Long span low sloped trusses 6

[RontheRedneck]

Someone brought me a truss design to look over the other day. One of our salesmen sent it in to the office for review.

I know that some of you probably know this stuff better than I do. But others may not deal with trusses a lot. There are new engineers every year. And not everyone who reads the thread is an engineer. So if you feel like I'm talking down to you or telling you things you already know, it's not intended to offend anyone.

Here's the basic truss:

I know you can't see a lot of detail. The span is 77'. It's 3' 6" tall at the low end and about 5' tall at the high end. Spaced 2' OC with 30-10-10 loading.

A few things jump out at me right away. Looks like it meets the deflection criteria:

L/240 for live load and L/180 for total load are allowable in this area. So what's the actual deflection?

The live load and dead load together total well over 4".

So what kind of ceiling does the building have? The salesman doesn't know.

Are there interior walls? Don't know.

Is an architect or engineer involved? No, this is apparently someone winging it on their own. (Not at all uncommon in this area)

With L/180 allowed for total load, the truss easily slid by the design program. Total deflection could have had up to a little more than 5".


What else comes to mind?

Wind loading for end verticals is something you have to turn on for every truss. It got missed on this design.


So does my train of thought have a caboose?

There are 2 things I want to emphasize. I think there ought to be max deflection limits that are numbers rather than L/ numbers on long span trusses. L/180 is fine for a 28' residential truss. But in a situation like this it's not.

If any of you are involved in similar projects I'd suggest specifying max deflection as a number.


This truss should be deeper. The guy probably thought a shallower truss would be cheaper. But that's not the case in situations like this.

Another very serious issue I have with this truss is erection bracing. Trusses like these have a tremendous amount of force in the chords just with their own weight. With the TC in compression it will want to buckle awfully badly. Without a ridiculous amount of bracing it will be difficult to keep them upright until the plywood is applied.


I did a job with trusses similar to the one above a few years back. When I sent in the sealed drawings I sent a letter along with them stating my concern about erection bracing. They assured me they knew what they were doing.

Here's the result:

https://newschannel20.com/news/loca...f-building-collapses-near-carle-at-the-fields

 

[BulbTheBuilder]

Great thread RontheRedneck! I am a young/new engineer myself. For long trusses I don't use the L/span ratios. Example, that a 100' long truss, L/180=1200/180=6.6667" which is excessive. Things I look at when it comes to deflections are ceiling, piping, and other deflection criteria of elements supporting by the truss. For roof, I am particularly more concerned about ponding since that can be catastrophic. Appendix L of AISC 360-16 talks about that (I think you are a wood person, can't help with timber references)

I can remember off head but there are two design guides by AISC that have important information on the span depth ratio for long trusses. For roof I use 1:12, floor 1:10, bridges 1:5 to 1:10. Another important thing I have learnt so far is the importance of bracing and connections. They are key to the stability of the truss.

 

[driftLimiter]

Ron, thanks for sharing this.

I have a few questions.

1. What in your view is a proper depth for a truss of this span and loading? Do we have a rule of thumb on the span to depth ratio of trusses that gets us started on the right path?

2. Do the truss manufacturers have the ability to camber?

My thought is that if the dead load deflection was cambered out, and there isn't serious level of snow loading, the the live load deflection on this long of a truss would probably never be realized. Certainly if there are utilities or ceilings that have a deflection limitation then we need to adhere to that, but if its an open ceiling perhaps then maybe it doesn't matter as much.

 

[1503-44]

When a truss deflects, what it really wants to do is roll over at the first hint of misalignment.
Camber does not eliminate deflection, it only makes it look nicer when you can't see that it has a serious amount of deflection and it is a lot less scarey to look at it. Brace the hell out of it. IMHO 1:10 rule of thumb.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[JLNJ]

For a steel joist, the standard camber would be 2.5". This makes me feel better about the possibility of ponding issues. No camber and 5" of defection on a 1/4" per foot roof? You better have a slick membrane with few rooftop obstructions. And no roof sag or creep over time.

 

[jayrod12]

I've always ran with 1:12 for economical wood trusses, so for a 77 foot span, you'd be looking at a minimum of 7'-6" truss depth. Using the same slopes as above, you get an increase in stiffness for the truss equal to 460% using the identical lumber. Now all of a sudden that 6" of total deflection is 1.5". Far better.

Obviously we usually get beat down on truss depths, but if I ask for 1:12, and accept 1:16, in this case that would equal a stiffness increase of 184% using the same lumber and the truss is 4'-9" deep at the low end. But if I start with going "It's most economical at 7'-6" deep, however we could push the trusses to 5'-0" deep at the cost of some higher grade lumber." Then I've found the truss suppliers can make tolerable deflections work without too much pain and the architect thinks I'm doing them a solid because I saved 30" of truss depth.

 

[jittles]

I echo jayrod12's comments

Start the negotiation at an inch depth per foot span and usually land somewhere in the 75-90% of that depth range.

