Compression in Wood Diaphragms Induced by Basement Walls 2

[AZengineer]

When designing the floor framing and diaphragm for a typical residential basement (restrained design with walls laterally supported at slab and at floor), how are nonbearing walls adequately braced? I would think blocking would need to be installed between each framing member far enough to develop the load into the floor diaphragm, but once it is developed, how do you insure that the floor sheathing will not buckle under the compression load (i.e. in a square basement where the floor acts as a compression member bracing the tops of the walls)?

Similarly, how would you insure that the slab on grade is sufficient to brace the reaction from the bottom of the retaining walls without buckling? The lateral loads induced into slabs and floors in these types of walls can be quite large. On my specific project, I have 1000 plf into the floor, and 2000 plf into the slab on grade (10 ft walls with 60 pcf at-rest pressure.)

 

[JAE]

AZengineer,

For the floor strength and stiffness you can utilize the horizontal diaphragm tables in the IBC to verify that the floor can "span" horizontally from wall to wall when taking on lateral force perpendicular to the floor joists.

Blocking between joists parallel to the basement wall should be extended into the diaphragm far enough to develop the load into the sheathing. If you space the blocking every 4 to 6 feet, usually you can get the load developed in 4 to 6 joist spaces...depends on the sheathing and the depth of basement wall.

For the slab, the concrete floor, if flat, can take quite a bit of load before it buckles upward. You might try to develop a buckling capacity of a 1 ft. strip of floor slab via traditional buckling equations but stop and think about it for a moment. 2000 plf / 4" slab / 12" width = only 42 psi. This is a very small stress. If the floor is flat enough, buckling shouldn't be a problem.

 

[AZengineer]

If the basement is square (or rectangular), then theoretically the diaphragm won't have any shear in it since all of the walls have equal and opposite reactions. Thus, the diaphragm would not need to be designed to "span" horizontally as it normally would if it were were taking wind or seismic loads. I agree that IBC Table 2306.3.1 would be utilized to determine the length of the blocking required to transfer the load from the top of the basement wall to the diaphragm, but I am concerned about the compression in the sheathing once that load has been developed. In this case the diaphragm behaves more like a compression element than a conventional flexible diaphragm, which theoretically only transfers shear.

 

[JAE]

AZengineer....good point that its theoretically in compression, but if you have a perpendicular to joist compression in a diaphragm, there theoretically isn't much stiffness there relative to stiffness in the "shear" direction.

But if you really want to check it - you could go to the APA specification for the design of plywood sheathing and determine a compressive allowable stress in the plywood for a 16" unbraced length (assuming joists spaced at 16") and then get a number to compare.

But not too many people bother with this as the stresses are low. Again...1000 plf / 12" / 3/4" sheathing = 111 psi - pretty damn low - even for wood.

The key is to ensure that the wall has something to bear against with relatively significant stiffness so the wall doesn't lean in. With the between-joist blocking you offer the wall a fairly stiff load path.

 

[AZengineer]

True that the normal bearing stress is relatively low, but because the plywood is built up of several plies (some of which are perpendicular to the load), I would think that the plywood would have a relatively low radius of gyration and thus a fairly low capacity to take compression loads in the plane of the sheathing.

I'm not extremely concerned with the problem on the current project at hand, but at what point does this become a ligitimate design issue? It doesn't make sense to me to assume if there is "something" stiff for the wall to push against, then you can assume that the plywood will sufficiently handle the exact same load that you are reinforcing it for at the nonbearing wall location.

 

[RARWOOD]

AZengineer

I think you are looking at the problem wrong. In the direction parallel to the floor framing the framing acts as compression members. In the direction perpendicular the walls are braced by the floor diaphragm made up of the floor sheathing and the diaphragm chords.

I think if you try to consider the sheathing as a compression element I think you are in uncharted territory and your question about the panels buckling is a legitimate question.

The published tables for the use of APA sheathings as diaphragms are not based solely on the use of engineering mechanics. Instead they contain a lot of data that is based on load tests of different systems.

Your problem is similar to the transfer of seismic forces from tilt up concrete panels into wood diaphragms. AS JAE indicated most of the times your loads are going to be small and will not present a major concern.

However there may be times when you have much larger loads and need to be a littl more concerned. In those cases you may need to use subdiaphragms or drag struts to distribute your load in to the floor framing.

JAE's advice is excellent. In particular, his advice about talking to some one at APA-EWS

 

[PMR06]

We have used the same approach as discussed, i.e. drag strut the load into the diaphragm either with perpendicular joists or blocking. In our experience though, both in design of new and investigation of failed existing foundations, the problem has not been the capacity of either floor diaphragm, it has been the load path from the top of the concrete wall into the wood floor diaphragm that controls.

