Interesting Break in Truss 2

[SinStrucEng]

Hi folks. I've got an interesting break in a truss. One of the webs is completely split along its full length. All other truss sections are solid and in good condition. No idea how this would have even happened. Perhaps someone hit the web member during a reno, or sat on it, or something else. No signs of stress in the roof otherwise and it's mint throughout.

How would you personally go about fixing this?

I have ideas but I am concerned about constructibility. In my opinion the whole web member needs to be replaced. I propose cutting it out, carefully, without damaging the nail plates, then inserting a new piece of matching size (2x3) in its stead. Then install plywood/OSB gusset plates at the top and bottom connections over the existing nail plates as surety.

My concern is the contractor:

1) damaging the nail plates and most importantly
2) getting the new member in at full length, particularly without bending removing the existing nail plates.

Maybe cut the nail plate to be able to slot the new piece in and then gusset over it all?

 

[KootK]

@KootK, Your question confuses me slightly.

I asked, anticipating that your answer might be useful to us as forensic evidence here. And it is. If that bracing is on the longest compression web on both sides of the king post then I deem it highly likely that bracing was specified at design and not installed as some kind of post install repair.

In the pictures you can see plenty of bracing as is.

I wasn't sure if this was deliberate bracing or something else. The two rows of bracing installed adjacently like that is uncommon and pointless. Perhaps it was just a way to lap one bracing member to the next.

 

[KootK]

As XR250 knows better than most, fasteners installed below the centroid of a wood member cause problems: [link]https://www.eng-tips.com/viewthread.cfm?qid=507651[/url]. This situation has the appearance of exactly that condition when one considers the demand that the bracing would induce upon the braced member when that member tries to buckle.

 

[SinStrucEng]

I wasn't sure if this was deliberate bracing or something else. The two rows of bracing installed adjacently like that is uncommon and pointless. Perhaps it was just a way to lap one bracing member to the next.

Totally see where you were going with this now. I see this as deliberate, not something done on site during construction willy-nilly. I doubt that the bracing was doubled for the purpose of lapping, however. On each side of the roof the bracing members practically spanned the whole roof width. And yes, they were doubled on both sides as well. Odd.

 

[KootK]

I doubt that the bracing was doubled for the purpose of lapping, however. On each side of the roof the bracing members practically spanned the whole roof width. And yes, they were doubled on both sides as well. Odd.

I wonder if the truss drawings showed those slender 2x3 compression webs braced by two lines of 2X bracing at the 1/3 points. Then they got installed as shown, failed to serve their intended purpose during a snow storm, and created the predictable result that we see.

 

[3DDave]

"It would only be the part of the one web that moved laterally when it ripped apart in tension perpendicular to grain."

The part that moved laterally is the part the horizontal board is nailed to. It remains displaced.

My contention is that the horizontal boards were added after the rupture because otherwise it would also displace all the other webs it is nailed to.

 

[KootK]

The part that moved laterally is the part the horizontal board is nailed to. It remains displaced.

I disagree and see no compelling reason why it could not be the part of the web not nailed to the brace that has displaced.

My contention is that the horizontal boards were added after the rupture because otherwise it would also displace all the other webs it is nailed to.

I get it but your contention is not supported by the fact that that boards very much appear to be deliberately installed bracing that was specified on the truss drawings and installed at the time that the trusses were installed rather than as a later repair. If this was a repair, we would not expect to see the bracing installed symmetrically on the compression webs on either side of the ridge.

I will add that the bracing was symmetrical on both sides of the roof but was only applied to the first web members (closest to the king posts) in each truss.

That sounds like designed bracing to me, not a one off fix.

 

[jerseyshore]

Typically for compression members, like KootK said, I'll specify a new full length web jammed in there as tight as possible. We've all seen the older trusses that have no pressure plates for the compression members so once they take out the old broken pieces and plates, they only need one 2x4 and a simple plywood gusset and they're done.

 

[3DDave]

I am mostly struck by the appearance the lower part has cross grain fractures and looks like someone used a sledge hammer on it. Also the end of the part that is attached at the upper end doesn't seem to touch anything at the lower end.

 

[jerseyshore]

If I had to guess on the cause, this one just feels too broken to be caused by a snowy winter. One truss totally exploded, but every other one perfectly fine... We've all been in 100 attics with these roof trusses @ 24" o/c and how many times has someone seen a single truss absolutely shattered (and not obviously cut) with no issues to the others.

 

[XR250]

Such a scheme is often infeasible because it would induce a tension perpendicular to grain failure in the truss chords from the axial coming out of the web.

Welcome Back!

Don't get this. There is always tension perp. to the grain in a truss. The full length sister attachment is no different that that of a truss plate.

 

[RontheRedneck]

Off the top of my head I'd venture to guess that truss is around 36' span. Looks to me like a forklift hit is when a bundle of trusses was being loaded on a truck. A very common occurrence.

I would just sister it on both sides, and extend the scabs to the chords. That saves cutting through plates or peeling them off.

The web is in compression 99.99% of the time. Forces are not too high. No need to over-think it.

 

[IRstuff]

Seems like an overabundance of shingle nails, though, particularly next to the truss in question; one might expect reasonably aligned and orderly rows, but these seem a bit haphazard.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[KootK]

Don't get this. There is always tension perp. to the grain in a truss.

No, there isn't. At the least, not in the joints that do the primary "trussing". In fact, detailing truss joints to avoid that exact outcome is a fundamental aspect of designing wood trusses regardless of the scale or the connection typology.

As shown below, the idea is to pass the truss vertical shear from one web to the next through the joint alone, not allowing it to "live" in the chord beyond for any appreciable amount of time. That, precisely because it causes the tension perpendicular to grain issue which is a nightmare for wood members. You'll see this concept manifested in heavy steel plated truss joints, dowel pinned truss joints... you name it.

It's no different in a metal plate connected wood truss where you want the load path for the vertical truss shear to be:

1) Compression web.

2) Connector plate.

3) Tension web.

 

[KootK]

I would just sister it on both sides, and extend the scabs to the chords. That saves cutting through plates or peeling them off.

The full length sister attachment is no different that that of a truss plate.

That path likely pans out as shown below which creates the tension perpendicular to grain problem shown below.

 

[KootK]

Conceptually, the issue is similar to trying to splice two, vertical tension members across a transverse member, as shown below. Obviously, no one here is gong to be okay with that regardless of whether or not some numerical voodoo can be managed.

 

[KootK]

An extreme example of this concept is those rebuilt truss things. Obviously, those wouldn't have a prayer if the vertical truss shear were not being passed from web to web via the steel pins, without engaging the chords in tension perpendicular to grain. That also renders them quite difficult to reinforce / shore at times.

 

[Tomfh]

It looks like the diagonal brace perpendicular to the truss is pushing the bottom chord laterally. This has caused the truss web to fracture, rather than deflect laterally, due to the restraint provided by the nailed cross members.