GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X 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

Deflection for a Stiff Floor

alexdorf | Posted in General Questions on

In looking at joist span options, what is the ideal deflection # to hit for a “stiff floor” ?I’m trying to better acquaint myself with the understanding and importance of deflection when considering a span.

I have a general understanding of L/360 : L/480. However, is there a number that one would want to achieve within these ratios that would equate to a stiffer floor.

Reason I ask is that I am pushing the limit on these plans I’m toying with to do joist/floor truss that would span the 40′ and meet the PSF load requirements. All the bearing/support would come form the exterior walls (pole & barn) style, no internal beams to support the joists. This puts my deflection at 480/480 – basically an 1″ if I’m not mistaken.

Any insight would be fantastic! Thanks!

Is this super bouncy (not very scientific, I know) – again, simply looking to get what you guys would consider “stiff” / high performance.

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.


  1. Patrick_OSullivan | | #1

    Both Weyerhaeuser ( and Boise Cascade ( have online design tools that allow you to spec the span, members, and subfloor and then get a quantitative rating on the performance of a floor. I suggest checking these out.

  2. Expert Member
    NICK KEENAN | | #2

    Keep in mind that the rated deflection is at rated load -- typically 40 pounds per square foot live load plus 10 pounds dead load. A 40-foot joist spaced at 16" is supporting 53.3 square feet so that deflection is at a load of 2666 pounds. A load of one person would give about 1/16" of deflection.

    Also look if your joist manufacturer has a recommended depth to length ratio. For that length you may have to go with quite deep joists.

    1. alexdorf | | #8

      How did you guess my weight, albeit I am 10 lbs heavier right now ;)

      This gives me a bit of better understanding. This is one example of a Joist I was looking at - If I were to go 24" in depth, I'd be able to comfortably be at 16" O.C.

      And for some reason, based on my limited knowledge of this area, I still feel a bit unsure about the stiffness/comfortability of the floor joists, albeit it would seem to be struct sound. Again, I wouldn't want the chandelier on the floor below to be swinging all the time.

  3. AlexPoi | | #3

    L/480 is what you should aim for if you want a stiff floor. L/720 if you want to install a stone floor. I think L/360 is fine for a roof.

    1. alexdorf | | #6

      Right, so in my case, I'm 480/480 , which feels like at the bitter end of the allowable deflection amount, and again, concerned about quality of stiffness / vibration/ etc. - looking to better understand the number to hit here.

      1. burninate | | #19

        The flooring tile manufacturers demand L/720 or better. Code typically demands only L/360 or better for living space, some even go down to L/240. Truss manufacturers tend to put L/360 and L/480 in their tables as the main use-case.

        Human beings and joints in drywall care less about relative deflection than about absolute deflection. Absolute deflection of 1"-2" at full load rating is likely to feel perceptibly bouncy even under a single person's weight, and may lead to significant differential deflection around furniture & walls. I've read of custom builders who tell clients they need to reduce it below 1/2" absolute; Otherwise they get callbacks. Also a forum post mentioning a 0.4" limit proposal. Or a HUDUser document suggesting limiting it to 1/2" in long spans.

        Personally, I will take that advice and aim for 1/4" at 40psf, which is L/960 at 20ft at 40psf, or alternately it's equivalent to L/360 at 20ft at 104psf. The cheap & easy way to do that in the 20ish foot span range seems to be deep (24") parallel-chord floor trusses. For substantially longer than this, as you get into the 30-40 foot range, you may need to upgrade to steel; Cold-rolled steel joists are just barely capable of this.

        The effects of inadequate joists can be mitigated partially by tying joists together horizontally with strongbacks and thicker sheathing, so that point loads end up spread out over multiple joists.

  4. natesc | | #4

    This is a great question. People with I-joists and floor trusses that span long distances, and are engineered for safety and code compliance have actually found the livability "annoying."

    "However, the traditional deflection criteria resulted in total deflections of long span floors that caused occupant annoyance, and owners began to complain about the vibration performance of wood floors. In other words, the requirements of design codes provide a safe structure, but fail to provide a serviceable structure"

    "A new design criteria for preventing annoying vibration in wood floor construction has been presented. The criteria requires that the fundamental frequency of the floor joists and combinations of joists and supporting girders be greater than 15 Hz. The criteria was developed using vibration test results of both Minatory and in-situ floors, is relatively easy to apply, and uses information that is readily available to the average architect, engineer, or contractor. The reader is cautioned that the design criteria is to be considered prelimay due to the effect of occupancy loads has been neglected. A study of the effect of occupancy loads is currently underway to verify the design criteria’s effectiveness under typical load conditions."

