Steel Truss

Another question I have is if cambering the truss for the dead weight of the concrete would be appropriate in this situation.

1) I would prefer to brace the bottom chord with a maximum L/r of 300 even if there is no compression in the bottom chord. Otherwise the truss is too limber.

2) Splicing is a good idea. It makes handling much easier. Splicing at midspan is an idea I have used in the past with a similar project. In your case, there would be two sections each 45' long. Top and bottom chords may be spliced using end plates welded to the HSS and bolted together. Bolts in the top chord splice would be nominal as the chord force is always acting in compression. The bottom chord splice will carry the full chord tension so the end plates will be quite thick. Bolts penetrate both end plates and act in tension. Large diameter A490 bolts are needed for this purpose.

3) Cambering for dead weight is normal practice but with a central splice, both sections can be built without camber, then connected to provide any slope you wish. My preference would be to provide 1/4" per foot of slope after the dead weight is in place...provides good drainage for the roof.

BA

Is your roof a gable roof or a monoslope roof?

Depending on your project and wind loads, you might want to re-consider the concrete on the roof. At that span and trib area, you can use MWFRS wind loads for your truss uplift.

If your trusses are spaced at 20' on center, are you planning on having infill framing in between? If so, are you using regular wide flange beams spanning the 20' or using bar joists that are perpendicular to the roof slope? You will need to consider the roof slope when designing beams perpendicular to the roof slope.

Below is a document I have found useful.

The Importance of Tension Chord Bracing

It doesn't mean that you absolutely must have bottom chord bracing. But, if you don't have it, you'll want to justify that decision carefully with self weight etc. In this regard, extra weight on the roof actually makes the problem worse.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

Very interesting. Does anyone have any accounts of truss failures from gravity loads due to this issue?

I, for one, do not. But, then:

- The overwhelming majority of trusses out there do have some tension chord bracing.
- When I've tinkered with the numbers, it doesn't take much to get this done with self weight and other tough to quantify "stuff".

When you look at bar joists anecdotally, you almost always see at least two lines of bridging, uplift or no. To my chagrin, there are some at my local fitness center that have only one line of bridging at midspan. That baffles me, for either uplift or tension chord buckling.

You can create an interesting example on your desk:

- set a pen up vertically on your desk.

- drape a tensed, broken rubber band over the top of the pen and hold it down at the level of your desk with the drawing end at the bottom. Like an upside down truss. Stable, if marginally so.

- Now put an eraser with a decent thickness under the pen and do the same thing. Should be armageddon unless you're picking up some unintended restraint. I consider this to be a "slightly encouraged" form tension chord buckling.

Another interesting example is just a wide flange beam. There's no such thing as a free lunch so, if this is a "thing" for trusses then I'm sure that it's also a thing for beams under the right combination of load and support conditions. I don't know of anyone actually considering that of course.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

My bet in the wide flange beam scenario, is for the most part the web is stiff enough to brace the tension flange appropriately. In a truss however, it's just a piddly little connection between the web and chord that never really had lateral loading taken into account when designed.

The opposite I think. In both cases it's the web that wants to buckle and the flange/chord doing the restraining. From time to time, I'll wind up with a wide flange beam that ends up needing to be heavily coped to match joist seat depth. In those situations, I always wonder if one ought to have bracing/bridging near the ends for this reason. I've not done it yet though. I'm not all that worried about it for a few reasons and, mostly, I'm too lazy to create a new detail.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

I concede you are correct now that I think about it more. Where the difference between beam and truss lies is the strength of the connection between web and flange. There's significantly more capacity there to prevent the armageddon type buckling in your real world small scale test scenario.

I actually disagree with that too although I'd be lying if I pretended that I had this all sorted out with great confidence. Both the truss and beam need to be restrained from what is, fundamentally, a rigid body twisting. As such, I don't feel that connection quality would come into play that much. My guess is that the difference is this with beams:

1) The ratio of torsional stiffness to flexural stiffness is much higher than with your average truss.

2) When a beam is top loaded, there's usually a relatively wide loaded surface (flange) over which the point of load delivery can shift to rectify a twisting tendency. This, so long as the member delivering the loads extends past the supporting beam center line.

I suppose that could argue that #2 is helped by a stiff-ish web/flange connection. That said, even though they're non-solid, most truss webs possess more lateral stiffness than beam webs I suspect. You're comparing HSS/WF/LL discretely versus a thin web continuously.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

You're talking about rigid body twisting. However typically the top flange is restrained from twisting significantly by whatever is framing onto/into it even steel deck has the ability to restrain this twisting at the magnitude of load we are likely discussing. Then the only thing left it the remaining web and tension flange buckling/twisting. The connection from flange to web definitely comes into play in my eyes.

Sort of tying back to BA's point, I feel that it's generally good practice to brace all truss chords generously. Fundamentally, that's part and parcel with translating these idealized 2D things into 3D realities. If aesthetic concerns mean that's not possible, then it's time to bust out the calculators and sharp pencils. And to gauge one's own tolerance for risk. And I don't mean that as a deterrent. Merely as a nod towards sage business practices. You can't be an effective structural engineer without some tolerance/understanding of risk.

