Door Construction and Wood Movement

Experienced pros try to educate a newcomer about the power of wood movement and the necessity of allowing for it in constructing complicated assemblies such as architectural doors. May 12, 2013

I am in the process of replacing some vintage 70's sliding patio doors. I have done all sorts of construction all my life and can't even try to guess the number of doors I have made new framing for and have hung, but I have never made the door itself. The tacky broken 70's style sliding doors will be taken out, and replaced with some rough sawn beams and such to frame the opening and make the casework for the door - that I have no trouble making. It will also have some homemade stained glass windows on both sides of the door.

The door itself is new to me. I have huge stacks of awesome weathered boards. They were 2 x 6ís (1.5 by 5.5 inches roughly). I want to make the door out of them so the boards will be side by side, and create a castle door look. Does anyone have any suggestions as to how to assemble these boards to make the door itself?

Click here for higher quality, full size image

Forum Responses
(Architectural Woodworking Forum)
From contributor O:
If you search the Knowledge Base you will get some good discussions of the basics on these types of doors. Search for "plank doors" and "ladder core" doors. I have recently come across some great history on these types of doors. The earliest English cathedral doors of the 1400's and such were a series of large planks with interlocking stopped tongue and groove arrangements that kept the planks in alignment. Wood battens and strap hardware also did some of that work - primitive, but sensible.

Plank Doors

From contributor G:
We built several of these plank style doors. Although much of the work was done on the CNC the joinery can be duplicated using conventional tools. We milled the planks with a modified tongue and groove. The boards were 1 3/4" x 5". Be sure to joint them first. On the tongue side the tongue was about 3/4" wide and projected about 1/4". The groove side was milled deeper to 3/8 ". This left a cavity of 1/8". We inserted a strip of neoprene 3/4" x 1/4" in the cavity. When the boards were pulled together the neoprene would compress and create a weather seal that could expand and contract with the changes in the wood. We made three or four (depending on the door size) through tenons that were 3/4" x 6" along the milled sides of all the boards. The boards were then laid down on the work table and the through tenons slid through each plank. We used bar clamps to pull all the joints snug. We used a wedge tenon on the end planks. You can also through-dowel the face to hold things together.

From the original questioner:
I'd like to be able to keep the weathered edges also, so jointing and T and G joints would remove the rounded off corners, the weathered edges, small cracks along the edges, etc. These ad even more character to the wood. The wood came from a deck that was taken down. It is cedar that was never sealed or preserved and was in the full weather and sunlight since 1972. It looks ancient, yet it is solid strong wood. Iím just looking for a way to put it all together.

From contributor S:
The crucial thing in the design is to allow for cross grain wood movement. Look up cedar in the Shrinkulator and check the current MC with anticipated installed MC in your area. Look at your wood - quartered or flat sawn? You need to know how much that much wood will want to expand or contract - both from now to completion, and seasonally. If you don't allow for movement then there will be problems. Either you allow the wide board, at 36" width or so, to move at will with a huge frame rabbet (and all sorts of attendant problems), or you allow each plank to move in width independently of each other.

From the original questioner:
The boards are standard 2 by 6 material size (1.5 by 5.5). I originally thought of just gluing them all together side by side, one layer thick. Then I thought of a double layer system almost like brickwork with the joints overlapping, but that would make a super heavy door three inches thick! Then I thought a single layer side by side attached to a 1/2 thick plywood, with something else on the opposite side.

From Carl Hagstrom, Systems Administrator at WOODWEB
The comments regarding combined movement issues are correct - if the door is not constructed to allow for each board to move independently (cross grain movement caused by fluctuations in MC), then the door will behave as one very wide board, and with the exterior side exposed to different EMC then the interior side, movement problems are inevitable.

If you review our Shrinkage Calculator (link below), you can get an idea of how much a door that width would move if constructed as one wide plank. From the picture, it appears to be a 40 inch wide door. If the door were Western Red Cedar, it would move nearly 3/8's of an inch with a 4 point change in moisture content. If it were white oak (I wouldn't want to have to hang it), it would move nearly 5/8 inch in a four point MC change.

Bottom line - make sure to allow the board to expand and contract independently, or you will be facing serious movement issues. And, as mentioned, review our Knowledge Base on this topic - there are examples of how to correctly construct this type of door.

WOODWEB's Wood Shrinkage Calculator

From contributor D:
My suggestion - buy a solid core flat slab, edgeband it with about 1/2" of the rough material you have, then saw the faces off in 1/4" veneers. Glue the veneers onto the door with Excel Structan adhesive. Veneers won't move as bad as solid 6/4 boards, and they'll be bonded to a stable substrate. Plywood doesn't work- ever see a lone sheet of plywood left outside? Warps like a potato chip.

