Tongue & groove on plywood, glue or no glue?

  • Advertise with us

« back to Designing Woodworking Projects forum

Forum topic by Kabashu posted 05-02-2013 11:28 AM 9172 views 0 times favorited 7 replies Add to Favorites Watch
View Kabashu's profile


50 posts in 3233 days

05-02-2013 11:28 AM

Topic tags/keywords: question

Can anyone tell me the best way to fasten 3/4” tongue and groove cypress to plywood for a barn door? I’ve been told to glue the heck out of it and use screws from the back so it wont move, also been told just use screw slots so it CAN move and never face glue any wood, ever. Thanks

7 replies so far

View Buckethead's profile


3194 posts in 1895 days

#1 posted 05-02-2013 11:57 AM

I like the idea of a couple screws from th back side as a secure means of attachment. I don’t like it as a practical means of installation. If you have a brad nailer or finish nailer, I would apply the tongue and groove first, using that. Blind nailing, which is done by angling the nail gun and shooting the stock in through the back side of the groove end.

Maybe one fastener in every 8” to 12” of linear board. You can use the same method at the tongue side, if you are careful not to shoot the nails in a manner that impede the tongue and groove fitting together. I would install the stock, groove side down, to allow for better drainage.

(I wonder if there is a simple CAD app for my tablet?)

After that, do the screwing.

-- Support woodworking hand models. Buy me a sawstop.

View Cosmicsniper's profile


2202 posts in 3184 days

#2 posted 05-02-2013 12:19 PM

+1 to Buckethead

-- jay,

View Kabashu's profile


50 posts in 3233 days

#3 posted 05-02-2013 12:37 PM

Thanks….my question is… do i glue it to the ply or not.

View Buckethead's profile


3194 posts in 1895 days

#4 posted 05-02-2013 12:43 PM

No need to glue. You could do it, but movement might break the seal or crack the wood. Not a major concern IMO, but I would not bother. Just put enough screws that it doesn’t clap against the plywood. Even that isn’t an issue, but it will make the door feel more solidly built. Do use outdoor rated screws. Not drywall screws… Stainless is the better choice, but costly. Coated deck screws would suffice.

I also saw the board oriented horizontally, for whatever reason. It is most likely oriented vertically, which is obviously fine, but all the same principles still apply.

-- Support woodworking hand models. Buy me a sawstop.

View Cosmicsniper's profile


2202 posts in 3184 days

#5 posted 05-02-2013 01:01 PM

The cypress will expand. When it does, glue will not permit it to expand and it will buckle. Nails alone allow expansion. Skip the glue.

-- jay,

View GregD's profile


788 posts in 3162 days

#6 posted 05-02-2013 02:08 PM

So the plan is to have a door made up from a layer of plywood and a layer of 3/4” solid cypress? The plywood layer will be dimensionally stable and the cypress layer will want to expand and contract so I would expect the door is going to want to cup unless you accommodate the wood movement somehow.

If you use Buckethead’s suggestion of blind nailing then each cypress board is fixed to the plywood only at the groove edge. When the board shrinks and expands the tongue edge can slide in/out of its mating groove provided you’ve left enough space for expansion but not so much that the tongue pulls out of the groove when it contracts. I don’t see any harm in running a bead of glue near the groove edge when installing the board. I’m an amateur and I’d probably take the extra time and do it. It would be best, I think, if the first board (the one with no tongue) was relatively narrow since both of its edges will need to be fixed to the plywood unless you do something clever on the edge opposite the groove.

If instead you’d like the tongue/groove joints made up tight you will end up with a solid panel that will move quite a bit relative to the plywood. In that case I’d suggest breadboarding the ends of the panel and attaching only the breadboard to the plywood. This could be as easy as cutting bevels on the ends of the panel and the edges of the breadboard and trapping the panel bevel between the plywood and the breadboard bevel.

-- Greg D.

View srzsrz's profile


37 posts in 1893 days

#7 posted 05-03-2013 05:20 AM

There are, as far as I know, two standard ways to build a solid wood door. They both deal with the fact that wooden boards like to expand and contract with changes in humidity and temperature.

