deflection limit 2

[Thomas2121]

I have this 12m long rafter supporting light weight roof (2m span) on one side and terrace on the other. The max deflection of the rafter is 36mm under live load, which passes the limit check for the rafter itself.

However, if I look at the 2m long purlins of the roof that span from wall to the mid span of rafter. The deflection of the purlin will be 0 at one end (wall) and 36mm at another, which is L/55. Is this acceptable? Will it cause water leaking issue and the like or am I overthinking?

 

[Thomas2121]

I know typically we just check the deflection to its own length. But what is the reason we don’t check the deflection to the secondary members that connect to them?

 

[phamENG]

Not sure about Australian standards, but in the US the codes only dictate deflection along a member's length and don't consider secondary effects (support deflection increasing overall deflection of a member). BUT...I've always like the saying "The code is the worst building you can legally build." We should absolutely consider the overall deflected shape of the entire system. For a roof, it may not be a big deal. What kind of roof surface do you have? Can it accomodate that kind of deflection without damage to membranes, etc.? If so, then it may be acceptable. Otherwise, you may need to stiffen up the system near the wall to smooth out the slopes near the sides.

 

[Tomfh]

The deflection checks are rules of thumb to help satisfy the user. If you are satisfied with how it is and it’s not mis performing then there’s no need to worry about this number or that.

If the roof already has a slope in that direction then you probably don’t even notice.

 

[jayrod12]

Put a roof drain there and the potential issue of ponding goes away.

For what it's worth, I'll often keep the first two joists away from a rigid support a half joist space, but maintain the stiffness (halving the deflection for the first joist and reducing it 25%ish for the second joist) allowing a more gentle slope when deflected.

 

[HTURKAK]

I looked to the BS 5950-1:2000 for the deflection limits.. the following excerpt gives an idea ..

and clause 4.12.2 states;

( The deflection of purlins and side rails should be limited to suit the characteristics of particular cladding . )

If the tip deflection of purlin ( L/55 ) suits the characteristics of cladding and does not hinder the flow ( ponding etc..), should be acceptable..

 

[haynewp]

Depending on the slope, water ponding should also be checked which will take into account the deflection of the roof bay.

 

[ProgrammingPE]

The limits are given in terms of the member lengths since that limits the curvature along the member, which is what causes finishes to crack. Ponding checks are different which require checking the stiffness of the primary and secondary members.

Structural Engineering Software:

http://www.structuralcentral.com

Structural Engineering Videos:

http://www.youtube.com/channel/UCOj2mXXf3ZbZtgxTc6BtkGg

 

[haynewp]

No, it's not the same as member deflection checks but ponding is very much a function of deflection.

 

[JoshPlumSE]

Typically, we check a member's "relative" deflection. Right? Where we check how much a member deflects relative to it's supports. If the support deflects.

However, keep in mind that these deflection "checks" are serviceability issues, not strength issues. We could allow any level of deflection as long as the structure is capable of handling the deflection without causing other problems.

The "other problems" we're trying to avoid are typically related to other materials connected to the structure. Drywall, roofing, tiles, masonry veneer, etc. When these materials may be damaged by deflection, we will typically implement a more stringent deflection criteria to protect them.

This may be the case at the location between your wall and your first rafter or two. How you deal with this is up to you. You might be able to camber the members such that the dead load deflection is mostly eliminated.

PS I agree with the others who also mention ponding (which is a structural and stability issue), even though it is a different subject entirely from my response above.

 

[MIStructE_IRE]

“The code is the worst building you can legally build!” - I’m stealing that!!

 

[phamENG]

Steal away. The more people that realize that the better.

 

[Thomas2121]

Thank you.
So when we say relative deflection to the member itself. What if we take the entire roof as a ‘member’?
For example, if this is a 4m x 12m two way concrete slab on walls, we are not going to compare the 36mm deflection to the long span but the short one when doing the deflection check, right?

 

[Ron]

WOW! This strikes a nerve. About 7 or 8 years ago, I gave an opinion in a litigation case that the engineer who designed the light gage framing did not meet a reasonable standard of care because he did not check deflection in both the vertical and horizontal directions. The issue was that individually, the studs met the deflection criteria (L/360 because of stucco cladding) when checked vertically; however, there were double studs at the windows and studs fully attached to the columns, so the parabolic horizontal curvature did not meet the L/360 criterion.

Because of my opinion, they filed a complaint against me with my state board of engineers. Fortunately the state board agreed that I could hold such an opinion and they threw out the complaint, but it was the first time in over 35 years of practice that anyone had ever complained.

To me, that falls under fundamental engineering judgment....you have to check the worst case conditions as you know them.

 

[McSEpllc]

When a relative long spanning beam is close to a stationary wall, I considered the deflection of the beam relative to the wall. If the finishes require it, I use The span perpendicular to the beam in question to figure the allowable deflection.

 

[Settingsun]

Infinite stiffness to suit brittle cladding. Point owners to Ron's comment here to explain why they must pay through the nose for choosing brittle finishes.

 

[XR250]

I run into this in houses frequently. When floor joists change direction, the sag of the joist adjacent to the stiff, parallel girder is very noticeable - even if it meets code deflection along its length. Can wreak havoc on hardwood floors. Do I ever check this when designing houses? - uh no. However, I do try to run all the joists the same direction when practical.