You really should read all the old posts on here about this, as it has been discussed many times. Another option you may consider is actually planing and sanding your weathered material and roughing it up afterwards. Sounds counterproductive, but the wood has suffered some deterioration already and you make be sealing in some dry rot. I'd always prefer joining and sealing freshly cut surfaces myself. Exposing new wood like this would surely help its longevity.

From the original questioner:
I am very grateful of all the help that has been suggested so far. Many terms I'm not familiar with, but thatís ok. I was wondering what it would be like stability-wise if I were to just nail and glue it up as seen in the examples here with the grain pattern arranged as seen. Or would a different pattern be better for stability?

Click here for higher quality, full size image

From contributor D:
I think most door-builders on here would agree these plank doors are deceptively simple, when in fact they are much more problematic than stile and rail construction. I've built several thousand doors and can say matter of factly that your sketch will not work. Warp is more than just ugly- you won't be able to get a weathertight seal, and it affects greatly the operational performance.

From the original questioner:
Sub it out is not really an option. I have the tools and the time. Just need the assistance to tell me the best way. I'm disabled and on a fixed income of just over $800 a month, so obviously little is left over for anything, but I do have time. Even the framing out is salvaged wood from an above ground planter made from 4 x 6's and they also look awesome.

The way I laid out the grain - is that the reason it would not work? What is happening? This is very seasoned wood and it has been outside in the form of a deck without any type of finish at all since 1972. It is solid as if it was just installed, but with the awesome surface weathered look.

From contributor S:
Contributor D is correct. These look much simpler than they are. You really need to grasp the idea of cross grain wood movement and design for it, or they are doomed. You have received very good suggestions, but your last post indicates you are ignoring the movement issue. Did you do the research? Do you understand? As Carl illustrated, the door will move quite a bit - grow in width - with just a small increase in MC. You cannot build it the way you propose, and have anything like success. With these very basic questions unaddressed, I suggest you think about having a professional - with experience in these types of doors - make these for you. You haven't even made it to latches, hinges, weatherstrip, sills, glass setting and finishes yet.

From the original questioner:

Yes I did have to grasp the idea. I do understand now. What wasn't clear was allowing it all to move, yet be a useable door. The allowing it all to be able to move and yet be a door that works was baffling me. I am still a little concerned on one or two aspects of this.

Right now and in the future I won't have the money for a professional. Plus I have been working with wood all my life. It is just this door that was throwing me for a loop. I've even built homes from demo materials after a tornado went through. I had the demo companies dump the ruined homes in the back 40 and Iíd pick through it and make a 1200 square foot house for nothing other than fasteners and time.

From contributor D:
Just trying to steer you away from a known headache. I'm not ashamed to admit we had many callbacks in the early years with these kinds of doors. In our shop that door would involve hundreds of tiny pieces in veneered stave-cores, each one tongue and grooved and through mortised top bottom and center. Massive internal horizontal tenons, vertical grain 6/4 x 6" would be in those mortises, fit tight in thickness, loose in width.

Others employ a ladder frame, planked shiplap style. We stop the v-groove joints at the top edge of the external side, keeping them flat for the last half inch so they seal tight against weatherstripping. No need to get defensive - my first set was simply tongue and groove vertical grain edge glued planks and they're still flat, but that's an exception, not the rule. As I understand it, those doors get stripped and repainted yearly.

From contributor E:
I've placed a double row of biscuits down the length of the boards 4 to 6 inches on center and glued up the panel and itís worked pretty well. You can glue up pairs to start with, then the whole assembly. Normally I would joint all of the edges first, but that wonít work if you're trying to preserve the entire face and edge of your boards. You could also spline the boards together with a dado blade and some Baltic birch plywood strips made on the tablesaw. I use the 1/2" Baltic birch because it's a 7 or 9 ply - very stable plywood. T and G is the other option, but you lose wood. Your steel straps may help keep this door flat, however make sure your fasteners allow for wood movement.

From contributor D:
I will stress that you can get lucky sometimes. There was a guy on here a while back who welded a ladder core out of steel angle and planked it with oak. The board's warp power was such that it bent the steel and the door wouldn't close right, opening up in the dead of winter. I prefer staved core and veneers for each "plank." We have some fun distressing them with wire wheels mounted on small angle grinders to mimic an aged look, if the customer desires it. Notice the small moisture change mentioned in the posts above, and then imagine the raw cedar the OP intends to use after a good soaking downpour. Warp city.