horizontal deflection of scissor joist spported on masonry wall

[Philboyd]

I am designing a gymnasium with load bearing masonry shear walls. the architect wants to use scissor joists with a 2-in-12 pitch for both the top and bottom chords. The joist span is 100 ft and the masonry bearng walls are approxomately 30 ft high to joist bearing. The joist manufacturer has informed me that the joist will deflect a total of 1-1/8 inches in the horizontal direction. I cannot allow a slip connection as the masonry walls need to be attached to the roof diaphragm for support.

I am concerned about the deflection of the joists cracking the masonry. I have considered allowing "slip" to occur at one end until the dead load is in place and then making a positive connection. I am still concerned that possible live load deflection added to possible uplift deflection in the opposite direction may cause cracking in the masonry over time, particularly near the corners where the masonry is very rigid.

Should I be concerned or can the masonry accommodate this movement without cracking? If the live load deflection can be limited to 3/4 inch that would only be 3/8 inch at each wall.

 

[jike]

I would think the walls would be able to flex to accomodate that movement except at the corners. You may need to set the first joist far enough away from the corner to be confident that the wall can move without cracking.

 

[BAretired]

I assume you mean 4/12 and 2/12 in top and bottom chord respectively?

A horizontal deflection of 1/2" in a 30' high masonry wall should not cause cracking in the central portion of wall, but if the trusses are spaced at two foot centers, you could have a problem near the end walls.

I think it would be a good idea to provide horizontally slotted connections at both ends of the first three or four trusses from each end wall. These connections must be able to resist uplift, of course. The bond beam at the top of the wall would then span, say 8' to the first tied down truss.

Best regards,

BA

 

[msquared48]

BEretired:

Intreresting thought here. What problem{s} have you run into in the past if this modification was not incorporated?

Mike McCann
MMC Engineering

 

[hokie66]

Mike,
He can speak for himself, but I doubt BA has had a problem, he is just trying to avoid one.

I agree with jike and BA that the roof members near cross walls should be allowed to move. And I would allow them all to move until dead loads are in place.

 

[BAretired]

Mike,

Hokie is quite right. I have never had a problem because I cannot recall ever having designed a scissor truss spanning 100 feet between masonry walls. The detail which I am advocating is one which I believe would help address the concern expressed by Philboyd.



Best regards,

BA

 

[msquared48]

OK...

I guess the problem I was dealing with here is how do you transfer lateral forces perpendicular to the trusses into the long bearing/shear walls if movement of the truss ends is allowed. In that the concensus is to allow for only dead load movement, fixing the live loads, I can accept that. Thanks.

Mike McCann
MMC Engineering

 

[BAretired]

mike,

There are three degrees of freedom at each end of each truss. I am proposing that two of those will be restrained. Only one will be allowed to move, namely the direction normal to the wall and only for three or four trusses at each end. Lateral forces perpendicular to the trusses can still be transferred to the wall.

If you know of a better way of achieving the desired results, please let me know. It is certainly open for discussion.



Best regards,

BA

 

[DaveAtkins]

First of all, we need to remember the horizontal movement at the "roller" end of the truss will really be divided between BOTH end supports. So the top of each exterior CMU wall will only move 9/16".

I think if you put a vertical soft joint (that can accommodate the 9/16" movement) in the CMU near each corner of the building, you don't need any special truss bearing details.

DaveAtkins

 

[Philboyd]

All,

Thank you for your responses, they are appreciated.

Just to clarify, the joists will be 60 inches deep and have parallel top and bottom chords,(both 2/12). The joists will be spaced at 10 to 12 feet and the thrust force causing the deflection will be approximately 82 kips horizontal, at each joist.
I have 2 choices for what I can do at each end wall.

1. Place a joist near the end wall to pick up the deck

or

2. Span the deck to an angle or channel attached at the end wall.

In either case, I need to provide for attachment at the deck to transfer the diaphragm end reactions into the shear wall. I will, of course also need to provide wall anchorage for uplift bridging on the botoom chords. Under normal circumstances I like to put an extra joist next to the end wall. Should I provide the extra joist and let the end slip? or should I skip it and go with option 2, spanning the deck to a wall mounted member.


How do think either of these options will affect shear transfer?

 

[jike]

I recommend a deck bearing angle in all cases. It transfers the in plane and out of plane shear directly.

If the end wall follows the slope of the roof and the deck laps onto it or continues up one course, you will have problems trying to match the top of the wall with the top of the joist because of camber or deflection. You may not have as many problems if the wall extends up (like a parapet).

 

[BAretired]

Philboyd,

If the joists are spaced at ten or twelve feet, I would not think you would need any special detail at the bearing points. The walls should be able to move 1/2" in ten feet without causing excessive cracking. Maybe using DaveAtkins' soft joint in the masonry is still a good idea.

You will not have a hinged and a roller end to any of the joists. Both ends will be roller ends and will each accommodate half the deflection. Lateral resistance will come from diaphragm action of the wood deck.

You could use a joist at the end wall, but you would have to connect it intermittently to the end wall to transfer the diaphragm shear and to laterally support the wall itself. In doing this, you would effectively prevent the joist from deflecting so that thrusts would not develop at the bearing points.

You could use a wood ledger or a steel angle or channel at the end walls too. That would likely be the more economical solution.



Best regards,

BA

 

[SteveGregory]

I would suggest using a flat-bottom chord truss with a pitched top chord. If the owner needs the extra 8' of headroom in the center, you could make the bearing walls taller by 4' or more.

Plead with them if necessary.

 

[Philboyd]

Steve,

Thanks for the advice, I am inclined to argree to with you, but the designer in question has done the same type of design in the recent past and since they haven't experienced any cracking yet, they don't see any reason to change the design.

I'm wondering if cracking that is due to this type of situation is something that will show up immediatley or if it is more likely to occur over time. I know of one building with this type of joist and wall design that has been occupied for two years with no apparent cracking and two others that are nearing completion of construction. Have you had any experience with this type of design or other types of masonry cracking that makes you reluctant to let the wall deflect?

 

[JStructsteel]

82K reaction, I would think you would have trouble with your connections onto the wall. How thick is the masonry?

 

[COEngineeer]

I would think adding butresses / shearwalls going the other way would solve the problem. Perhaps add horzontal steel on the top to span from butress to butress?

Never, but never question engineer's judgement

 

[BAretired]

I think the 82k is the horizontal reaction required to reduce the horizontal deflection to zero. It will not happen, because the walls will simply flex as required.

Best regards,

BA

 

[Philboyd]

jrisebo,

12" CMU. 1900 psi, Type S mortar, f'm 1500 psi, #7 @ 48 vertical reinforcing, 4" brick veneer, 1'-6" total wall thickness. There is no way that I know of, besides buttresses, that the wall can be designed to resist the forces, so I must allow for the deflection, because if the roof ever "sees" it's design load the deflection is going to happen.

 

[SteelPE]

Why don’t you model the whole system as it would end up in the field? Model the truss with on end pinned and the other on a spring. You should be able to calculate the stiffness of the wall… because it is only a cantilever beam. Just make sure you multiply stiffness by 2 to account for each bearing wall. Then you will be able to get an accurate representation of how much force goes into the wall.

 

[JLNJ]

Maybe a good share of the deflection can be taken care of by letting the connx slide until all the dead loads are applied. Now you are down to 3/8" or less each side. With 10' to the first joist it seems to me you'll likely be OK.