I'm in the middle of a full renovation of a 100 year old house. I'm keeping the original floors on the first floor, and patching them with pieces of the original flooring with the second floor. The flooring is, for the most part, in decent condition. However, the sizes are not consistent, many of the tongues/grooves are damaged, and some of the gaps between boards are up to 1/8". There is no subfloor. When you step on a single board between joists, you can often see/feel movement of up to 1/16".

While I have access to the bottom of the floor, I am considering stiffening it in high traffic areas. My first thought was to smear subfloor adhesive onto 14.5" x ~4' sheets of 3/4 ply or OSB, and then hold those up with 2x4s sistered onto the 2x8 joists. however, I'm not sure how the adhesive may affect expansion/contraction. is that something I need to be concerned about? or do y'all have a better idea for stiffening the floor from below?

  • Blocking may suit better than sistering, but it depends on the orientation of the floor boards w.r.t. the joists.
    – Huesmann
    Feb 29 at 17:13
  • @Huesmann are you implying that it would be possible to run planks parallel to the joists with no subfloor? that would certainly be something 😂
    – ickybus
    Mar 1 at 14:04

2 Answers 2


It sounds like you have a feasible solution to the problem that you've described. I would use 2x2 ledgers instead of your 2x4 ledgers. They're more flexible, so you'll get a better fit without fighting the ledger's bending stiffness. A 10d box nail every foot of ledger will give it plenty of strength and stiffness for this job, but I would install one every 8".

I would use a 3/16" v-notch trowel to apply setting type thinset instead of your subfloor glue. That's about the cheapest gap filler that you're going to find. Be sure to orient the OSB's strength axis correctly.

The temperature stuff is nothing to worry about. Based on 2.8e-6 in/in/°F for pine and 3.1e-6 in/in/°F for OSB, a 10 ft board length, and 40 °F of temperature change between glue setting ambient temperature versus some other ambient temperature, I get a differential lengthening (or shortening) of

(3.1e-6 in/in/°F - 2.8e-6 in/in/°F)(40°F)(120in) = 0.0014 in ≈ (1/40)(1/16").

I have low confidence in the creep characteristics of subfloor glue, so I think of it more as a stiffener against vibration and short time-horizon loads than as a stiffener against sagging under large, long time-horizon loads. To solve your soft foot press problem, I would implement ledgers like you intend, but by the numbers you could just glue the 3/4" panel to the underside without ledgers. ASTM D3498 establishes a 100 psi shear strength for subfloor glue with a 1/16" gap. I get a 16 psi demand for the glue when there's a 400# point load at midspan. That's a (16psi)/(100psi) = 16% demand/capacity ratio, but that's not just randomly squirting some glue on there. That's with a 1/8" v-notch trowel applying the glue as a continuous film below the floor. That would require about 3-1/2 gallons of adhesive per 100 square feet.

The glued composite is so expensive that it makes more sense to buy the cheapest gap filler possible and make the OSB paneling strong enough without any composite action. This is where my thinset idea came from. One bag of thinset applied with a 3/16" v-notch trowel will cover 100 square feet. 3/4" OSB with ledger supports is plenty stiff over 14.5" openings even ignoring any composite action with the pine boards, so the system is fine without glue.

  • i don't know enough about the math to doubt yours, but if 0.0014" of expansion/contraction over 10' is correct, why are flooring installation manuals so insistent about gapping the edges under the baseboard and caulking it? is humidity the larger factor?
    – ickybus
    Feb 29 at 1:33
  • and thank you for the ideas and the math regarding the thinset vs. the adhesive. the cost difference isn't a huge concern at this scale (i only intend to do 100-150 square feet), and if i'm the next guy having to cut a hole in the floor to patch a board or something.... i think my saw blades would rather it be glue than thinset.
    – ickybus
    Feb 29 at 1:45
  • 1
    Axial elongation from 30% moisture content down to 0% moisture content is typically spec'ed as ranging from 0.1% to 0.2%. Assuming a 10% change in wood moisture content (which is very large--see the table on page 316 from fpl.fs.usda.gov/documnts/fplgtr/fplgtr282/fpl_gtr282.pdf), that's about (10%/30%)(0.15%)(120in) = 0.06" ≈ 1/16". Caulk isn't perfectly flexible, so I'll speculate that they anticipate compressed caulk exerting too much force. Double the caulk width, halve its force. Maybe they anticipate 1/8" without caulk getting clogged with inflexible gunk?
    – popham
    Feb 29 at 1:59

You can certainly glue down (or up) pine flooring. It's often done to reduce movement. However, you should use adhesive engineered for the task. As much as I like construction adhesive for many things, it's not ideal here. Use something with higher bond and tensile strength.

Or use a sheet underlayment between the OSB and the planks, such as you would for engineered, vinyl plank, or laminate flooring. It'll allow support without introducing noise. Key here is getting the subfloor pressed up tight. Without the adhesive to take up gaps you need precise positioning. Consider jacking or clamping as you install the support cleats.


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.