Wood Truss Connection Failure

[CBSE]

I did an investigation into a wood truss failure and am providing a fix (see attached photo for failure). It's an entire section of the roof system that failed at the same connection point...probably about (35) trusses total. The request has been made to remove the deflection in the trusses such that they are at approximately the original state before failure. The deflection cause interior partition expandable doors to be inoperable due to the deflections in the trusses.

Apparently, the doors stopped being use about 12 years ago, and someone recently went into the roof system to see what was going on.

Anyone have any guides or recommendations on an approach to remove the deflection? The roof has a Spanish clay tile roof so it's pretty heavy. Maybe it would be easier to not remove the deflection and readjust the partition doors?

Thanks,

 

[CBSE]

I'm not smart enough to figure out how to upload multiple files under one post. Attached is the picture of the failure point.

 

[Ron]

Jack the bottom cord back into place with no deflection or even a little upward beyond flat, provide a repair to the location of failure (I would use plywood gussets on both sides of the truss to restore the connection and would check the other end of the diagonal as well.). Looks like the truss plate might have been undersized at this location.

 

[CBSE]

Thanks. Plywood gusset on both sides with clinched nails is what I'm using. The plate was certainly undersized.

Any thoughts on how quickly to jack the truss back into place? I'm concerned if I go too fast it may cause 2nd and 3rd order effects to connections. I was thinking like 1/8" every week or something like that.

 

[dik]

It's a tension failure of the plate and likely underdesigned. It's a matter of shoring up the roof trusses for the full width and jacking them into place. best to do with scaffolding. Jack the trusses a bit higher than original location and then it's a matter or replacing the plate with plywood gusset. Glue and nail it in place. If there are shop drawings, they should have the force in the member... if not, contact the suppler to see if they can provide the information.

Dik

 

[CBSE]

Thanks. The supplier does not have the information, neither does the contractor that built the building (it was built in 1987). I had to analyze it in Risa and then applied some additional safety factors.

I originally thought about bracing locations at each intersection of webs/bottom chord on the bottom chord would be the best option for jacking it back up.

 

[Triangled]

yep, that's your biggest tension web force that failed... just pulled right out.

one question, although you've likely already confirmed it because you seem to be in contact with truss supplier and contractor.... question: was Spanish tile the original roofing? If yes, so be it. If no, trusses may have additional problems.

second, if you're in contact with 'supplier', assuming truss supplier, they could/should/almost-certainly-would supply you with efficient fix. .. ?

third, looks like 2x4 web and at least 2x8 chord and, like everyone says, a too small plate. It seems to be the heaviest loaded tension web in the system. is the web member adequate, ok?

forth, or fourth or.... whatever.... assuming only the panel point connector is the thing: It also appears difficult to access in your photo. I'd prefer a metal plate connector as it will be stronger in a smaller 'footprint' than a plywood plate. I've had truss yards with mobile hydraulic truss plate press machines rent their equipment to a contractor to field 'press' plates. I'd anticipate the possibility of removing the web completely as the top end is a bit shredded, and field plating both ends.

fifthish, regarding shoring/jacking.... i'd be looking at shores/jacks at bottom chord panel points as follows: 3,4,5, assuming left wall is panel point 1. special emphasis on 4 and 5.

 

[dik]

Ron: I really didn't copy you... when I replied, there was no comment...

Dik

 

[Triangled]

additionally:
you're web pulled out about an inch or so it appears from the photo, and, your door isn't working, and that makes sense.

I'm guessing you're not in snow country?

 

[Ron]

Dik....honored to have parallel thoughts with you....but it probably should scare the hell out of you!

 

[Ron]

CBSE...no real limitation on jacking speed just take it slowly and listen carefully. You'll hear it if you pop something. I would also renail all truss plates after you jack the truss.

As for jacking location, I would place the jack at the panel point just to the right of where you have deflection noted on your sketch. You can also place a jack at each panel point and do them sequentially. That will minimize your risk of overstressing any single point and will spread the upward load on the truss a bit better.

 

[oldestguy]

I'd look at ring shank nails that penetrate almost to other gusset, since nails long enough to get out the other side likely to be larger diameter and splits are more likely to occur. Gluing as an added feature might be done easily also. I've built many a truss with plywood gussets and ring shank nails and they (ring shank) work fine.With tight quarters there a nail gun WITH RING SHANKS would do great.

 

[CBSE]

Ron: Hadn't thought about re-nailing all the plates. I can see that now that you mention it. There most likely is some movement in the connectors...I didn't see any indication when I looked at the joints while I was in there, but I likely wouldn't have seen it because I wasn't thinking about it.

oldestguy: Great idea on the ring shanks. I will look into that as well.

 

[Ron]

CBSE...any rotation of the plate causes the little triangular barbs to loosen in the holes they create in the wood. That compromises their embedment, thus their capacity. Just 1/16 of an inch outward movement of a plate reduces its capacity by 20-25 percent.

 

[dik]

It may be difficult to correction: leave the plates in place and it may be necessary to remove them to get a neat contact surface for the plywood gussets.

Added: You should also review the other plated connections to see if any have failed. Plates should be in neat contact with the wood members. If you can slip a business card between the plate and the member, you should consider that joint to have failed also.

Dik

 

[struct_eeyore]

You might have a bottom chord splice plate that's become loose and contributing to deflection...

 

[Triangled]

since your chord and other web axial forces are rather small at your damaged panel point and your main issue is getting the tension forces of your web back into the truss, I might consider also widening your web with an additional 2x4 edge to edge so new web is 2x3.5" wide, giving you more surface area to accomplish panel point connection.

some eccentricity is introduced into the panel point that adds secondary moment to the chord etc, true, but since your chord and other web axial forces are rather small at this panel point..... just saying . Your main issue is getting the tension forces of your web back into the truss,

 

[Triangled]

your jacking and repair will definitely remove some of the deflection. It'b be hard to say that it'll remove sufficient deflection to make the moveable partitions fully operable. After 12 years, I could imagine that at least a little of that deflection has become permanent. I suspect the solution will be the repair plus some adjustment of partitions...at least I'd prepare the client for that possibility.