GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Reducing annoying floor vibrations – interesting paper

andyfrog | Posted in General Questions on

I found an interesting paper on reducing annoying floor vibrations. I thought it might be useful so I am sharing it here.

https://www.woodworks.org/wp-content/uploads/seaoo_2012_conference_design_of_wood_floors_to_mitigate_floor_vibration_problems.pdf

The gist of the recommendations are:

1. Use a higher live load of at least 40 psf in calculations.
2. Go up one size in joist depth over code min or use L/600 deflection limits.
3. Glue and screw the floor sheathing.
4. Use strongbacks with floor trusses.
5. With TJIs, discuss specific vibration concerns with manufacturers.
6. With solid sawn joists, multiply the maximum L/360 span by 0.91 / Use 40 psf live-load and live-load deflection of L/480 for joist spans up to 20 ft.

The main reason for these methods is to increase the stiffness of the floor to increase the natural frequency of the floor above 15 Hz. Humans are apparently most sensitive to vibrational frequencies around 7-10 Hz. The stiffer the floor, the higher the natural frequency.

2012 was a bit ago, so maybe some of the recommendations have changed.

I’m editing in Akos’s notes because they seemed rather salient:

What is not intuitive is that when you are looking at longer spans, even something considered stiff (ie l/480) can be bouncy. A small increase in span can significantly reduce the natural frequency of the assembly, it doesn’t scale linearly at all. Once you get 20’+ clear span, deflection criteria no longer works, you have to check natural frequency. Manufacturer’s design software is definitely a big help.

The biggest issue is that sometimes getting non-bouncy floors can double the floor joist budget. I’ve ran into and it took a lot of self convincing that is worth the cost.”

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.

Replies

  1. Patrick_OSullivan | | #1

    Great reference!

    Years ago, in my mechanics of materials class, the professor said something that stuck with me. He explained how in his rather new house, the second floor was bouncy. The floor was plenty strong as far as load goes, but the deflection was not really reasonable from a subjective point of view.

    Years later, I realize that this is something that good residential designers have known all along. But his research was carbon fiber for submarines, so he was probably not hip to such things. ;-)

    > 2012 was a bit ago, so maybe some of the recommendations have changed.

    The biggest change is that the engineered lumber companies have free web based software that can evaluate your framing. They will go beyond code limits but also provide a rating on the floor comfort based on overall construction. Weyerhaeuser ForteWEB is one such tool. Note that ForteWEB, as well as BC Calc from Boise Cascade allow you to specify dimensional lumber members as well, so their utility is not predicated on using engineered lumber.

  2. user-723121 | | #2

    I have always used 3/4" t & g plywood glued and fastened withs screws for subfloors. I even glue the groove of the plywood as sometimes there can be some squeaking without it. This makes for a quite an airtight floor system and can serve as a primary air barrier for rooms over unconditioned spaces. My lumber supplier normally stocks a 7 ply subfloor that works well with 24" oc open web trusses. The construction adhesive really stiffens up this floor system.

  3. DC_Contrarian | | #3

    Not to be too Contrarian, but don't all of those recommendations boil down to "build the floor stronger than code minimum"?

    1. Expert Member
      BILL WICHERS | | #4

      Looks like that to me too. If you want a less bouncy floor, built it "stronger", which will also mean stiffer, with less deflection and less bounce. If you want it to be quieter, the same applies. The gold standard here I suppose is a concrete floor poured over steel floor pan, which is typical of commercial buildings, which tend to have very stiff floors and you don't hear much sound penetrate them from the other levels.

      BTW, the difference between building "safe" and building "looks safe" is well known in the engineering world. That's why there are tables for "safe" based on loading, and tables for "looks safe" based on deflection over length, which is what people see. The "looks safe" tables are much more conservative.

      Bill

      1. DC_Contrarian | | #9

        The reason I harp on this is that in construction, cost is always a factor. Some people like to say that code is the worst house you can legally build, but I'll counter by saying that the goal of construction is to build as cheaply as possible and still meet the requirements. So overbuilding this unnecessarily defeats that goal.

        The problem I have is that it's not very quantitative. Are we supposed to do all of these things, some of them, one of them? If some or one, how do we pick which ones? "Go up one size on joists" seems like an awfully blunt instrument.

        The problem is the code is designed for one requirement -- deflection -- but there is another requirement -- resonant frequency -- that the code doesn't exist. And yes, an assembly that meets the deflection requirement doesn't always meet the resonance requirement. But without being able to model resonance the same way that deflection is modeled there's no way to design for it.

        1. andyfrog | | #10

          The rules of thumb were provided for convenience. If you would like to model resonance, the equations are provided in the paper.

    2. andyfrog | | #8

      It's more about the ratio of deflection to span, as Akos has mentioned. The whole point is that you can build a very strong floor that is still perceived as bouncy if you don't take measures to raise the natural frequency of the floor.

      Notably, and this is counterintuitive--going from 24 O.C. to 12 O.C. doesn't do nearly as much for bounciness as increasing the depth. But it does make for a considerably stronger floor.

  4. user-723121 | | #5

    This is good information. Just a short while ago a GBA poster referenced trying to span about 25' with dimension lumber and I think 24"oc. If the builder or owner had access to the above they may have avoided the trouble.

  5. Expert Member
    Akos | | #6

    This is the most important part:

    "increase the stiffness of the floor to increase the natural frequency of the floor above 15 Hz"

    What is not intuitive is that when you are looking at longer spans, even something considered stiff (ie l/480) can be bouncy. A small increase in span can significantly reduce the natural frequency of the assembly, it doesn't scale linearly at all. Once you get 20'+ clear span, deflection criteria no longer works, you have to check natural frequency. Manufacturer's design software is definitely a big help.

    The biggest issue is that sometimes getting non-bouncy floors can double the floor joist budget. I've ran into and it took a lot of self convincing that is worth the cost.

  6. Expert Member
    MALCOLM TAYLOR | | #7

    Akos is spot on. Increasing the natural frequency above 25 HZ is key.

    You can find a good summary and discussion here: https://www.finehomebuilding.com/forum/floor-vibration

    A couple of other useful rules of thumb:
    - If you double the span of a joist, it will defect eight times as much.
    - If you double the depth of a joist it will defect 1/8 it's original amount.

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |