question is how much load can safely be suspended from this steel.
> The building was built about 1940.  The slope on each sawtooth is 5 on 12.
> The building bays are 24' parallel to the ridges and 30' parallel to the
> slope.  The building is approximately 1000' long (with expansion joints)by
> 270' wide.
> The roof cladding is 2' x 6' x 1 7/8" reinforced precast concrete panels
> (135 pcf).  Each panel is attached to purlins with a pair of 14 gage
> butterfly clips at diagonally opposite corners (specified to have a
> driving
> fit on the beam flange).  The allowable stress for the supporting
> structural
> steel is noted as 18 ksi (the AISC code permitted 20 ksi for A7 steel at
> that time). The supporting steel was originally sized assuming that the
> individual beams were laterally supported, but there are NO SAG RODS. For
> a
> roof system of this type it has been our long standing practice to require
> sag rods between the purlins.  On a smaller building we might simply
> recommend to the owner that he add sag rods but on a building of this size
> the cost would be very substantial.  Moreover this roof has successfully
> stood for almost 60 years with no signs of distress.
> The original roofing is composition.  Above that at least one layer of
> asphalt shingles was eventually added.  Above these layers a sprayed-on
> membrane has been added.  We assume that the joints between the precast
> panels were pointed with a mastic when originally installed.  We are
> inclined to accept that the precast panels with the butterfly clips,
> mastic
> pointing, and adhered roofing acts as a diaphragm within individual bays
> of
> the building providing lateral support to the supporting steel up to at
> least the original 18 ksi stress level.
> Comments would be appreciated.
> Ed Marshall, PE
> Atlanta