12

The conventional wisdom is that hardwood floors need a 1/4-3/8" gap to the wall (or bottom plate) around the perimeter, hidden under the baseboard. Ostensibly this is for thermal and/or humidity expansion.

For a floating floor, I get needing a gap, it may move a little even just by being walked on.

But for a floor glued or nailed down to a subfloor, what's the point? Your subfloor is subject to the same environmental effects, so it's going to expand too, isn't it? I mean, an OSB subfloor may not expand at exactly the same rate as hardwood, but it should be close—it's only 5% "not-wood."

Thoughts?

Rohit Gupta
  • 2,494
  • 2
  • 16
  • 26
Huesmann
  • 1,214
  • 6
  • 16
  • 4
    OSB and plywood have their grains going in different directions within the sheet. Hardwood is all the same direction so it will expand in direction. Your location might be a factor in how much the wood expands, South east states different from north east and south west. – crip659 Feb 10 '23 at 13:41
  • 2
    Because of the way they're made, OSB & plywood do not (effectively) expand and contract with moisture. – FreeMan Feb 10 '23 at 18:02

3 Answers3

21

When an irresistible force meets an immovable object, in this case you get floor buckling. It's been observed in practice countless times.

Feel free to repeat the mistakes of others, or learn from them, as you like.

If it's the baseboard you don't want, you can hide the gap under drywall or plaster, with a little more finishing effort, whether that be time or money.

What actually happens with a nailed-down floor is that everything shifts a bit (nails have limits) when it first swells, and then as it shrinks tiny gaps open up (one reason that narrow boards are more commonly used than wide boards - the gaps between are smaller and less noticeable for narrow boards) and after that the gaps between boards change size with the seasons.

Ecnerwal
  • 174,759
  • 9
  • 212
  • 440
6

Humidity is a much bigger factor than temperature in wood expansion. And most of the humidity-related expansion is across the grain, not along the grain.

The floor and subfloor are subject to the same environment, but no, you can't assume that they'll react the same. If your subfloor is solid boards (common in old houses), then the flooring is likely perpendicular to those, so it'll expand in a different direction than the subfloor. If the subfloor is OSB or plywood, then it won't expand much at all. Also, different types of wood expand at different rates.

Mike Baranczak
  • 1,637
  • 11
  • 12
  • 1
    If that's the case, why bother with a gap on the *ends* of the boards? – Huesmann Feb 11 '23 at 12:36
  • 1
    ...because while limited, and much less than crossways, there is change in length with moisture, (0.1% in the previously linked study) and normal hardwood installation puts the long dimension of the board the long way in the room, so there's quite a bit of length for that small percentage of movement to occur over. – Ecnerwal Feb 11 '23 at 15:52
-1

Good question.

Oak has a coefficient of linear expansion along the grain of 0.0000027, which is per deg F. for a 20 ft run of oak that equates to .000054 ft per deg F, or 0.0011 ft (0.013 in) for a 20 deg F change in temperature. So yeah, I agree, that the 1/4" to 3/8" gap recommendation around the floor seems like it's about 10X bigger than it needs to be.

I would like to see the rationale, with some analysis to back it up, for the 1/4" -3/8" gap recommendation.

SteveSh
  • 6,076
  • 1
  • 8
  • 25
  • 2
    that's temperature. what about humidity, and what about expansion/contraction of the subfloor that pulls the boards (minutely) apart? – P2000 Feb 10 '23 at 18:02
  • 6
    Long grain expansion isn't the thing to watch for, it's the cross-grain expansion that'll get you. – FreeMan Feb 10 '23 at 18:03
  • No if the temperature and humidity never change otherwise yes. – Gil Feb 10 '23 at 18:06
  • The cross grain and along grain CLTE (Coefficient of Linear Temperature Expansion) are pretty close to the same. 0.0000030 vs 0.0000027. I would expect the subfloor to behave about the same. Don't know about the effects of humidity changes. – SteveSh Feb 10 '23 at 19:05
  • 6
    Humidity is by far (orders of magnitude) the greater concern here. Temperature shouldn't change much in a conditioned space. – isherwood Feb 10 '23 at 20:49
  • 1
    So 20F is the maximum range of temps the inside of a house might see, even if the house is left empty for a vacation during a heatwave or freeze? I'm skeptical, and wouldn't like to be the family coming home to buckled floors because their flooring guy assumed the AC would always be running. – Dewi Morgan Feb 10 '23 at 21:56
  • If the 20 foot floor I used in my example changes by 0.013" for a 20 def F change in temperature, then even a 100 deg F change is only going to increase the expansion to 0.065", or 1/16", or 1/32" on each side. This is still well under the recommended 1/4"-3/8" gap on each side of the floor. – SteveSh Feb 11 '23 at 01:54
  • As already mentioned, moisture content across the grain is a much larger issue. https://hardwoodfloorsmag.com/2018/12/03/expansion-solid-engineered-rift-white-oak-flooring-increase-moisture-content/ – Ecnerwal Feb 11 '23 at 02:15
  • Dump some water on the floor and you'll see why you wanted that gap. – gbronner Feb 11 '23 at 05:17
  • @Ecnerwal if one goes by that study, it means a 10' course of hardwood planks will expand by nearly 2" when you double the moisture content. Sounds like you'd need a helluva lot bigger gap than 1/4" on the ends... – Huesmann Feb 11 '23 at 12:39
  • @Huesmann - Yeah. I suspect a lot of that expansion, in a flooring application, is taken up by the small spaces between the boards. Or the moisture change isn't near as much as was stated in the study because floors are usually sealed on one side. – SteveSh Feb 11 '23 at 12:51
  • @Huesmann - only if they were installed in the driest condition of the study, utterly tight. – Ecnerwal Feb 11 '23 at 14:17
  • @Ecnerwal it would be nice if that study was done for boards in an interior environment with a humidity varying from 30-70%, not outdoors. – Huesmann Feb 11 '23 at 15:31
  • Feel free to hunt for a different study. Or infer from this: https://www.wagnermeters.com/moisture-meters/wood-info/how-rh-affects-wood-mc/ – Ecnerwal Feb 11 '23 at 16:00
  • @Ecnerwal looks like actual moisture content only varies by a couple percent within the likely humidity zone of a house. Based on the first study you linked, a couple MC% causes at most 0.18% increase in the width of a board. Over a 10' wide run of boards, that would be 0.2". – Huesmann Feb 11 '23 at 20:32
  • Guess I don't understand the reason for the downvote on my answer. So far no one has been able to dispute or find an error in the simple analysis that was presented. – SteveSh Feb 13 '23 at 13:04