    1. Expert Member
      NICK KEENAN | | #7

      My understanding is this is where depth/length ratio comes in. If the length is too great for the depth the resonant frequency becomes annoying low, even if the deflection of the joist is fine.

    2. alexdorf | | #9

      Exactly, this is my concern.

      I suppose partition wall, while nonbearing would provide some relief if positioned correctly.

      Again, my goal was "is it possible to span 40' and not compromise livability in the form of stiffness/vibrations/bounce"

  5. alexdorf | | #5

    Thanks. Need to pay around with it. I'm still quite hazy on the stiffness meets deflection area.

  6. Expert Member
    Akos | | #10

    I doubt you'll get 40' without bouncy floors. Resonant frequency is effected by both span and deflection, as you go longer, you need much lower deflection to have similar resonant frequency. Without doing the math, it is hard to say what deflection you need.

    Interior walls only help if they are connecting the floor to a solid roof structure, otherwise they are only adding more mass, thus lower resonant frequency.

    I think for your span, it is probably simpler to go with a couple of steel wide flange beams and hang the floor joists off that. The joist would now run at 90 deg to the 40' span and can be much shallower. Besides getting you floors that feel solid, for a multi story structure, you won't have each floor taking up 24"+ of vertical space. This can make a big difference in ceiling height if you have a height limit.

    1. alexdorf | | #11

      Got it sort of.

      My dimensions are 40' x 96' . So, If I were to add an interior load bearing beam, it'd probably be around 18' in supporting a 22' span instead of the 40'. I'm gonna discuss with some MFR's as well. I'm considering cold formed steel joists as well, which I'm guessing can make a difference.

      1. Expert Member
        Akos | | #12

        It would be pretty expensive to span the 96' without any columns. Easy to do 1/2 that if you don't mind a post.

        If you want full clear span without any posts, you can set the wide flange in the 40' direction and split the 96' up into four sections. The floor joists would now run in the 96' direction between these beams. This way the floor joist only need to span 24' which is easy with I-joists.

        I don't think you'll get 40' out of CFS unless you get some pretty tight spacing. At that point you are spending way more on them than the three wide flange you need above, plus I still don't think it will be stiff enough.

  7. Expert Member
    BILL WICHERS | | #13

    There is a table in the Strut catalog that lists a few different deflection numbers with 1/ values, and the last column is based on VISUAL deflection -- it's much more conservative than the others. I remember way back in college a professor saying that if you built a bridge with steel to be "safe", no one would use it because it would bend too much and everyone would be scared. It would LOOK like it was going to fail.

    With a very long span, you'll have more bounciness even if the load numbers are all good. If you put any walls on there, you could potentially have cracking corners (if using drywall) too.

    If you really want to have a very large clear span, I would use steel and build more commerical style. Steel allows much more stiffness in a given area compared to typical wood structures, and you can hang shorter wood joists between a relatively small number of steel beams to get where you need to be.

    A mid-span column is going to be cheaper though...


  8. Expert Member
    NICK KEENAN | | #14

    I have a nagging feeling the OP might be mis-reading the span table. I'm looking at the biggest joist in my table, the 16" TJI 560, and it has a maximum span of 32' 11" at 16" spacing and 36' 1" at 12" spacing.

    1. Expert Member
      MALCOLM TAYLOR | | #15


      It's a bit unclear but I think he is considering parallel chord trusses. My builder friends on the Canadian prairies say they have become the go-to method of framing main floors over basements there to provide a clear-span usable basements. I have no idea how deep they end up being. Some time ago my truss designer gave me the rule of thumb for roofs which was about 1" depth for each foot of span.

      1. Expert Member
        BILL WICHERS | | #17

        Malcolm, assuming you have some expierience here, are those parallel chord trusses cost competitive with something like a few big glulams with "regular" joists hung between them? I don't do enough light construction to have run into stuff like this. In the commercial world, big clear spans like this pretty much always end up as a heavy I beam, a big truss, or sometimes prestressed concrete beams. Basically the same stuff you see used to build highway overpasses and stadiums.

        In the telecom facilities I'm most familiar with, we almost always find it's cheaper to just use columns to shorten the spans, but our concerns are usually holding up very heavy things (generators, large UPS systems), not so much clear spans.


        1. Expert Member
          MALCOLM TAYLOR | | #18


          All I know comes from what my friends have told me. I don't know if floor trusses make sense economically, or it is one of those regional practices that take hold and are never really questioned.

          Here houses are probably 50/50 slabs and crawlspaces. The interior bearing in the crawlspaces is usually provided by knee walls on strip footings, with either lumber or TJs above.