Frame and panel

The first is used when the door has to fit precisely in a door frame: frame-and-panel. To understand why that design works, you have to consider first that wooden boards do not expand and contract much lengthwise. The expansion and contraction is mostly perpendicular to the grain, i.e., in the width and thickness of your board. Also, the amount of movement is proportional to the dimension, i.e., a 12” board is going to expand and contract 12x more than a 1” board (approximately).

Now with frame-and-panel construction, the only boards that run the entire width or length of the door are the rails and stiles, which do so lengthwise. Suppose you have a 3’x7’ door with 4” rails and stiles, with the stiles going all the way up and down and the rails terminating at the stiles. That door is going basically to not vary in height at all, and vary in width only as much as an 8” board!

Once you have that dimensionally stable frame, you’re going to make panels by gluing up boards like for a tabletop, and those panels are going to be horribly dimensionally unstable. But you hide that fact by burying their edges in grooves in the sides of the frame. The only effect of the wood movement is how tightly the panels fit inside the frame, but because the thickness is pretty stable, you can make the grooves just the right width and the friction will prevent the boards from rattling around while still allowing for seasonal movement.

Alternatively, you can make the panels out of a dimensionally stable material like plywood, in which case I suppose you can glue them, although it’s still not usually recommended.

Board and batten

All this is complicated, and for simple barn doors, people often use an older trick: board-and-batten construction. The idea with board-and-batten is that you are going to make a door that does expand and contract in one dimension (typically width). And you are just going to live with that. How? Either by not fitting it inside the frame but overlaying it on the outside of the wall, or by undersizing it and just living with the gaps.

Unfortunately, just gluing up a bunch of boards, tongue-and-groove or not, you are going to get not just expansion and contraction but also warping, so you attach battens to prevent that. Problem is that battens are lengthwise dimensionally stable and if glued on counteract the wood movement, causing the boards to instead buckle, probably at the joints, which is not a good thing. So you don’t glue them—you use metal fasteners. Metals fasteners can bend a little, and if you want to be really careful you can always make the holes in the battens a bit oversized to give them extra play. Also, you commonly see elongated hinges that double as metal battens.

Here are some pictures to illustrate.

Shutters with traditional wooden Z-battens

Shutters that overlay on top of the wall, so they don’t need to fit precisely

Shutters that overlay into a very clever rabbet in the window frame

The same tricks apply to barn doors, but it was easier to find these pictures of shutters, and the principle is the same.

Your idea

Gluing a bunch of wooden boards butted up to each other onto a sheet of plywood is not a good idea. The plywood will be dimensionally stable and the boards won’t. That does not end well, I would think. Maybe I’m missing something, but when it comes to wood movement, I would be very careful doing anything that I’ve never seen used, preferably on something very old that still looks decent. At that point, you are left with only first principles to rely on, and the road becomes rather slippery.

The only exception, by the way, to the rule of not gluing boards to a dimensionally stable surface, is if the boards are very thin, like 1/8” at the very most. At that point, we call them “veneer.”

I never quite found out what the mechanics of this are, but it appears to work. I think that what’s happening here is that the veneer is being held fixed width-wise by the glue and thus forced to expand only in thickness, not in width. If the board is thick, then only one side of it is held by the glue, and the other side is not, so the board wants to be wider away from the stable substrate than at the interface with the stable substrate, which causes unpleasantness. If the board is thin enough, then any point in it is pretty close to the substrate and the glue joint successfully holds it in place.

But what about 3-ply lumber-core plywood?

I didn’t think you would really ask this question, but I have, and it intrigues me ;-)

There exists 3-ply lumber-core plywood, and engineered flooring, that works, with plies considerably thicker than 1/8”. There must be considerably stresses inside of it when humidity is far away from manufacturing conditions, but if it’s made carefully enough, the stresses apparently cancel out so as not to cause any warpage, and the glue is strong enough so that the stress doesn’t break the bond. Obviously, something like that is much riskier than a conventional plywood where the stress is limited by each veneer being pretty thin. But again, it appears to be possible.

Have your say...

You must be signed in to reply.

DISCLAIMER: Any posts on LJ are posted by individuals acting in their own right and do not necessarily reflect the views of LJ. LJ will not be held liable for the actions of any user.

Latest Projects | Latest Blog Entries | Latest Forum Topics