Often here, folks will post about large trusses where the bottom chords are tied to the columns. The issue dujour will be whether the trusses should be modeled as pinned... blabbity blab. My supsicion is that, often, these trusses were detailed this way so that the bottom chords could span laterally between columns like girts and thus brace the bottoms of the compression webs. Surely no numbers on that, just considered detailing. You can get the same thing done, theoretically, with an axial slip connection. They didn't seem to fuss around with that froufrou stuff back in the day though. They just tied things together with a vengeance and dealt with the implications accordingly. Or turned a blind eye to the implications. Seems to go both ways.



I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

Google "Sidesway Web Buckling of Steel Beams" and some papers come up that discuss this.

jayrod said:

Then the only thing left it the remaining web and tension flange buckling/twisting

Click to expand...

Yup, I see what you're getting at. If top flange rotational restraint is given as a prerequisite to the problem then, certainly, the stiffness of the web and it's flange connections would come into play. Effectively, the restrained top flange reaches down via the web to laterally restrain the bottom flange which then restrains the bottoms of the compression webs.

This same mechanism can work in trusses too. Sometimes you'll have beams or secondary trusses tying into the girder truss webs with nominally capable moment connectins that will brace the girder bottom flange the same way. Basically AISC style rotational bracing. Of course, with big trusses, the relative efficacy of torsional restraint offered by the framing tying in, and the chords themselves, tends to be much less than is the case with beams. Things are just of a different scale and character. And that takes us back to the need for discrete bracing.



I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

Keen observation XR. Did a little of that Googling you suggested: [1].pdf. I'd do a proper hyperlink but the brackets screw it up. Clips from the intro below. At the least, this is a close cousin to my concern. I was really thinking of a case with minimal rotational support at the top flange, minimal rotational support at the supports, and a tendency towards a more true rigid body rotation as is the case with trusses I believe. Admittedly, that's a pretty contrived situation in normal structural steel.




I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

12x6HSS chords could pose a small problem of wall stress where joined by 5" or 6" HSS compression web members but this will not be known until an analysis is conducted. It may be better to reduce the width of the chords.

BA

Thanks for the replies. I’ve attached a schematic of the truss, it’s spanning over a large fire truck bay. We are still early on in the project, so I’m not married to anything yet. I’m assuming the preference architecturally will be to have the secondary framing concealed by bearing it on the top chord of the truss, I believe they are trying to go with a wood soffit. I think I’m left with a few options to pitch to the architect:

1) Run HSS members between the top chords (flush) of the trusses, maybe at quarter points, to provide axial bracing for the top chord and, ideally, rotational bracing for the entire truss

2) Use diagonal bracing (probably cables) for the bottom chord, maybe at midspan, to brace the whole truss. In this case I think I would size the top chords as a composite section unbraced for its entire length between columns.

3) If I cannot justify the first 2 options after running some calculations I will try to use a combination of the above options.

I’ll need to read through the posted literature and get educated but I think the options I described above are viable. I’ll look into the wall stresses at the web connections but if it’s an issue I think I can get around it by knifing a gusset through the top chords.

[URL unfurl="true"]https://res.cloudinary.com/engineering-com/image/upload/v1513387954/tips/Truss_Screenshots_fjnxz9.pdf[/url]

After giving it some thought, I think the cable bracing (option 2) is the best solution that satisfies both architectural and structural needs. The plan is to brace at the midspan splice location and where the bottom chord meets the end diagonals if needed while keeping all of the secondary framing concealed. Hopefully I can justify the out of plane strength aand stiffness of the chord/web connection at that location to omit diagonal bracing there but I will determine that after doing some calculations. I’ll follow up with with some connecton details but I think I have some in mind that will satisfy all required load paths and ideally brace the top and bottom chord for compression and eliminate the concrete on the roof. The dead weight of the concrete is a lot more “real” than wind uplift and I have seen deflection issues with it on long span trusses in the past due to bad construction and misfabrication so it would be preferable to avoid. Thanks for all the responses.

Kootk- Nice link on the tension chord bracing, I will be using that for sure.

At the risk of being a Debbie Downer, I don't love cables as bracing for a major truss like this. Consider:

1) Brace stiffness is just as important as brace strength. More so probably.

2) All of your cables need to be tension capable, all of the time, for this to work.

3) Cables need to be pre-tensioned to keep them tension capable after creep stretch etc. It can be tough to get/verify that robustly in the initial install.

4) If one truss deflects more than its neighbors for any reason, two of the cables coming off of that truss may loose their tension.

In my heat of hearts, I feel that cables probably would work. A little buckling movement would surely draw out any slack and get things taugh again. For me though, the scale's just off with this. I'm happy to cable brace some little curtain wall support truss etc. Not a large primary girder though. My preferences would be:

a) Justify no bracing.

b) Horizontal truss the bottom chords at the end bays and run two strut lines in between. I know... you've got an architect. All this stuff could be done in sexy timber though. And it would tend to look more "right" architecturally in my opinion. Just gotta sell it. If the strut lines could hit solid wall or columns at the ends of the truss runs, you could eliminate the trussing too.

c) Run you bottom chords out to supporting columns if they exist.



I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.