The floor on which I am installing 3/4 oak has 1/2" plywood subfloor nailed to joists. To keep same height as hallways and other rooms, I am installing 1/2" plywood to the subfloor before installing the hardwood. Should I fasten the 1/2" plywood sheeting through the subfloor to the joists (i.e. ring nails withframing nailer), or should I only fasten the sheeting to the subfloor (i.e. short ring nails or screws)? Also, direction of long grain perpendicular to subfloor grain, or parallel? Thanks.

 

Wow, it's difficult to believe there is only 1/2" plywood as a subfloor. Minimum, IMO, would have been 3/4". If you add subflooring layer, you will screw your old subflooring to the joists, then apply new subflooring, screwing it in the field of the lower layer, intentionally missing the joists. Keep your grain perpendicular to the joists, and stagger the joints. Screwing the original subflooring will help prevent squeaks.

 

Thanks. What would be best option for the type of screw?

 

1 1/4" torx decking screws, IMO.

 

Chandler, why do you intentionally miss the joists when nailing or screwing the top layer of subfloor? I am just curious. Thanks

 

With just 1/2" subflooring there's going to below spots between the joists.
Your trying to pull the two pieces together to get rid of that gap.
Not sure what the logic was but almost every house I've seen built in the 70's had 1/2 subflooring, then 3/4"on top of it. (often times just useless particle board) To me that's just backwards.

 

As I understand it the reason is to allow both layers to "slip" in relationship to each other during changes in temperature and relative humidity. My personal thought is that this is hogwash. With the number of fasteners and how closely spaced they are the two layers should be monolithic and NOT move independently.

I feel the same about tile floors that have a wooden subfloor, thinset mortar, backer board fastened to the wooden subfloor and then more thinset and finally the tile. All those different materials have bonded to a single monolithic floor and ANY movement of the individual layers is going to result in severe cracking or other problems.

I'm quite certain that all the professionals who do the work will disagree vehemently but have you ever seen a problem when the rules you post are not followed explicitly? How about some details that prove your rules are not just the arbitrary rules that have been passed down for decades without any scientific basis? Change comes slowly to construction methods and some people never change.

I'll offer some personal experiences. In my old house I nailed down 1/2 inch exterior plywood to my existing floor and then glued down a parquet floor. I put 6-mil plastic under the plywood for a vapor barrier, another no-no according to the accepted way of doing floors. I lived in that house for another ten years and had ZERO problems with the floors or the underlying structure. In the bathroom I laid 1/8 inch tempered hardboard into a flooring adhesive with the hard side up and then added sheet vinyl. I never had one iota of problem with that floor either. In my present house I had a high-end laminate installed in the living room (floating floor) and then proceeded to install a floor-to-ceiling bookcase on top of the laminate. According to the theory this should have negated any movement of the laminate where the several hundred pounds have effective "nailed" the laminate to the subfloor but I have had no problem whatsoever with the laminate flooring.

So how about some scientific studies proving the necessity of NOT nailing an additional sub-floor layer to the joists? Even some examples in the real world that are not scientific would be welcome. Prove also that adding a vapor retarder/barrier between layers of subfloor will cause problems. Show me some laminate floors that have buckled because of a heavy weight being set upon the floor and restricting the free expansion and contraction.

 

@Furd,

I'm not sure that your gripe can be "proven" already. Ten years is too short a time for anywhere near the significant amount of damage necessary to prove if any of that is problematic. After all, it's taken over a century for my house to show that its original design for this area would lead to horribly sagging floors.

But, I do agree with your point about proof over theory. I, as a non-carpenter, do not know why the second subfloor would need to be decoupled from the joists. But, if we take your position as true; screwing into the joists will have the same effect. If the options are 1) it might be better to do A, or 2) it's the same thing if you do B, I'll just do A and know that I didn't mess it up, and might have made it better.

 

I've read Larry's advice for years and trust that he knows what he's talking about BUT when I overlayed my floors with new plywood, I screwed them directly to the joists. That was 18-20 yrs ago and I've not had any squeaks or evidence of the 2nd layer of plywood coming loose.

 

I wish I could take credit for the methodology and/or theory, but it supersedes my 20 years in the industry. The way it was explained to me was, the plywood or OSB will tend to expand and contract at different rates. Tying each to the same member (joist) could allow for warping in the field. Keeping the fastener pattern different can allow expansion and contraction for each sheet separately, allowing slip. I am not the one to argue the point,and have no intention to do so. I can see both sides of the argument. Similarly to embedding cbu in thinset. I did it without for years without failure. Don't really see the need in it on a flat floor,but industry standards change, with knowledge far superior to mine.

 

I'm not "griping" but I simply do not see any scientific reasoning behind the methods offered by those that do this work for a living. As a slight analogy it has been known for 60 years that the expansion tank connection for a hot-water heating system belongs on the suction side of the circulating pump yet even today there are hundreds of heating professionals that pipe them backwards. Most of the time there is no discernible difference in the operation of the system but when there are the installers are usually clueless as to why. There ONLY reasoning seems to be, "This is the way I was taught." or some such lame excuse. Even when the boiler manufacturer states in the installation manual the proper method as often as not the installer ignores that method in favor of the old method.

If a subfloor of plywood is fastened to the floor joists, whether with adhesive or not, by screws or ring-shank nails, with a fastener every six inches that plywood is NOT going to move between those fasteners. If a second layer is added and fastened with screws or nails in a six inch pattern all over the field, but missing the joists, that second layer is NOT going to move independently of the first layer. So what possible difference could it make if the second layer ALSO had fasteners into the joists?

Same principle applies to backerboard set into a layer of thinset and then screwed down to the original subfloor in a six inch pattern. The thinset bonds the subfloor to the backer board and the fasteners add to the strength of the bond making the subfloor>thinset>backerboard into a monolithic assembly that simply cannot move independently without breaking the bonds between the layers and destroying the integrity of the sandwich. Add another layer of thinset and then tile and you have a multi-layer sandwich of dissimilar materials, all with different rates of expansion and contraction yet IF they did move in relationship to each other the result would be a failed floor. How often does this happen in the real world where some other agent isn't really to blame? I can definitely see that insufficient strength of the subfloor, from using too thin a material or having the joists too far apart or having joists over-spanned can cause problem but I see no reason why fastening multiple layers of subfloor to otherwise adequate joists should cause any problems.

I can also see where an incursion of water, not just a change in relative humidity and/or temperature but ACTUAL water getting in between layers of plywood could definitely cause a problem but that would cause problems regardless of how the subfloor layers was fastened.


Back some thirty years ago I was a new employee in a high temperature hot water heating plant of a large manufacturing facility. After taking a boiler out of service for cleaning and inspection the "old timers" had the most convoluted method of warming that boiler and putting it into service I have ever seen. It made absolutely no sense and after studying the method for a couple of years I pronounced it as hogwash. I devised a much simpler method and reduced the time and effort by about 80%. It was even worse when we took the entire plant down over a long weekend to inspect and repair parts that required a total shutdown. Prior to my time it took something like sixteen hours to get that plant up and running and there was a huge amount of running around and leaking expansion joints to contend with. After adoption of my improved method it took less than six hours and most of that was spend sitting at the desk watching the various pressures and temperatures stabilize.

Bottom line is, just because something has always been done in a certain manner does not automatically mean that manner is either the best or the only manner that works.