Existing Trusses Failing 2

[HeavyCivil]

I was asked by the architects at my firm to measure and analyze some existing trusses for proposed solar panels.

The trusses are 36' clear span , 6pitch roof and every member is a 2x4.

I analyzed them for 40psf snow load (1992 pg - todays is 60psf for the area but existing building code, I think, lets me use 40psf). I found top and bottom chords failing (bottom chord F.O.S. is nearly .5!). Say nothing about additional loads - this roof, as I have analyzed it, is not even strong enough to support its original design loads.

The question is how to get trusses "up to code". By my analysis a 2x8 bottom chord and 2x6 top chords will work with additional bracing (existing bracing is minimal). But that is in theory. Do any of you have experience with "building-up" existing trusses with deeper members? Will simply nailing on needed members (I'd assume it'd have to be a very aggressive nailing pattern) be enough to assume those members are working compositely? - Seems like this could be a dangerous assumption.

There is also some ethical baggage too. I believe the statute of limitations has expired for the owner to take any action against the fabricator or contractor (or even determine who is responsible). Still I'm not sure where the engineers responsibility extends to beyond making any truss affected by additional panel loads up to code.

The caveat is that this is the first truss analysis I've done professionally and it was done with Risa 3-d, a program I am familiar with but am not an expert. I am pretty sure, however, that they are not and have never been up to code.

 

[miecz]

Taking a quick look at your model, I notice that nodes 8 and 13 don't land on the top chord. Is that right?

 

[jike]

First, I must a few questions:

Do you have a copy of the truss shop drawings? This will give locations of required bracing as well as the grade and specie of wood used for various members. Often higher grades of wood are used (like MSR, etc.)

Make sure the bracing and member sizes and species are properly modeled.

It is advisible to go to the job site to observe the actual bracing installed (many times this is not installed or incorrectly installed) and note grades and spicies of various wood members and measure truss plates sizes.

 

[miecz]

Same for nodes 6 and 12.

 

[miecz]

Next, I would expect the dead and solar loads to be applied in the vertical direction.

 

[HeavyCivil]

Miecz- No, that is not right. All members definitely connect at gusset plates. I will try to fix. Your also right about loads, although I think your discovery about the nodes is more related to the results as the snow load really has the biggest impact.

Jike- I don't have a copy of show drawings. I did go to into the attic and measures all the trusses (members and node layout). I modeled the members as No.2 , and from what you have said I should (fix the model per miecz's comment) and reiterate with a higher grade lumber to see what the difference is.

 

[Splitrings]

VTEIT,
I am helping a buddy build a garage in VT. It happens to be 36' wide with a 6:12 pitch. His truss looks identical to the one you show. He lives where the GSL is 60 psf. I will grab his design info that came with his trusses and let you know. From memory, I think the TC & BC where SYP 2400 Fb. The chords are 2x4's in his truss, also.

 

[HeavyCivil]

Splitrings- Thanks- that would be great.

Miecz- As I said, I'm pretty new to Risa. I modeled the top chords as continuous members because my understanding is that modeling separate members into each node will complicate bracing (it will assume pin, braced connections at joints I think). I'm not sure how to make an exact intersection on a continuous member, but I think that the way I modeled it Risa thinks they intersect anyway. Maybe I'm wrong though.

See attached photos of members. I'm not really familiar with lumber grades outisde No2,No1, SS and others listed in NDS. I would expect a very special grade to be stamped or labeled, which I did not see.

When I change the material to Select Structural, I still have Euler buckling code check for top chords, but bottom chord is now okay. I guess I really need to track down the lumber grade.

 

[HeavyCivil]

It just occured to me that SYP is probably southern yellow pine and not a different grade of visually graded lumber.

 

[Splitrings]

VTEIT,
A couple of things I noticed with your model. You aren't using a Cr factor, which you can. (Asssuming a 24" truss spacing, of course). You also have your top chord braced at 5' o.c. in the weak direction. I'd be willing to bet the original truss design required bracing at 2' o.c.

 

[HeavyCivil]

Repetitive member factor...true...Okay will do.