With the standard detail of 8” concrete wall, 2x6 sill with anchor bolts, 2x joists positive bearing on the sill (with a rim joist), and floor sheathing:

Using 1000#/ft top reaction, that's 1333# per joist or blocking (assuming 16" oc). Even with anchor bolts between every joist, a 2x6 sill is not capable of 1333# shear and there's no way toe-nailed connections will transfer that force into the joist ends. I have seen the tops of walls pushed inward cracking the sill in half, and I’ve seen both the sill and wall push in together. In both cases, neither the joists nor the floor diaphragm budge a bit!

We have tried using a detail in which a continuous pocket is cast into the top of the wall that the joists and blocking bear in. With this configuration, the wall is pushing directly into the end grain of the joists, then drag strutted into the diaphragm. It works great on paper, until the contractor sees it!!

We always been curious how other engineers have approached the problem…

 

[JAE]

the problem has not been the capacity of either floor diaphragm, it has been the load path from the top of the concrete wall into the wood floor diaphragm that controls

PMR06 - my point is the same..I agree.

 

[AZengineer]

I agree, PMR06... I usually detail the walls extending up to the finished floor elevation rather than relying on anchor bolt/strap connections. The basement I'm designing right now is constructed of 12" CMU, so I'm bringing it all the way up to finished floor. The client wants to use a 2x4 stud wall for insulation, so I am using it as a bearing wall for the floor joists.

RARSWC, I understand the concept of the floor acting as a diaphragm, but before it does that, the load needs to be adequately transfered. I am looking at this analogous to a steel beam for example. If the loads are heavy enough, web stiffeners need to be provided to avoid web crippling failures. Similarly, I would think that the diaphragm would need to be checked for this same sort of "web crippling". Dragging the load into it via blocking insures that the load is transfered, but it does not insure that the sheathing is rigid enough.

Am I looking at this incorrectly? It seems logical to me, but I know thousands of basements have been built without analyzing this...

 

[RARWOOD]

AZengineer

You are correct the first problem is adequately transforming the load into the floor sheathing. You also make a good point about the square basement with equal loads applied on all four sides.

I will chat with some APA-EWS people I know about the problem. I believe part of it is semantics. I think we arre both talking about a force applied to the floor sheathing. The force is in the plane of sheathing.

When one has a drag strut, one looks at it as dragging shear out of the sheathing into the strut. Your problem is the opposite, you want to apply a force and distribute it into the sheathing.

Floor sheathing depending on the sheathing type, thickness and nail spacing is capable of resisting a load in its plane of X number of lbs/ft. So I believe that if your sheathing has an allowable load of 400 lbs per foot you can put that much "compression" load into it. If you try putting 700 lbs/ft into the sheathing then the sheathing may buckle or fail through nail shear. To prevent buckling or "web crippling" instead of adding stiffners you increase the panel thickness and or decrease the nail spacing.

Returning to your question about transferring the load from the wall to the floor framing usually, when possible, the best way of transfering load to wood members is by bearing. So your solution of bring the cmu wall up to finish floor is a very good approach.

Yesterday I was trying to think of some good ways to transfer higher loads out of the cmu wall into the floor framing, when the framing and rimjoist bear on a treated plate sitting on top of the wall.

PMRO6 is correct that at times the amount of force that can be transferred into the floor framing is controlled by the plate capacity. I've seen applications where joist hangers are laid flat on top of the plate. Load is transfered through nail shear into the joist hanger, then by bearing of the joist seat against the end of the floor joists. However as you loads increas this won't work.

I deal a lot with glulam hip systems supported by cmu walls or piers. In a lot of cases a masonry hanger is embedded into the top of the wall, such as a Simpson GLB beam seat. I have been trying to think of some ideas of doing something similair with floor joists.

The problem I run into is that the floor framing is sitting on tip of the 2x plate. So anything one would use has to be anchored to the wall by going through the plate firs.

 

[jhoulette]

I'm not sure where most of you are designing at but in Colorado we use buttresses and counterforts to brace the basement foundation walls. Instead of spanning top to bottom rotate your span side to side.
I agree with the problem is the "load path" from the top of foundation into the floor sheathing. It's not a easy path and it gets costly to have a framer do it right.
The counterforts we use are typically 3' to 4' long (out from the foundation wall) and uses the weight of the soil above (footing) or a drilled pier to resist the uplift. Or use a buttress that works in compression on the inside of the basement. But usually builders hate to have a 3' foundation wall jogging into the basement. But compression is always better to work with.
Only an idea, but we have had great success with it.