Sistering Existing Floor Joists

[Lion06]

I have a small (but large scale) residential project where I need to sister a lot of existing floor joists. There are two flavors of these joists - the first pockets into a CMU wall at each end, the second pockets into a CMU wall at one end and frames to a stair header at the other.

We will be jacking these joists to relieve the existing DL stresses. How would you typically determine how far to jack them? I've considered doing a DL only deflection check, neglecting creep and using that value (rounded down to the nearest quarter inch).

We will be stopping the sisters short at the ends where these joists pocket. The biggest reason for this is that there is existing HVAC ducts under the joists at one end that we don't want to have to drop, and there are a lot of wires at the other end that we don't want to have to cut, pull, and re-wire. So the fix looks like an existing joist with a sister that stops roughly 2'-3' from each end.

There are several of these joists that we're not going to count on for any strength in bending at center span (this is mostly because of large splits or shakes at midspan - we had a grader come out and grade in-place all of the existing lumber so we have a very good idea of what is ok and what isn't), but the sister is still only coming to 3' from the ends. This clearly requires a connection at the ends of the sister capable of transferring the moment at that location back into the existing joist to take into the wall. What I'm struggling with is is this required where the existing joist has bending capacity but needs to be sistered for strength? In my mind it wants to be just to maintain consistent rotations bending moment, but how important is that really? The moment transfer is critical at the joists with no bending capacity at midspan to remain stable. In the case where the existing joist does have bending capacity at midspan is there anything wrong with just designing the connections (which will be nailed) at the ends of the sisters to dump a shear only (with no moment) back into the existing joist?

Any thoughts?

 

[paddingtongreen]

Does this work?
Figure how much shear at the center if the sisters were separate beams, bolt the sisters in place at the middle, the beam is bent but the sister is straight. Jack the ends up till the sister is back in alignment with the beam at the ends, make the connections each worth half the shear in the middle.

It has the advantage in that you don't have to guestimate how much deflection to take out of the beam. I used it on a steel structure that was under load and was going to get more. With luck, if the sister is not as deep as the beam, you can fix one end without having to jack it up and then just jack the far end.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[Ron]

SEIT...have you considered flitching these with steel rather than sistering. That way you could take it all the way to the supports and not have to worry about the moment continuity at the 3-foot point.

 

[Lion06]

paddington-
One problem is that the joists likely were installed with some type of crown, so the joist could potentially be perfectly straight right now, but is obviously under load.

Ron-
I did consider the steel option, but to attach it to the joist back to the face of the CMU still requires dropping the HVAC at one end (which we're trying to avoid). Also, the steel is more expensive, right? Probably not an issue on a single unit, but this is a pilot unit for what will be somewhere on the order of 200 units by the time they're done. What are your thoughts on that? The structural fixes are going to be very inexpensive compared to the enabling work of moving wires/ducts/pipes, so we're trying to limit that as much as possible.

 

[BAretired]

Prior to joining this forum, I had never heard of the term 'sistering joists'. Does anyone ever brother, mother, father or cousin joists?

BA

 

[TXStructural]

I assume you are talking about plate-connected wood joists.

Anything you do which is discontinuous at the ends will result in odd inflection points in the joists, and unless the manufacturer really messed up the original joists, the plates at ends will not have excess strength. You cannot safely just sister part way.

As for lifting out the deflection, you can take out the load-induced deflection, but not the long term creep and the deformity at the connections.
Jack using shoring posts at each panel point. There is typically not enough capacity in the plates to take the stress reversal when jacking at a single panel point at midspan.
The stiffer truss or member will take proportionately more load, so try to design your repair to either take the load or match the existing joists.

There are other methods to reinforce under-strength joists. Web members are typically short enough and lightly-loaded enough that they are not the limiting members, it is the chord and connections. I used field replacement of plates on about 200 joists on one project. Using plates, you can add a second bottom chord. We also used sistered LVL lumber to take the load off the joists in many instances.

 

[Lion06]

TX-
They are 2x12 dimension lumber. I should have stated that in the OP.

 

[TXStructural]

Ah, good thing I put that in my response, else we would have been talking past each other.

I think you will create a serious discontinuity at the ends of the new board unless the connection is made along the length. The hard part will be getting the connection at ends without killing the capacity. If you are considering this solution, you have obviously determined that the existing wood has adequate shear capacity.

You might consider using manufactured lumber (LVL, etc) which might allow a shallower member to be used where needed. I began using these almost everywhere since dimension lumber quality is so unpredictable here, and the cost increase for the strength gain proves economical.

Sandwiching the existing, one end on one side and the other end on the other side, might allow the boards to be pushed to the ends (if the load is sufficiently heavy to require it.)

Last time I did a repair like this, they ended up removing most of the utilities, attaching ledgers or making room in pockets, and carrying the boards the entire length. The old structure was under-designed by a significant measure (accounting for the collapsed roof we were fixing.)

 

[msquared48]

SEIT: If you're worried about developing the moment at the end of the sister, rather than rely on side fasteners such as nails or bolts, try fabricating a light gage side mount, top flange hanger in a bucket design that would be able to be placed at the end of each sister - right and left.

Nail and glue between the sister and joist, clinching nails on the far side.

You might even sell the idea to Bart Simpson.

BA: This forum is wearing on you.

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

 

[BAretired]

In the case where the existing joist does have bending capacity at midspan is there anything wrong with just designing the connections (which will be nailed) at the ends of the sisters to dump a shear only (with no moment) back into the existing joist?

Even if the existing joist doesn't have bending capacity at midspan, it is not necessary to develop moment at the end of the sister joist provided the existing joist has adequate capacity at the end connections of the new joist.

