Creep in 1920 wood truss

[tdse]

Without any forensic structural engineering experience I've been asked to look into an HVAC remodel on a roof with problems. The building is a 90 year old theater with wood trusses spanning 50' between brick piers, 13'6" on center, about 4' tall, Howe style (I think, and top chord bearing from what I can see). T/B chords are 4-2x12 through bolted, angled webs are 4x4 and 4x6, vertical webs are steel rods through T/B chords with steel plates/nuts. The owner tied a string to monitor movement several years ago, and there is siginificant deflection at midspan (4 3/8" = L/137) under dead load (no snow when observed, design sl is 30psf). He's noted 11 7/16" movement since 2003, 1/8" in the last two years. It last went under a remodel in 1997 when the current HVAC was installed. There's no previous information on the project and the "structural plans" from the remodel are by an architectural firm. My question is, how much creep is too much. The plaster is cracking along each truss at or near midspan. There is also cracking at or around each brick bearing pier. I observed three trusses and all look Ok with one exception where there is splitting along a top chord through bolt connection.

There are several "collateral" loads which could not have been originally designed (mechancical ducts, lights, stage lighting, sprinklers). In addition I can only assume the "attic" is accessed for the recessed stage lighting possibly more often than any original intent as well. Lastly, before the current owner bought the building there was water damage, and the tar roof was replaced with membrane.

I plan on taking measurements and trying to find the capacity of the existing trusses. Does anyone know a good refence for calculating a decrease in capacity of the wood due to the creep? thx

 

[OHIOMatt]

Can you clarify something? You mention that the owner noted 11 7/16" of movement since 2003. Is this accurate?

 

[tdse]

Typo, 11/16" movement total, 1/8" in two years

 

[msquared48]

If I read it correctly, that's L/50, way too much. With plaster, the total load deflection should be no more than L/240, and the live load (snow) deflection L/360. That's 2.5" max under TOTAL load. The dead load deflection should be no more than 1".

"there is splitting along a top chord through bolt connection." BIG red flag here!!!!!!!!!!! That, coupled with the roof leaking and wetting/drying the wood under load could be the main problem. There may be slippage in the connection. IT NEEDS TO BE CHECKED ASAP!

When was the last time that the old roofing was removed from the structure prior to re-roofing?



Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[msquared48]

Well, I guess things change... I was typing as the update cam in...

Just check the connection you mentioned real close.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[tdse]

The re-roof was around 2003 and the owner is trying to find out what (if any) of the existing was removed prior to the new membrane install. From the attic/underside the original planks are visible.

 

[BAretired]

Sounds like it's deflecting too much under dead load. I would strongly consider strengthening all of the existing trusses and devising a repair detail for the split top chord.

BA

 

[hokie66]

This type truss deflects primarily due to loosening of the joints. As the chord members shrink across the grain, the rod tension members loosen, and of course nobody retightens them. This causes deflection and slippage of the fitted compression webs. The strength of the timbers is not likely to have changed appreciably unless the water damage was substantial...it is all in the joints. So creep is probably not the best description of what has happened. To confirm whether or not the timber member strength is an issue, you should probably have them assessed by a grading expert.

Where there has been splitting of a chord, that can probably be rectified by adding bolts or screws to pull the members back together.

The L/d of the trusses is 12.5, which is quite comfortable. Is the roof essentially flat or sloped? If flat, ponding could be an issue with the deflection which has occurred over the years.

 

[csd72]

I agree with hokie, this level of movement is most likely due to the connections loosening up.

Creep does not usually weaken the timber, although the opposite could be true i.e. weakened timber would tend to creep/deflect more.

Nailing plates may solve the problem of the splitting, I would be careful about adding any more bolts unless this is rectified.

 

[hokie66]

csd72,
Yes, nailing plates may be the answer for the splitting. I was thinking he meant the 4 chord members were separating at the hanger, but you are probably right, there is splitting on the wide face.

Another issue is to check the joints in the chords. The 2x12's wouldn't be in 50' lengths. If only one is discontinuous at a given location, this may not be a concern.

 

[tdse]

OK so connections, I did see some gaps (not large, maybe 1/4") between the the bottom of the top chord and the angled 4x webs which I assume to be s result of the rotation due to deflection. I can clearly see the rods and threaded portion with nut above the top chord (at least in some areas), bottom is mostly inaccesible. Can that be tightened on a truss this large?

Also on the L/d comment, the 4' height was not measured but only estimated ( and poorly, apologies). Am heading out to take measurements, inspect all the connections I can get to.

The chords 4-2x12 are not continuous, however the breaks are staggered. Added pic of the split at connection...

Each truss steps with the pitch of roof, probably about 0.5/12 pitch or larger, we already discussed ponding.

 

[a2mfk]

Not that it has much of an effect, but moisture in wood will exacerbate creep. Just did some research for a job and found a couple articles on it. Wood is one odd building material when you really get into it.

