Span/Deflection Ratio 1

[XR250]

Say in a wood framed structure, you have floor joists that change direction at a beam in the middle of a room. The joists that run parallel to the beam will deflect more than the adjacent beam as the beam is alot stiffer. Say you get 1/4" of deflection in the joist from dead load and/ or occupancy load. If the first joist is say 15" away from the beam, the span/deflection ratio of the floor sheathing is going to be super low - like 15/.25 = L/60. We can call it L/120 if you want to consider it a cantilever. I see this cracking tile all the time, but is it a code violation?


Inquiring minds want to know. It is certainly a good reason to keep all joists running the same direction if possible or change directions at a natural floor finish stopping point.

 

[KootK]

- I know of no code limitations.

- I think that consideration of issues like this is great wood frame engineering.

- I've tried showing the first few joists doubled up but, as you'd expect, get contractor push back.

- This is one of the few things that makes me long for old fashioned cross bridging. Of course, squeaky floor are not.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JAE]

Not a code (life safety) issue so not in the code.

The only life safety issue would be a potential tripping hazard should the differential in slope be too severe.
We've had residential structural warranty claims come up on this issue in the past. The homeowner is pi$$*% off due to a big perceived "hump" in his floor.

Steel wide flange beam supporting perpendicular wood trusses on one side and parallel with trusses on the other with a huge difference in vertical dead load deflections.

My only answer is to really just avoid the change in span direction.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[XR250]

Not a code (life safety) issue so not in the code.

I was more concerned about the code mandated span-deflection requirements of L/360 live and L/240 total.


Thanks!

 

[bhiggins]

In my mind it is sometimes impossible to not switch joist span direction due to design constraints. Would using a thicker subfloor help soften the effects of this sudden jump in floor stiffness? I typically specify 1 1/8" T&G subfloor on all projects now. Just as an extreme example, I suppose if all projects used a 2x4 deck or something super thick, this effect would be less pronounced. I also design all floors to L/480 LL deflection, which would help also. I would try to use a combination of efforts to make this effect minimal or imperceptible.

 

[jayrod12]

I typically spec the first few joists after the switch at a tighter spacing, essentially your double joist idea, and when the contractor complains I explain the reasoning and attach the client to the email. Then let the client decide. In my experience the client typically goes for the couple extra joists since in the grand scheme the additional costs are negligible.

 

[JAE]

I don’t think it is the concept of L/240 or L/360 that matters here so much as the difference between a stiff girder or steel beam and more flexible adjacent floor members right next to it. You could meet L/369 or even L/480 and still have a problem hump in the floor.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Ron]

There has long been a large gap between strength and serviceability in our codes. Lots of attention to strength....little to serviceability. Deflection limitations are typically done for serviceability.

JAE is exactly right. Stiffness absortion!

 

[XR250]

I typically specify 1 1/8" T&G subfloor on all projects now.

I believe if i did that, the framers would force me to help unload and install it. Do you get any pushback on that?
I usually do not go past 19.2" O.C. for anything better than a shack so I don't feel that need.

 

[JAE]

XR250
We typically uses 3/4” subfloor with 19.2” oc. I think the solution for a change in span direction is to use X bridging between girder beam and first joist/truss to help minimize the differential. At least the best solution with minimal push-back.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Jerehmy]

I guess I'm in the minority, but I definitely think it's a code violation. IBC Section 1604.3 is pretty explicit.

1604.3 - Structural systems and members thereof shall be designed to have adequate stiffness to limit deflections and lateral drift. See section 12.12.1 of ASCE 7 for drift limits applicable to earthquake loading

1604.3.6 Limits - The deflection limits of section 1604.3.1 shall be used unless more restrictive limits are required by referenced standard for the element or finish material

I mean, it's plain as day to me. Floor sheathing is a structural member. Even if it wasn't, the structural floor system has a deflection issue at this location.

Say this occurs in a house with really expensive tile. Cracking in the tile occurs, so they call their insurance company. Adjuster sends his structural engineer out. The engineer is competent, takes one look at the floor system, and easily diagnoses the problem. He writes a report citing this section and showing the deflection is obviously exceeded not only by the L/600 required by the tile manufacturer, but by the L/360 required by IBC. The insurance company then subrogates against the engineer for his error. The engineer will lose as he should.

 

[XR250]

vSay this occurs in a house with really expensive tile. Cracking in the tile occurs, so they call their insurance company. Adjuster sends his structural engineer out. The engineer is competent, takes one look at the floor system, and easily diagnoses the problem. He writes a report citing this section and showing the deflection is obviously exceeded not only by the L/600 required by the tile manufacturer, but by the L/360 required by IBC. The insurance company then subrogates against the engineer for his error. The engineer will lose as he should.

Huh, I think home insurers simply deny claims. It is up to the homeowner to sue the engineer. This situation occurs enough that it is likely considered standard practice. I imagine you have designed a few like this yourself. Just about any structure is susceptible to it. Think about the condition when the joists are parallel to an exterior foundation wall - a similar phenomenon occurs and affects almost every joisted structure.

 

[JAE]

We are not talking about an excessive deflection here due to a member being designed with more deflection (>L/360) compared with some limit.

What the OP was discussing was a situation where a very stiff member was adjacent to a very flexible member and the deflection between them was noticeable.

The very stiff girder/beam could have been designed for L/1000 and the adjacent floor truss designed for L/480 or something well within the code.
And the plywood/OSB/decking between them could also have been designed at L/360 or less.