Span ratios only go so far and absolute deflection accommodation must be considered. You want to stick in the bounds of typical products - fastcap screws (or whatever the Simpson title is now), CFS deflection tracks, etc.

Just working on a similar project, ~50' span, 42" deep trusses to match some existing truss depths. I'm unsure of the economics, but the existing trusses have DF select structural grade for the chords, so it's already pushing the envelope IMO.

(40 psf snow loading)

 

[RontheRedneck]

driftLimiter said:

"What in your view is a proper depth for a truss of this span and loading?"

I don't have a specific rule of thumb that I use. If I were actually designing the job I'm mess around with the depth and see how the deflection looked. Maybe talk to the owner or GC and see what their thoughts were. See if there's a height limitation for the building or other restriction we're dealing with.


"Do the truss manufacturers have the ability to camber?"

Yes. But I have not seen it done for ~35 years.

When you try to camber a truss you get into issues with webs being the wrong length. Joints don't fit as tight as they should. In theory, it should be easy to do. In reality - Not so much.


BTW - I didn't mention that I wasn't the one making decisions on this quote. Someone just asked me what I thought of the truss design. I told them no way in hell would I allow it if it were up to me.

Maybe next week I can find out how things turned out.

 

[1503-44]

Good on you.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[BAretired]

If any of you are involved in similar projects I'd suggest specifying max deflection as a number.

I agree that L/240 live and L/180 total are inadequate for long span low slope wood trusses. L/360 seems to be standard live load deflection in North America.

Trusses cambered for dead load will deflect, but camber compensates for deflection, so it doesn't show without live load. Trusses without camber deflect under total load and the full deflection may be seen by an observer.

For this reason, I suggest limiting the total load deflection to L/360 for low slope trusses without camber.

 

[RontheRedneck]

BAretired said: "L/360 seems to be standard live load deflection in North America."

For floors - Yes. For roofs - No.

At least in my experience, with the codes I've dealt with.

Obviously different numbers are sometimes called out on plans if an engineer is involved.

 

[BAretired]

@ RontheRedneck,

I am suggesting that Dead plus Live load deflection be limited to allowable live load deflection when trusses are not cambered. If that means L/240, then L/240 is my suggestion for total load deflection for long, low slope roof trusses.

 

[XR250]

My experience with long, shallow floor trusses designed near the bottom of the code is that they creep over time due to gang plate nail slip. I imagine this will suffer the same fate. As such, I would be looking to design to double or triple code.

 

[BAretired]

The OP mentions a creep increase factor of 2.00 for dead load.

I'm not sure how that factor is intended to be applied, but maybe I should revise my recommendation to the following:

(2*DL + LL) deflection shall not exceed L/240 for low slope roof trusses without camber.

 

[RontheRedneck]

Yes, there is definitely a creep factor in truss design.

I have never seen that addressed in the specs on a job.

 

[BAretired]

I have never seen that addressed in the specs on a job.

Neither have I, but maybe it's time we did. Long term deflection is addressed in concrete codes for the design of concrete members, and it makes a considerable difference.

 

[Aesur]

I'm going to play devils advocate here with the caveat that I would have been deeper in my design for said long span truss. If this truss has 4.32" of total deflection, but is sloped 1/4:12 we would have approximately 19" drop from high to low side. This theoretically would still allow for the truss without ponding issues as the low side is still lower than the deflected member.

As for what we do: we specify Live load (floor and roof) deflection as L/360 for simple span and 2L/360 for cantilevers and require camber of 1.5x design dead load deflection which is the local industry standard, we use more stringent criteria when required for floors.

 

[BAretired]

@Aesur,
In this thread, we are talking about low slope trusses, but flat trusses are not only low slope, they have no slope at all, so ponding is a consideration. Flat trusses are sometimes specified by architects for some pretty long spans. Some slope may be created by varying insulation thickness, but it always seemed to me that sloping trusses was a more sensible and economical solution. From the standpoint of drainage, it seems a better idea to put roof drains near the supports rather than at midspan, so building a crown on long span trusses is my preference.

Like you, I have used L/360 for roof live load deflection, not L/240, but you can be sure that code minimums are going to be used by some. Lack of camber and creep are somewhat unique issues for timber trusses and should be given serious thought by engineers.

 

[lexpatrie]

As a side note, ponding for wood is to my knowledge only discussed in the old ASCE 16-95 LRFD spec for wood. Please let me know if it's been updated anywhere since.

https://ascelibrary.org/doi/book/10.1061/9780784400418

Building code is going to require special inspection or something like that as the span is equal to or greater than 60 feet. If it's deflecting this much, I'm not convinced it's the appropriate depth to span ratio for economical wood construction. If it were an open web steel joist, K series wouldn't even work, they top out at 60 feet long. The first on that works for 70' is a 26LH07. That's a lot of steel joist without even considering the loads.

Regards,
Brian