If the sister joist is connected in only three places, i.e. at each end and at midspan, it will help. Say P is the load at midspan and P/2 is the sister joist reaction at each end. It has a moment of PLs/4 where Ls is the length of the sister joist.

The existing joist must have adequate shear capacity in the end region and must have a bending capacity of R*a where R is the reaction and a is the distance from support to center of end connection of sister joist.


BA

 

[hokie66]

I would be reluctant to do much jacking on the existing joists. You could do more harm than good. How much reliable reaction would you have at the ends to jack against? Would you be able to jack all the joists concurrently? I would tend to just think about ultimate bending capacity and not worry about built in stress.

 

[msquared48]

If you could place the sisters with an upward camber, forget the jacking.

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

 

[Lion06]

BA- if the existing joist has no capacity at midspan then I believe that the end connection of the sister transferring moment to the existing joist is critical and here is why. For all intents and purposes the end is essentially a link. If the middle of the existing joist were cut away and it were two individual joists it would be easier to see. Even though that isn't the physical condition, that's why I'm assuming in my analysis.

Hokie-
I did think that jacking could potentially damage finishes on the floor above, but I'm struggling with how to justify the joists sharing anything more than LL without jacking. Wood doesn't have the same ductility of steel so they really need to reach capacity at the same time.

Mike-
crown or no crown they'll still share load the same way, right?

 

[rowingengineer]

I would do the same as hokie, timber is sufficiently ductile for my liking. Not to mention creep, shrinkage and all that jazz.

An expert is a man who has made all the mistakes which can be made in a very narrow field

 

[dhengr]

BA:
In the US, in our efforts at political correctness in all things and in our effort to involve more of the fairer gender in the engineering and construction fields, we have quit brothering and fathering joists. I did once hear a carpenter say ‘that mother something joist,’ when the joist fell on him. Finally, joists are kinda like cousins of beams so they could probably be side by side.

SEIT:
I’m not really sure what you are trying to do: are you adding load, so you need more member; or are these particular, individual joist defective and showing it in softness, deflections or growing cracking, etc? Do they all happen at the same structural location in 200 different units. Or are they truly bad pieces of lumber just randomly located throughout the job. What are the span lengths and spacing of the jsts., and the loadings? Is the ceiling under the jsts. still open? Do all of these same defects occur at mid span or is that variable too?

Is the horiz. shear stress and compression perpendicular to the grain (bearing area), at the reactions, of the existing joist O.K. without any additional work or reinforcement? Then you still have to transfer the portion of the shear from the new, shorter, reinforcement member (sistered jst.), back into the existing member, at the reinforcement termination points. Think split rings and bolts.

You must jack up the existing jst. to partially (mostly) unload it, without lifting it off its reaction points or else your sistered jst. will not really come into play, and reinforce, the existing jst. unless more load is added and more deflection occurs. And, you probably don’t want that. I would look at calculated DL & LL deflections and probably jack up for more than just the DL delta. Pay attention to joint movement and imminent cracking above when doing this. But, in doing this, I would want to account for some connection movement as the existing jst. moves back down and imparts the loads into the new jst. Existing jst. set with crown up still will have approx. the calced. delta; and creep or long term loading you really don’t know how to account for, but don’t ignore them either.

In its simplest form, why not think this way: 1) jack the existing jst. up DL delta plus; 2) apply the sister jst., 3' shorter on each end, with a bolt near each end and a bolt at center span; 3) as you release the jack you impart a downward load on the sister jst. at center span and this is reacted by and upward bolt load at each end of the sister jst.; 4) of course these bolt loads are in the opposite directions on the original jst., but you should be able to draw shear and moment diagrams for each of the jsts. and superimpose them for the total member. You are really dealing with concentrated load transfers (bearing and shear stresses) at the end of the sister jst., not a moment transfer (a moment connection) as such. If you cut the original jst. at mid span, then you would need two bolts there, one on each side of your cut, to start to transfer the moment.

 

[hokie66]

It is not only ductility that I would rely on in this decision, but rather the higher safety factors used for wood than for other materials. I suppose it depends on whether you believe in ultimate design for wood or not.

 

[BAretired]

dhengr,

Thanks for the short course on joist genealogy.

I agree that we are dealing with concentrated load transfers between existing and new joists. If a section of the original joist is removed, the sister joist must be connected each side of the two cuts and at each end.

For any particular situation, the forces can be resolved by statics.

BA

 

[hokie66]

The load from the floor sheathing will continue to be transferred into the existing joist, so continuous fastening between the joists is desirable to uniformly load the new joist. I don't know why you would want to apply the load at centre span. I agree that the end connection should be for the full end reaction of the new joist's portion of the load.

 

[msquared48]

SEIT:

Right, but if the upward camber of the sister is greater than the current "deflection" seen by the current joist, when you force the sister tight to the top of the existing joist, it not only sharea the load, but brings has the effect of lifting the existing floor structure. If you can tolerate the weak analogy, it is a primitive form of post-tensioning

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

 

[dhengr]

I wasn’t necessarily suggesting the three point load transfer as the best solution, but rather trying to draw the simplest picture, statics and shear & moment diagrams to make my point about the force transfer at the termination of the sistered jst. I agree that a more distributed load transfer would be better along the whole length of the jsts., but I think you will still have a concentrated loading at the termination of the sistered jst. to deal with. And, this distributed connection arrangement is certainly not as determinate as the simple example BA and I used. I would not press the sistered jst. up to the floor sheathing without jacking the existing jst. up first for fear of starting to push the sheathing up and away from the existing jst. Then, certainly put the camber up on the sistered jst. and fasten it to the existing jst.