I have the top of the top chord braced at 6"oc for roof sheathing nailing, but you're right, bottom is modeled as 5' o.c.. If you look at the attached picture there really isn't ANY weak-axis bracing for the top chord on the bottom "flange". If anything, 5'oc isn't conservative enough.

You're probably right that 2'oc bracing was required but it was not installed.

 

[Splitrings]

VTEIT,
My mistake. I didn't see the decimal in your model for bracing the top of the TC. I saw the 5 and just jumped to 5'. I agree with your bracing assumptions. I typically use continuous bracing on the top when sheathing is applied but based on the nails I see that missed the top chord in your pictures..maybe they are roofing nails? I usually will not show bracing the bottom flange of the TC.

I think the piece we are missing is the species and grade. From the pictures the members look like SPF. I have several truss design here that use SPF 1650F 1.5E. If I change you members to this species and grade, it is adequate for your snow loads. Again, I'll bring my buddies design info with me tomorrow.

 

[HeavyCivil]

Okay- I updated the model to show all members as No.2's. Top chords and bot chords are SYP and webs are SPF. I had them at No.3.

What I do not understand is the allowable stresses (fb') risa uses. It lists f'b for SYP as 1984 psi. Even with a Cr and load duration for snow (say 1.15 also), I cant figure out how fb' gets so high. This is why I had them at No.3 before.

If I brace top chord at 2' o.c. I get a Bending check of over 2.0 (they are not braced at 2' o.c. but this is easy enough to do, easier than building up chords.).

If I build top chords up to (2)2x4, and brace them adequately, I get no code checks, even with my panel loads.

Thanks for your help everyone. Splitrings- that would be helpful to see your friend's design- thanks.

Attached is updated model of existing truss

 

[HeavyCivil]

Attached is model of modified truss and panel loads.

Note that LC4 controls.

 

[Splitrings]

Fb'=Fb*Cd*Cr
Fb'=(1500psi)*(1.15)*(1.15)
Fb' = 1983.75psi

Fb in NDS '05 is listed as 1500 psi for SYP #2 in table 4B

 

[BAretired]

The second photo (SDC1790.JPG) shows no bracing on the compression diagonals each side of the ridge, yet they appear to be pretty long.

I would run hand calculations on this truss as a check against RISA.

BA

 

[HeavyCivil]

Okay - I'm working out of 2001 NDS so that explains it. I should update our library and suggest a purchase.

BA: You're right, the truss has almost 0 bracing. In the modified model (and in the existing just to get reasonable results) I fudged some in. My design will include bracing on the webs you mentioned as well as on the top chord and everywhere else that has le/d>50 or approaches buckling stresses.

I did change members back to SPF because I cant confirm that it is SYP, which lowers fb' but it still works.

Thanks again for the help - super resource in lieu of in-house mentor.

 

[HeavyCivil]

n/m - 01 still lists fb = 1500psi for SYP my fault.

 

[MiketheEngineer]

And BTW - if the orignal truss was designed by a truss plate mfg - you will probably never be able to duplicate their work. They have proprietary software that makes extremely good use of all factors available.

Good luck and I am sure that the truss is at least near its maximum with a 40 psf snow load. And mayvbe someone jus found some old trusses and put them up - I have seen it done more than once.

Classic case: A farmer bought some 24' trusses designed for residential use. He hung his chicken coops from them. The chickens grew up nice and fat until one day one got just a bit too fat. KFC had a big sale the next week.

 

[steve1]

I recently had some issues on an existing steel roof truss that I was checking for new loads. The top and bottom chord members were continuous WT sections. I used RISA 3D for the analysis and found that the trusses (1969 era) were failing under the original design loads. Upon further investigation I determined that the chords were failing in a combination of axial load plus bending. Now all the load was applied at panel points so I was wondering where the bending forces came from. It turned out that I modeled the chords as continuous members and not as discrete members that were pin connected at each panel point. In my conversations with the folks at RISA we came to the conclusion that the truss chords were attracting shear forces due to their continuity, and that since this was the reality of the situation we could not discount those forces.

What I'm getting at here is that if you assume that the members are all pin connected but you use a continuous chord then your design no longer matches the assumption.

Perhaps your chords are also failing in a combination of axial force plus bending.