Truss Reactions

[JStructsteel]

So got truss reactions as a lump sum, i.e. not broken into DL, LL, WL, etc.

Is it worth figuring out the components, or just design for the total reaction using ASD?

I asked for the component reactions, but got no reply.

Thanks

 

[XR250]

I just use ASD.

 

[phamENG]

Yeah, I don't think the MiTek and other similar softwares are capable of breaking it out. And you're not working with an engineer - it's a designer who may be very knowledgeable (@RontheRedneck ) or may have no clue what load combinations are. The engineer is just rubber stamping the output after the fact. They probably couldn't tell you where a single truss they've "designed" is going to be installed.

Just apply it with ASD and roll with it. If you have complex framing with lots of transfers, then it might benefit you to work out loading for yourself so you can break it down.

 

[Eng16080]

I'd probably break it down into the components. This will also serve as a good double check of their loads/reactions.

 

[RontheRedneck]

No one has ever asked me for reactions to be broken out before. The program does not give us any control over what appears on the drawings.

It might be possible to get them by printing out calcs for each load case. I've never tried it. If I think of it tomorrow I'll give it a shot.

 

[EngDM]

No one has ever asked me for reactions to be broken out before. The program does not give us any control over what appears on the drawings.

It might be possible to get them by printing out calcs for each load case. I've never tried it. If I think of it tomorrow I'll give it a shot.

In Canada where we use LSD we get them broken up by type all the time. If/when you guys ever switch to LRFD only, you'll likely see more of this.

 

[RontheRedneck]

Just for the heck of it I ran a 40' truss today. This is what the program came up with for a shop drawing.





There's an option to print calcs to a PDF file. You should find it attached if you care to take a look. But it's almost 40 pages of stuff based on all the load cases. And for the most part reactions are not spelled out.
.

 

[SE2607]

Maybe I am oversimplifying the problem, or maybe I'm just dumb, but can't one just take the tributary load to one truss and use statics to get the reactions by load case?

 

[EngDM]

Just for the heck of it I ran a 40' truss today. This is what the program came up with for a shop drawing.


View attachment 7684


There's an option to print calcs to a PDF file. You should find it attached if you care to take a look. But it's almost 40 pages of stuff based on all the load cases. And for the most part reactions are not spelled out.
.

Can't you just take the VERT(TL)-VERT(LL) to find VERT(DL) anyways with this? It basically is broken up into components.

 

[phamENG]

And for the most part reactions are not spelled out.

Yeah - looks like you have a load case for "IBC" which I'm guessing is the worst case gravity combination for the overall truss, about 14 different wind cases, a dead load only case, and a wind and dead load case.

but can't one just take the tributary load to one truss and use statics to get the reactions by load case?

Yes and no. You can do that and get close enough to design a safe and reliable structure, but I've never managed to get it perfect - especially when dealing with more than 2 bearing points. It's just a bit irritating that they can run these trusses in such a powerful program, but there's no option to give the raw data needed to produce an efficient design for the rest of the building.

 

[XR250]

Yea, I mean look at this truss. The right end (14) is supported by a girder truss. I asked them to provide a bearing at 15 (interior wall) to take some load off the girder but it is barely taking any load (131 lbs). If their software took into account the actual deflection of this truss and the carrying girder truss (14), the reaction at 15 would be a lot more.

 

[SE2607]

Yes and no. You can do that and get close enough to design a safe and reliable structure, but I've never managed to get it perfect - especially when dealing with more than 2 bearing points. It's just a bit irritating that they can run these trusses in such a powerful program, but there's no option to give the raw data needed to produce an efficient design for the rest of the building.

Wow! After 48 years, I'm finding out statics is no longer a valid analysis technique! I'm not sure what your definition of perfect is, but one should be able to get reactions of either a statically determinant or statically indeterminant beam by statics.

 

[phamENG]

So you know what the stiffness of the truss is without knowing the manufacturer's profile and material choices?

 

[ChorasDen]

Throwing my hat in here, I always just ratio the given reaction. Is it a perfect solution? No, but it's close enough to satisfy me.

If a reaction is give as 1387lbs, and the design loading overtop is 15psf TC DL, 5psf BC DL, 20 RL, 35 SL, I'll just assume a worst-case load combo (DL + SL) and us the total roof loading ratioed down to break down the reaction into component loads.

 

[bones206]

I asked them to provide a bearing at 15 (interior wall) to take some load off the girder but it is barely taking any load (131 lbs).

Actually looks like that 131# is a tension reaction

 

[EngDM]

Yea, I mean look at this truss. The right end (14) is supported by a girder truss. I asked them to provide a bearing at 15 (interior wall) to take some load off the girder but it is barely taking any load (131 lbs). If their software took into account the actual deflection of this truss and the carrying girder truss (14), the reaction at 15 would be a lot more.

They don't take into account deflection/stiffness or 2nd order effects?

 

[KootK]

They don't take into account deflection/stiffness or 2nd order effects?

Not sure what second order effects you are thinking of but I doubt they are accounted for.

Also no to the situation that XR250 mentioned regarding the relative stiffness of the wall and the girder. And, frankly, I feel that's forgivable. Analytically, that's a pretty big ask and I doubt that many EOR's give similar problems that level of attention. Also, how far is the truss supplier to take it? What, then, if the wall is on beam that flexes? Or the girder is on a post on a foundation that settles? On and on... At some point you gotta just fish or cut bait.

I would actually be in support of a code requirement stating that light frame wood trusses not be allowed to avail themselves of supports that would make them statically indeterminate. That's another conversation though.

 

[XR250]

I agree with ya Koot. I have also seen enough creep in attic trusses like these over the years to know that wall is going to take more than 131 lbs.

 

[ANE91]

Here's a slick one from Paragon. I recall being told that the reactions per load effect are available in the software. Haven't worked up the nerve to buy it, yet...

 

[RontheRedneck]

Yea, I mean look at this truss. The right end (14) is supported by a girder truss. I asked them to provide a bearing at 15 (interior wall) to take some load off the girder but it is barely taking any load (131 lbs). If their software took into account the actual deflection of this truss and the carrying girder truss (14), the reaction at 15 would be a lot more.

Since trusses are considered as individual components, there's no way to make one truss take deflection of another truss into account.