deflection limits...not just l360

huangyhg

deflection limits...not just l/360
i know typically deflection limits for ll is l/360 and for d+l it is l/240 in general for stell framed construction(offices, apartments, etc...).  i have always used a rule of thumb of limiting deflection to 1/2" total, which usually causes a ratio in the neighborhood of l/500 or so and gives me much heavier sections.  
am i being too conservative limiting to 1/2"?
find a job or post a job opening
yes, you are too conservative, unless there is a special reason for limiting deflection to 1/2", such as being above a window, or something like that.
daveatkins
i generally agree with dave. when a beam is above a window, curtain wall or non-load bearing wall, you might want to coordinate with the architect to make sure that the slip track at the window head or wall is sufficient to accomodate the expected deflection. i suggest allowing some additional for construction tolerance.
i usualy use l/480 for live load and l/240 for tl.  unless i will have tiles then i bump it up to l/600.
so if you are allowing for l/240 for total, then you are allowing a 1" deflection on a 20ft beam, 1.5" on 30ft?  it seems like too much to me.
l/480 seems to keep floors under control...  vibration, etc.
l/600 for masonry or tile floors
l/360 for wood floors is ok....
if you have walls or partitions on the floor in question - these help a lot in killing vibrations and bounce.
30' x 30' great rooms without any walls are murder.  a 250 lb man walking across that will cause havoc unless you have at least l/480.  all the dishes in the hutch will rattle!!!
i am dealing with a 4story apartment above 3 story parking.  framing is steel beam & column, with a few masonry bearing walls, a few masonry partition (fire) walls, brick exterior facing.  the residential floors are 4" conc slab, and the parking is 6"conc.  i'm using 50ksi steel to hopefully reduce section size and floor to floor height is limited.
after the discussion here, i am considering just using l/360 as the limit for d+l deflection, that way i know i have l/360 for ll alone, and it will cut down on my calcs (i am having to do this all by hand & spreadsheets i've made).  also, i will adjust my numerical deflection to 1" allowable.
sound ok?
well, it seems a lot but really.. in order to get 40 psf in a room, you have to cramp people in the room like sardines plus it has to snow at the same time.  so your live load will probably never reach 40psf on the floor.
plus, you will also add 3/4" plywood and nail it to the joists.  you are creating a composite section so your joists are actually stiffer.  i would just worry about the defelction ratio instead of the actual deflection unless you are sizing a header above a huge window or something.  
keeping the precomposite dead load deflection of 1/2" to 3/4" is done to possibly avoid cambering of the beam.  sometimes cambering can be more economical.
total deflection is important, but for floor elements, live load only deflection limits also shall be checked per code limits.
burgoeng,
your building sounds large to be designing by hand.  if you plan on doing many of these buildings, you might want to look at purchasing ram structural system.  very easy to use, very fast, and not that expensive.  no need to buy all of the modules though.  just a thought.  make some money.
twinnell,
believe me, you dont need to tell me we need some software.  our office is not equipped to be doing designs like this and i think we sort of got talked into taking this project on as "favor".  being that i am the only one with structural experience in the office, it has fallen to me.  i have made it pretty clear to those in charge that if we are going to be getting more of these, we must buy something.  it's a lot of number crunching and a lot of guestimating to carry all the loads downward to support network below....and a lot of my time.
while we're on the subject, any opinions on ram vs risa vs staad?  i've used staad in the past at a previous job, but it'd be rather pricey for this office to absorb.