bracing connections to hss columns

huangyhg

bracing connections to hss columns
i am designing connections for canadian steel fabricator, and very often come accross the situation where vertical bracing, w beam and column intersect. i think i read all books and papers ever written that mention brace/hss connection and all of them suggest shear plate or mounted tee or through shear plate welded to the column and field bolted gusset plate shop attached to the beam.
yet, in canada everybody still use "knife" connection for bracing connections-instead of shear plate, double angle welded to the column.
i asked the aisc solution center and reply was:
"double angle connections in which the outstanding legs of the angles are welded to the support are not suitable for axial load applications. this is because of the extremely limited capacity of the welds on those outstanding legs to resist the torsion due to eccentricity."
anybody help?
isn't a "knife" plate connection the same as a "through shear plate" connection?
term "knife connection" (locally) means double angle welded to supporting   
thanks for the explanation.  i work in australia, where we don't use double angle connections for anything.
the detail of this connection depends on if the pull or push of the brace goes thru the column (such as forces accumulating in a collector beyond the column or a knee brace frame or in a detail where the gusset plate is connected to the column instead of the beam) in those cases, there will be pass thru forces in the column or into the column which must transfer from the connection.
in the case of a knee brace or similar situation, we typically will use a tee shape built up out of plates so that the the pass thru force can transfer to the side walls of the tube and does not buckle the face of the tube.
i hope this helps!
thanks...
the thing is that most of the fabricators in canada still use double angle shop welded to column and field bolted to beam (knife)on all hss column connections (hss columns are used on 95% off canadian jobs)regardless if there is an axial force in joint or simple shear.
in my case, bracings are knee type and have reversible loads of 350kips (tension or compression).
the question is:
how you can prove that knife connection already used on previous jobs as default can take the axial load? and why is this connection "not suitable" for axial loads-don't quite get that "torsion on welds" answer.

do you have salmon and johnson's steel text?  they have a clear discussion of the strength of this connection under axial loads.  the response from the steel solution center doesn't quite do it.  the problem is not the weld, but the bending capacity of the angle leg.
if the outstnading leg is bolted, the bolts clamp the plies together and reduce the bending length of this outstanding angle leg.  this connection can get pretty strong under axial load.  if the tips of the outstanding leg are welded, they are effectively pins, and the bending length is now the full angle leg dimension.  you have simple supports and a long moment arm (for the welded configuration), versus fixed supports and a short moment arm (for the bolted configuration).  the welded scenario will be much weaker under axial load.
put some numbers to it and you'll see.
thank you!!!!
can you send me that text, it seems i didn't read all after all!
actually, i have that book!
thank you one more time-per this text,it seems that welded versus bolted is 4 times less capacity.