But with the proximity between them you have a severe slope in the floor because of the differential. No code provision was violated based on the OP.
But it was a system issue that the engineer should consider of course as there's a lot of things we need to consider and include in our designs that aren't "in the code".





Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Jerehmy]

XR250 - Depends on the policy I would say. I've inspected about 10 of them in the past 5 years, but they're usually not due to deflection.

It does, but usually it doesn't matter because the floor finish can handle the deflection. But, if we have a beam in the center of a room, I think we need to do what we have to to ensure the tile doesn't crack.

JAE - No, I understood what the OP meant, but let me explain what I mean a little better.

Let's say we have a wall, so it's infinitely stiff and then a joist 16" away designed for L/360. Then we have sheathing spanning between the wall and the joist.

Now we design the sheathing. We need to take into account the deflection of the joist when designing the sheathing. If we assume a rigid support where the joist is, well that's just incorrect. If you take into account the joist support is actually a spring, then the sheathing would have been designed correctly and the limits would not have been exceeded. But, we didn't and the sheathing deflection exceeds the table. So, I would say that no, the sheathing was not designed correctly since we assumed a rigid support where the joist is when it should have been a spring. We get away with it usually because the joists deflect the same amount so we can assume rigid supports.

 

[JAE]

Jerehmy,
In your wall/joist example, the sheathing doesn't sag between wall and joist.
In fact it might be absolutely planar between them, yet sloping - so there is really no deflection violation as there is no L/360 number there, only support slope.
The joist may be well within L/360. Also no deflection violation for the joist.

So what is the problem?
You correctly point out that the sheathing is not "designed correctly" with that significant SUPPORT deflection.
There would be stresses in the sheathing if the floor joist or truss was spanning, say, 22 ft. with an L/360 deflection of 3/4" and a resulting differential across 16".
Those would have to be considered in the design of the sheathing (and also tile as you indicated)

But I don't agree that the sheathing deflection "exceeds the table"...whatever that means, or any code provision directly.
The sheathing is getting bent due to support movement, but the sheathing itself, within the 16" span, would meet any L/360 type deflection provision.

As a side note, for a wall/joist condition, a proper detail is to provide solid bridging from the first several joists back into the wall - usually to tie in the wall's lateral reactions into the diaphragm. These usually go far in helping to reduce the deflection differential.

For the OP issue stated above (not your wall/joist issue) the problem is that every structural member (girder, joist, sheathing) in the floor where span direction changes does MEET the code deflection provisions; but you still have a deflection differential.

I think we all agree that this issue is a concern and one that an engineer should include detail/design provisions for.
But I don't see this as a "code violation". Your suggestion that sheathing is a structural member is correct...but there's no "L" span here to relate back to a code provision. It is an overall system that isn't defined by L/something as a limit. That's the only picky little point I'm making here.







Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Jerehmy]

JAE

Exceeds the table - I was referring to Table 1604.3 (deflection limits).

I understand your argument.

I looked to see the definition of deflection in the IBC but it's not defined. I'm just wondering about the specific definition. Do we compare the same point, once in the loaded and unloaded state; or is it drawing a line between two support points in the loaded state and measuring deflection from that? I believe you take it as the latter.

You don't think our scenario is covered in the IBC and thats fair, you're probably correct. My point is I believe it is covered. If some location in your floor exceeds the deflection limits, whether due to differential deflection or what have you, then I believe it's in violation of IBC deflection limits.

"the problem is that every structural member (girder, joist, sheathing) in the floor where span direction changes does MEET the code deflection provisions"

Everything meets their respective deflection limits yet there are deflection issues. That just doesn't sound right, does it? I think that's where the structural system verbiage comes into play.

Let's say the joist and sheathing do meet the deflection limits, I can get on board with that. I would say the structural system does not in the wall/joist/sheathing scenario.

Regardless if it is or isn't covered by the code, I'm going to be sure to watch out for it in the future.

 

[JAE]

Everything meets their respective deflection limits yet there are deflection issues. That just doesn't sound right, does it?

Just because a particular design challenge isn't purely codified, doesn't mean it's not an issue to be considered.
I'm sure we both agree on that.

I just don't see that the building code ties this "change in span direction/deflection" issue into an IBC table with an L/something limit. There is no L here.
I suppose you could argue that the deflection of the sheathing is "bad" using the value of "L" as two floor joist spaces (i.e. 32") and determine that the kink in the floor, across two joist spaces, creates an adverse L/ deflection condition.

This is simply where common sense and good engineering judgement come into play.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[KootK]

Really, I suspect that the problem is exacerbated by the goings on both sides of the beam. On the side with perpendicular framing, you're maxing out joist slope right against the beam. And on the other side, you've got this differential deflection business. All signs point to humping. While it's bad business to say no to your clients too often, we're probably knocking ourselves out trying to solve what is really an architectural problem. If the span switch can't be avoided, there probably ought to be a tile joint of some kind at the beam. Somehow things have evolved to the point where we're no longer expecting architectural detailing to intelligently accomodate natural structural behavior.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[XR250]

Here is another hump problem I see that cracks/pops tile....

Long span i-joists continuous across a 3 1/2" wide LVL beam. Typically blocking is installed between the joists over the beam to prevent roll-over. The skinny flange of the i-joist crushes and digs into the supporting beam while the blocking does not as it has no load. The blocking consequently causes a hump in the subfloor.
Make sure you specify that the blocking get ripped down 1/4"!

Just one of the many things i despise about i-joist framing.