 

[hokie66]

Yes, the rods should be able to be tightened, provided there is sufficient thread. Tedious but necessary if the trusses are to be saved. A 1/4" gap between the fitted compression webs and the chord makes for large deflections when geometrically applied over the length.

The joint shown needs to be reinforced, and I would suggest if other joints are fitted in that manner, all the joints near the ends of the trusses should be side plated. The cut in the chord is so shallow that longitudinal shear has split the chord, at least the outside member which is shown. The web is cracked also.

These trusses have no capacity to resist wind uplift. If you must follow current codes, does the dead load of the system provide sufficient ballast against uplift?

 

[Sawbux]

My experience with old wood trusses is that most are overstressed in the bottom chord. Older designs used higher allowable stresses for tension. Research in the 80's (?),( on full size members) has greatly reduced the allowable tension values for wood.

You also indicated that the bottom chord is made up 4- 2x12 laminated, the net section of the bottom chord is only 3-2x12 in that case. The tension in the member is then governed by how much load can be transferred by the fastenings between the laminations. The deflection that is measured is probably due more to slip in the inter-lam connections than creep in the wood.

The deflections in these trusses can be removed by tightening the vertical rods- if the nuts will turn.

I have reinforced many old trusses using cables, which I tighten to reduce a portion of the dead load stresses. I limit the preload to about 1/2 to 3/4 of the calculated dead load to avoid putting the bottom chord in compression, which can make it buckle. I size the cables to take the full D plus L load for safety. The wood chord will still carry the live load because the EA of the wood is much greater than the cables will be.

The split you have is caused either by horizontal shear due to the thrust of the diagonal member ( I suspect it is at one of the outer panel points?), or cross grain tension caused by rotation of the joint. It is best repaired by bolted clamp plates and drawing the split closed.

I recommend you check for slip in the lower chord. Look for gaps between the ends of the discontinuous 2x12's, any fresh looking wood - indicating movement, deterioration or slip in the fasteners, etc. It sounds to me like they might be a good candidate for the cable fix.

S. Wagner, SE

 

[racookpe1978]

I would be very suspicious of "nailing" the truss joints back together.

The wood is already noted as splitting, and the joints are noted to be separated by as much as 1/4 inch.

Instead, consider adding 3/8 (or 1/2) plates on both sides of each truss joint. Then drill through the joint (through the 1st plate, the wood, the then second plate) and bolt the joint back together by requiring at least 2x bolts in each compression member of wood going into the truss joint, and at least 3x in each tension member.

Painted flat black, the new plates will be visible but won't be too ugly - rather striking in their own way - and the repair will allow you to pull members back into position. For example, put a jack under the sagging pieces, and tension up the joint plates when the sag is removed.

Tension cables can be run between nearby plates if a tie point is provided.

 

[JAE]

I had a church project a few years ago with a similar truss - sagging at midspan and cracked plaster ceilings. There was concern over the ultimate capacity of the trusses and really no easy way to evaluate the safety.

We ended up adding W24 steel beams OVER the roof, directly above the trusses and dropping rods down from the beam to cross tube brackets - essentially picking up the truss at its panel points and taking the overall load into the above-roof steel beams which spanned across the main room and landed on the existing load-bearing brick walls.

An awkward, sledge-hammer fix but it gave us the assurance of safety and avoided massive interior work which would have cost a lot.

Here is the cross bracket detail/photo:

 

[JAE]

Here is the above-roof view. The beam was boxed in and roofing was constructed over it:

 

[tdse]

Thanks all for yout input, will keep you posted.

JAE the owner and I discussed a similar fix at the original inspection, using a steel truss conficuration over each existing. If it comes to that it's good to know it's feasible.

 

[tdse]

I reinspected the trusses that are accessible (about half of total) Found three connections in bottom chord where the plies were bolted together (staggered 18" o.c.) and bolt was too close to the end of the ply and thus failed. See pic with the gap in the ply. There's likely more at the inner plies. Several bolts appear to be bending (there's prying to one side of the washer/bolt). The bottom and top chords are both sitting in pockets in the brick wall each end, but there's a gap between brick and the top chords so all the support is now at the bottom? Bottom bearing is hard to see, but there's steel bolted plates and connection has varied gaps (can see bolts).
Sawbux, do you use the cables over the entire length of truss? The inacceblilty of the bottom chord of the trusses may make this impossible without demoing plaster (we have stacked/stepped ceiling joists on top of bottom chord, not to mention mechanical)
What torque would be required to tighten the nut on the vertical rods (provided there is sufficient thread)?
If ponding is not a concern, could I reinforce the truss in it's current deflected condition without returning it to it's original profile (assuming no additional deflection will exceed the connections once reinforced)
JAE, when you did the reinforcing from above did you attempt to "pull-up" the truss to overcome the deflections?
Will post additional pics

 

[tdse]

Here's a pic of the gap in the top chord bearing at brick. The outer ply has slipped to be offset-vertically from the others.