unbraced compression truss chord

huangyhg

unbraced compression truss chord
i am being asked to design a custom truss with a completely unbraced bottom flange. the dead load on the truss is so light that there is a net uplift with wind load. this puts the bottom chord in compression. i don't like this condition, but i would like to have more than "i don't like it" for a answer, and if there is a safe way to design for this condition i would like to know how. i checked the bottom chord as a unbraced column the length of the joist and it worked fine. also, kl/r would be less than 200 if k is 1.0. is there some reason that k would be more than 1.0? is there some reason other than structural stability to brace the bottom chord? the truss is basically a bar joist made out of tube steel. i have two competing ideas in my mind. the one is bar joists where any joist with uplift always has uplift bridging. the other is a crane beam where the compression flange has no bracing. is there a way to calculate a stiffness that would make compression chord bracing unnecessary?
  
check out our whitepaper library.
is the bottom chord braced at least at the ends?  if you have checked the bottom chord for the compression that it will see and it doesn't need bracing anywhere besides the ends, i wouldn't have a problem with it.
the ends are not braced. that is the problem. if i had one run of bridging at each end of the truss, i would be quite comfortable with the design, but the owner and architect doesn't want any bridging.
if it's not braced at the ends, how did you do a design check?  you have to brace the ends or else it's like a gravity column that is supported on a horizontal roller at the base.  the only thing you can really do is look at the end vertical (or diagonal) and check it for out-of-plane stiffness and strength as a cantilever off of the top chord to see if that will provide adequate bracing per aisc.  i wouldn't like that, though, and would really push for the bracing at the ends.
what is it framing into?  can you run the bottom chord to a column/wall and have a knife plate off of the column/wall that gets slotted into the hss (but not attached), to provide the out-of-plane stiffness without restraining the movement in-plane?
structuraleit,
i'm not understanding why the bottom chord has to have the "ends fixed" in order for it to go into compression. obviously the top chord on the common bar joist is the bearing element, with the bottom chord stopping short of any support. the two are connected via diagonal webs. when i picture a common truss with a uniform or point load going up, i can easily see how the bottom chord develops compression.
wcw, it sounds like a stick model (computer model) is the next step. your k = 1.0 question is not apparent to me since the bottom chord is connected to the top chord by webs. i would venture to say that k = 1.0 since there is no bracing, but where i'm confused is if part of the bottom chord goes into tension and part does not (unbalanced loads?) i think i'm reading to far into your post and starting to guess, so i'll stop at that.
bigmig-
i didn't say anything about fixing the bottom chord in order for it to go into compression.  i didn't even say anything about fixing the bottom chord.  i'm not even sure how to respond because i don't understand your comment based on what i wrote in previous posts.
also, he is asking about the bracing because of using k=1.  you can't assume k=1 if the ends are braced.
when there is wind uplift on the truss the entire bottom chord is in compression. the compression starts where the first web member connects to the chord and reaches a maximum at mid-span. my question is how do you design a compression member that has no bracing in one plane? since there is no more compression after the last web member can i take l as the physical length of the bottom chord? what should i use for k? 2.1 or 1.0 or is k something even greater since i don't have even one end of the member braced in one direction. i thought about using k as 2.1 and l as the length of the   
bracing (obviously) provides lateral stability.  maybe you can provide some moment capability (not to say "cantilever").
by the sound of it, you've got bracing in one plane.  for the chord to move in the unbraced plane the bracing in the other plane has to bend, no ?  but clearly this in not a particularly effective way of providing a bracing effect.  so the chord could attempt to buckle like an euler column, with some restraint at the out-of-plane bracing; i suspect that the critical mode of failure is the lower chord deflecting mid-span in the unbraced direction.  
you can design a beam without flange bracing because the section is restrained against twisting (and translation) at the ends, hence the unbraced length.  if you have a compression member (e.g. a compression chord in a truss), you have to look at that member individually in addition to the overall   
rb1957-
that is exactly what i was talking about looking at several posts ago.  the diagonals are bracing it in the plane of the truss.  the only conceivable braces for the out-of-plane condition would be those same diagonals bending (acting as cantilevers off of the top flange) out of the plane of the truss.