Need a book? Engineering books recommendations...

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

Re[2]: L in deflection criteria

[Subject Prev][Subject Next][Thread Prev][Thread Next]
     Didn't notice this before. How about giving me reference that I could 
     look up. This same formula has been adopted by our local code here 
     (actually, a copy of the UBC, ACI, AISC combined in one "Code") for 
     years (as I can remember).

______________________________ Reply Separator _________________________________
Subject: Re: L in deflection criteria
Author:  MIME:SHERMANWC(--nospam--at) at INTERNET
Date:    7/18/98 1:28 AM

Roger Turk wrote: 
>..."L" should be the distance between points of inflection.   
>With regard to a cantilever..."L" equal to twice the cantilever length 
>be a conservative assumption ...  
I do not agree with this.  I think that "L" should simply be the length of 
member relative to its support points, i.e. for a cantilever, the length 
support point to the free end of the member; or for non-cantilevered 
the length between support points.  The concern is maximum total deflection 
relative to the support point where "zero" relative deflection is assumed - 
why should the type of beam curvature matter? 
But that raises another question: what if you have a secondary framing 
supported by a primary framing member?  The total deflection relative to 
column supports would be the sum of the member deflections (but I would only 
check each individual span per the code). 
>The requirements of the UBC (Table 16-D, 1997 UBC) have, for non-cambered  
>construction, been in error for as many years as I have been an engineer... 
>In essence, the loading for the deflection limitation for L/360 should be 
>less than LL+K(DL), where K is the percent of initial DL deflection that  
>represents the creep that would occur after the brittle finish is applied. 
>In non-cambered construction, the loading for the L/240 limitation should 
>not less than LL+(1+K)DL. 
You raise a good point here - the code formulas do appear to be in error.