Deflection

[smokiibear]

A couple questions on deflection:
It's been my opinion that the deflection limits set by IBC 2006 (and other codes) are generally lower than most people would find acceptable. As such, for residential construction where spans generally are not too much longer than 20 feet or so, I've generally tried to hold total deflection to less than 1/4", and in other cases, just limited dead load deflection to less than 1/4".

I've often told contractors and architects than smaller beams or rafters would work, but how much deflection would their clients be willing to tolerate? .5" per 20ft? 1"? 2"

So, my question is, how much deflection should be considered tolerable for roof and for floor members?

Secondly, if a floor memeber ends up directly support roof members (due to load path), are deflection limits for floor or for roof?

Looking forward to your input.

 

[apsix]

You are using deflection limits of up to L/960, while the norm is L/250 - L/350. That's a lot of extra section.
Unless dynamics is a problem I will use as recommended by code.

In answer to part 2, I would expect the limits for floor members to be more stringent than for roof members and therefore the floor defection criteria is used.

 

[ishvaaag]

Traditional stances for steel here (Spain) were 1/500 for floor not supporting structural walls and 1/250 for roofs, both under total service level loads. Lateral deflection was not initially a concern; buildings were RC framed or masonry mainly, and industrial buildings most times tension braced frames. When first appeared, some controls were used to discern building structures as non sway or sway, and industrial buildings had their day with portal frames of unsightly lateral deflections allowed in practice due to unclear specification in codes.

By now a RC building is allowed to sustain a 1/250 total actual deflection under the applied total service level load with a proper calculation of the deflection, including stiffness degradation by mechanical cracking and creep (obviously checked generally in approximate ways) AS LONG what is called the active deflection doesn't surpass 1 cm or 1/400 of the span. This active deflection is that occurring after the floor gets the loads applied, say weight of the floors, partitions, mechanical services and live loads, of course as well properly accounting cracking and creep. The rationale must be all deflection except that active will be hidden under non-structural parts and a remaining 1/400 movement is unnoticeable at sight, reasonable in stiffness for confort respect vibration, and reasonably conservative respect breaking partitions.

Of course if one plans to add structural walls above some framed structure must be more conservative. The traditional assumption for steel was to keep total deflection under total service level loads under 1/1000. For this case, from the deflection viewpoint, the stiffer, the better.

Finally most modern codes reckon even if in hidden way that deformation controls are as much a matter of distortion as of deflection. Rules are more complex, in the case of the steel having been here significantly relaxed, for where before strength almost never controlled design, being the deflection stiffness controls those that governed the selection of the member, now are more in agreement and a selection made directly in terms of strength can be valid also from the deflection viewpoint. This surely has been made to allow for agreement with other less stringent european codes and made by the summarily way of reducing both the limit deflections (between the 1/500 and 1/300 range, there is a range of hypotheses to check) and the part of service level Live loads (50%, 30% whereas before 100%) being brought to the check. This of course means steel vibration control, ponding issues will be more of concern from now on, before almost never were, and even the stiff steel members being used were NOT tasty to feet accustomed to the rigid concrete floors, so how less now.

CTE, the current, code, also very properly extends the requirements of deformation be brought to foundation matters, so proper estimates of total and differential settlement (dishing action etc) are required. These deformations are directly compounded with the deflections under loads and may break partitions where a mere deflection check wouldn't show it be the case. I have seen this to happen, and can say even where: buildings with H plan shape under general dishing action founded with floating piles in a deep mantle of clays, and U buildings (likely also in clays or weak soils) in the arms, likely also dishing action. These midrise cases showed structurally non-meaningful diagonal cracking at some partitions and could be repaired to tolerable level after primary settlement. In clayey soils if a crack remains open midterm on, likely you will have the crack open and closing seasonally, with cycles of overall rains and temperature.

So you have an idea of what to make in almost every situation. Keeping 1/500 or 1/400 under total service level load will get you outside of most typical problematic situations, but then be sure to have accounted in adequate way settlements. To keep up stiff walls, loaded or not, better go up, masonries crack even before 1/1000 shows.

A proper control of lateral deflection in the same ranges for masonries, and to the required level in structural P-Delta terms plus requirements of cladding (mainly allowed movement of glass panels in their casings) complete the most of the significant issues -other than vibration- relative to deformation controls.

These things surely should be complemented with distortion issues here, but for now leave here to not make it too long.

 

[Ron]

Deflection criteria are put in place for a variety of reasons, including aesthetics, limiting cracks, mitigation of creep, and to force sections to be more resistant to "dynamic" deflection such as floor "sponginess".

It is better to stay with a rate of deflection rather than a set value as you've done, as this allows the designer to make deflections compatible throughout the structure, thus maintaining some semblance of strain compatibility as well.

For instance if you limit deflection to 1/4" and you have a 10-foot span, it will have a deflection rate that's twice as high as limiting a 20-foot span to 1/4" deflection.

Stick with the code requirements. They are defensible and in most cases reasonable.

 

[BAretired]

I agree with Ron. The deflection limit should be based on a fraction of the span rather than some arbitrary limit. But the entire question is a matter of opinion as to what is a satisfactory deflection...not really an engineering matter.

BA

 

[apsix]

Asking the client probably won't help.
Ask them if they want 1/2" deflection for a 20' beam and they'll probably say no.
Give them 3/4" deflection and they won't even notice it.