        2. Jon_Lawrence | | #20


          I used parallel chord floor trusses and back in 2017 they were a bit cheaper than TJI's. I don't know what the cost comparison is like today. Mine are finger jointed and glued, no nail plates. We designed to L/480, I don't have any spans longer than 22', 3/4" Advantech subfloor and 8" wide flooring, no noticeable deflection/vibration. I like that the trusses allowed me to easily snake the ventilation pipes through without drilling holes in joist webs.

          As far as the cost comparison between steel and glulam, I have both a 35' long w21 111 wide flange beam that is in the first floor ceiling and directly above it in the 2nd floor ceiling a 32' long Anthony Powerbeam that is 7" wide and 29" deep. The Powerbeam was much cheaper ($1800 vs $2900) than the steel girder - again at that time. My engineer told he prefers to spec wood over steel for the 2nd floors because it moves less on windy days. Sounds like he may have taken the same class as you in college.

          1. Expert Member
            BILL WICHERS | | #23

            That's interesting, I've always found steel to be cheaper than wood for larger beams (I was using comparing to big LVLs though).

            I'm surprised about the "moves less on windy days" too. My professor was talking about deflection and people's thinking about what "looks" safe. The same rules for "looks safe" apply to all materials. I haven't really compared glulams to steel for stiffness, but it's possible that steel could bend more before failure than wood, which would make wood "look" safer since it didn't deflect as much. The solution to that would be to use a steel beam with a greater distance between flanges, for more stiffness.


    2. Expert Member
      Akos | | #16

      I've never tried to use them, but 24" TJI can span 40' at L/480. I think they would be pretty bouncy at that span though.

      Since it is pretty close construction, I would guess 24" parallel cord truss would be similar.

    3. Expert Member
      NICK KEENAN | | #21

      The link the OP provided in #8 shows 24” RFPI 900 spanning 41' - 10" at 16" centers and 40 PSF live load. So I stand corrected.

  9. Expert Member
    Michael Maines | | #22

    I use code-minimum deflection ratios as a starting point, but for most floors I bump up to L/480. Any floor over about 14-16' is at risk of feeling bouncy at L/360. I learned from licensed engineers to never exceed about 5/8" deflection so regardless of the ratio, I don't allow more than about 0.6" deflection. In some cases, such as the bi-parting exterior pocket doors I'm working on now, with a 14' clear opening and 22' header supporting a roof with a 90 psf snow load, deflection ratios are useless and I just design to a specific deflection distance--in this case, 0.1", or L/2600.

    Something to be cafeful about when using engineered material for floor framing is that it's so consistent in composition that it's vulnerable to harmonic vibration--when you walk across it, even if you're not close to its rated load, you can trigger harmonic vibrations that feel like a bouncy floor.

    Edit to add: I see that others have already mentioned harmonic vibration. Here is a helpful table from the IRC about what deflection ratios are considered the minimum standard:

    1. Expert Member
      MALCOLM TAYLOR | | #24

      "such as the bi-parting exterior pocket doors I'm working on now, with a 14' clear opening and 22' header supporting a roof with a 90 psf snow load"


      I'm going to have nightmares for about a week after reading that sentence. :)

      1. Expert Member
        Michael Maines | | #25

        Haha, don't worry Malcolm, my first round of calculations show a W14x34 will handle the load but I'll check again before placing the order. The door is well inside a covered porch for a house that will be used mainly in warmer months.

        Do you have experience with exterior pocket doors? This will be my first time with them but the details (from Loewen) look pretty good, all things considered.

        1. Expert Member
          MALCOLM TAYLOR | | #26


          No, I've never even heard of them. Coniferous needles, which you would think should be a fairly minor concern, preclude all sort of things on many sites here. They make any sliding doors difficult, help water bypass gasketted closures, clog ridge vents, and make gutter guards ineffective.

          As you get more adventurous and explore new building assemblies and products, I seem to be going the other way and becoming more risk adverse and conservative. Age I guess.

          1. Expert Member
            BILL WICHERS | | #27

            >"As you get more adventurous and explore new building assemblies, I seem to be going the other way and becoming more risk adverse and conservative. Age I guess."

            Expierience sounds better than age, and is probably more accurate too :-)

            I am more reluctant to try the latest thing at work too, and a lot of that is because new stuff sometimes has exciting and unforseen ways to fail. My industry is all about uptime (preventing outages), so surprise failure modes are a BIG deal. Older stuff is better understood, and therefore more predictable. I like that. As new stuff gains time in the field, it gets more interesting to me. I have to be able to tell my customers with a reasonable degree of certainty what their probability of an outage is, and what their ongoing maintenance costs will be. Surprises aren't a good thing for long term contracts.


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |