Our floors are somewhat bouncy, especially in big rooms with long joists. I think that the main cause is that they used 2x8 joists on 15-20' spans.

We are going to replace/install hardwood in several areas. Given that we'll have access to the subfloor, what can be done to make the floors less bouncy?

Note: looks like it's different from the situation described in What are these spongy/soft spots on my hardwood floor?, since the floor moves as a whole and not just in some places and there are no hollow spots there.

  • Do you have access to the joists below the floor?
    – Joe
    Jul 24, 2010 at 2:23
  • @joe: No, unfortunately there are rooms (ie finished ceiling) below.
    – Rom
    Jul 24, 2010 at 3:35

5 Answers 5


It might be worth checking out the subfloor if you can get access to it in case that is what is not holding up to weight instead of the joists.

Otherwise, I know in your comment you stated that you don't have the joists exposed, unfortunately, I don't know of a way to resolve the issue without exposing them as anything you tried to do at the floor level would just be adding weight.

If you don't feel your joists are up to snuff, you can reinforce them by laminating them.

You would get additional boards of appropriate size to fit snugly next to the joists, then use an adhesive and screw the boards to the joists.

Make sure you are getting the additional wood over your load bearing walls, otherwise again you are just adding weight to the situation.

Remember though, even if you are going to make it a DIY project, structural work is a big deal and you should at least consult a contractor/architect and make sure you get proper permits and inspections.

You can save yourself the money on the labor, but don't skimp on the planning.

  • Laminating them is the best way if you really do not want to rip out the ceiling to get at them. Then you should place a few 4 x 8 sheets on the floor and see if it is still bouncy. If it is the next thing to try is to add some 2x8s between the floor joists. May 22, 2014 at 21:58

ManiacZX mentioned laminating the joists below, and in my opinion, if you had access to the joists, you'd be better off installing bracing between them (cribbing, bridging, whatever you want to call it) ... but that assumes that you actually had access to the joists.

as you're going to be gaining access to the subfloor, you might be able to stiffen the floor as a whole -- put down some adhesive, lay more plywood/chipboard on top of it (ply's more expensive but it flexes less ... I'm not sure how MDF compares to plywood on flexibility), and then screw it down tight to the original subfloor. (and a lot of screws -- might even be worth renting a belt fed screwgun, as you're going to want them every 6" (15cm) or so to make sure the new subfloor is bonded well to the existing one, so they flex as a single unit.

If you can get the floor stiffer, it'll distribute the load better across multiple joists, and therefore they should deflect less when loaded.


Your other option is to come in from the top -- strip the existing subfloor off, install the cribbing, then lay the subfloor back down. Depending on how the subfloor is laid, you might be able to cut windows into it (set a circular saw so it just barely goes through the sheet; you might have to drill a hole and use a wire w/ a bent end as a guage first), expose enough to set the cribbing, close it back up, then to the lamination thing I mentioned above to stiffen the subfloor.


It's possible that installing the hardwood floor might help to spread load better (especially if it's laid so it's perpendicular to the floor, and nailed in, rather than a floating floor), but it'd really suck to go to the work of laying the new floor, and then have to tear it up because it's still a problem.

  • 2
    Cross-bracing would provide significantly less extra rigidity than doubling the joists. Think about it: you're using the bracing to transfer some of the load on one joist to another joist in an inefficient manner (by pushing diagonally on the bottom of the next joist that's 16" or so away). With sistered joists, you're transferring that load much more directly to a whole new joist instead. Jul 26, 2010 at 17:23
  • 2
    @Mike : actually, the load transfer comes from both pulling and pushing, which is quite efficient (it's the basis for truss bridges). The advantage to the extra bracing is that you (1) don't have to figure out how to tie in support for the new joists (as if you can't get below the joists, you likely can't add new joist hangers) and (2) can install much easier as you don't have to tear up the whole floor to expose the whole joist and (3) you don't have to deal with pipes & cables that might be going through the joists already.
    – Joe
    Jul 26, 2010 at 17:46
  • 2
    I'm with you on (2) and (3), in fact (3) is something I hadn't thought of that might make sistering completely out of the question. But comparing cribbing to a bridge truss is apples & oranges -- in a truss, the diagonal members are parallel to the bending direction of the beam and serve to resist internal shear, which stiffens the beam. With cribbing, the diagonal members are perpendicular to the beam you're stiffening and are only there to transfer load across to the next joist. Jul 26, 2010 at 19:52
  • 1
    Oh, and regarding (1), you don't need to support the ends of the new joists if they're securely attached to the existing joists (e.g. by glue and screws as suggested by ManiacZX). Even sistering another 2x8 that only spans the middle third of each joist will provide a significant improvement in stiffness, since most of the bending happens in the middle of the span. Jul 26, 2010 at 19:55
  • 1
    @Mike : I get it, you don't like my suggestion. And yes, by stiffening the center of the beam you won't have as significant of moments at the ends (which is where a failure would occur); but you're off on the truss -- it's a three dimensional problem: you're creating a series of trusses that are joined by the joists; both are rigid in their own way. I admit, it's been more than a decade since I've had to calculate beam deflection or loading in a truss, but until you know what the size of the bridging selected is, there is no way you can say it is or isn't going to be stronger.
    – Joe
    Jul 26, 2010 at 21:39

Really, 2x8? That's undersized for a 20-foot span, even if your house is old enough for full-sized lumber. 15 foot would even be iffy in my book for a 2x8 (though I'm not a structural engineer).

Like Mike and ManiacZX, I would recommend tearing out the subfloor and laminating the existing joists across the entire span, both for better carrying the load and helping to decrease the give in the flooring. Modern flooring is much less forgiving than, say, old hardwood floors.

If you have large spacing of joists and sistering still doesn't get you down to 16" OC or less, add bracing as Joe recommended. Also, use two layers of new subfloor, perpendicular to one another and fastened with both lots of screws and adhesive to one another and the joists.

Also, especially if this is an older home, check the level of the home itself--that may be one source of the problems.


A 2x8 joist on a 20' span is extremely undersized. What is the center-to-center spacing of your joists? If I were you I'd seriously consider pulling up the subfloor at least in the center third of your longer spans, and sistering in extra joists as described by ManiacZX.


Sistering the joists will have little effect on reducing flexation. A 2x8 for example has a modulus of elasticity of 12,300,000 flb in-2 Two 2x8 will have the same modulus of elasticity. The easiest and better solution is to use metal strapping and run it diagonally nailed to each joist and bent around the edge and nailed. This practice will transfer the downward flexation into a radial flexation which in a floor system is much greater.

  • 1
    Assuming constant E and I, Flex is F(x)/EI where F'''' is the loading function. (Yes, 4th derivative) Take x as L/2 for midpoint deflection. Sistering the joist will not change E at all as you've stated. It will change I -- namely it will double it, or half the deflection. Feb 26, 2013 at 22:16
  • And 40 2x8's will have the same modulus of elasticity, as will a single toothpick of the same material. Feb 26, 2013 at 22:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.