 

[GregLocock]

Isn't there a standard based on the resonant frequency of the floor, so that it doesn't amplify footfall?

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[StructuralEd]

We can't get 40 ksi steel very readily around here.
Are you sure that's what was used?

 

[ishvaaag]

GregLock

in current (Spain's) code there's such section

CTE DB SE-A
7.2.2.1 Percepción humana that pertains to vibration control.

Prior to this code starting vigency in 2005 there was only that I remember reference to vibration for bridge and footbridge construction; the controls then for deflection were more exacting for steel.

Other sections in CTE (SE, that cares of required hypotheses to check, SE-A, when specific treatment of geometrical nonlinearity is mandated, and SE-C, that cares about foundations and alike, then total and differential settlement) are also use to ensure proper service level behaviour from a deformation standpoint.

 

[Ron]

Greg,
Good point. Codes in the US require consideration of vibration and dynamic loading, but do not provide specific criteria (except for seismic activity).

It's a relatively common issue for steel joist supported concrete floor slabs.

Ron

 

[smokiibear]

What about garage headers with stucco siding and a second story above? What about ridge beams of a roof? What about floor beams with gyp on the underside in the middle of a large open ceiling living room? How much deflection will really go unnoticed? What about large porches where beams are supporting heavy tile roofs? Locally, there are many a houses where deflection is quite visible, even though service loads are within code defelction limits.

I'm curious if some of responses are taking residential design into account, or large commercial buildings where this code deflection would never really be seen. Even 3/4" isn't a lot of deflection, especially if half of that is due to live load. But I've had plenty of residential beam calc out with 1" to 2" of deflection and meet requirements. I'd believe that would be plenty visible.

I suspect that there is a deflection for a given length that would be unoticeable, and my inclination is that it is usually much higher than L/360. Although I've heard of plenty of contractors call for help when their garage headers sag a bit on other engineers' jobs, I've never had the opportunity to see for myself the effects of design to the code minimums on deflection. I've thought even a bit of overdesign in these situations is better than a call back. But, it's been a concern to me for over five years as to what others are doing out there, and how much overdesign is really going into my design.

Another factor that enters into this is the derivation of dead load, especailly for floors. I've noticed that in most situations, engineers use a floor dead load of around 10psf. Over and over again I've gotten specs for a tile floor to come in around 27 psf, and for wood floor to come in around 16psf. Yet I see over and over engineers design for 10psf and usually that seems adequate. The only reason I suppose this is occuring is that the live loads are never being realized on those floors. But I've never thought it appropriate, certainly not defendable in court, to reduce the dead load because of possibly excessive live load issued by the code.

Anyhow, looking forward to more of your responses.

Thanks.

 

[Ron]

SB...your questions show a lot of naivete in your practice.

Why does there have to be a difference between commercial and residential construction? I know there are different codes, but if you read those carefully, there's not much difference. A residential structure is not spared the loadings that a commercial structure sees, particularly with respect to wind loads. The wind doesn't discriminate nor is it selective.

Yes, the structure might not ever realize the live loads for which it is designed, but can you predict that? No. I live and practice in a high wind area. Most of the time we have nothing near those design loads. But occasionally, those wind loads get realized and even more.

When you can predict and "Act of God"....let us know.

 

[msquared48]

I generally design to a certain span ration, but do have a limiting value when I am working with long beams over certain components such as windows or garage doors so that the components are never subjected to the loads of the beam.

I also put in higher constraints when working with special floor finishes such as tile, marble, and concrete containing infloor heating coils.

These applications I use whether or not it is residential, commercial or otherwise. Makes no difference.

Mike McCann
MMC Engineering

 

[geoffwatson]

In European Codes there is a final deflection value for situations where there are brittle finishes and a separate value for non brittle finishes. In the UK version it is span/250 and span/125 respectively for timber but the joists or more precisely the floor has to be checked for vibration and this usually governs the joist depth particularly for spans over 4m's (approx 13'). In the previous code there was a limit of 14 mm (.55") for deflection basically to limit "bounce" on long spans but a vibration check was not mandatory.

 

[geoffwatson]

As a rider to my last post European codes theoretically allow the client to set deflection limits but as a previous poster reflected most clients would not be aware of a reasonable deflection value and would would almost certainly turn the question back to the engineer.

 

[StructTaco]

Table 1604.3 in the 2006 IBC states deflection limits for various members, Roof (with and without plaster ceilings), Floors, Exterior walls, etc. There is no difference in Residential and Commercial.

There are different limits depending on dead plus live loads, live load only and seismic or wind loads. All are based on the span of what you are looking at, not an arbitrary dimension.

You have to design for the specific structure, whether it has carpet, tile, concrete,...whatever, you have to calculate the dead load. A 10 psf dead load is more like a wood deck, not a floor for a residence or a commercial building.

If you want to eliminate sag, then camber the beam, if it's big enough, typically 100% to 150% of dead load, depending on what material the beam is made of.

Dynamic loading or floor vibration is typically only addressed in steel supported structures, or rather, I've only seen literature on vibrations due to foot traffic for commercial buildings with steel framed floors with a concrete deck.

 

[csd72]

StructTaco,

Vibration criteria do exist for concrete and timber structures. Concrete vibration is usually only an issue for vibration sensitive uses, timber joist vibration on the other hand is a major issue.

I suggest you do some research.