Construction of Exterior Doors

Woodworkers discuss stave-core and ladder-core methods for building a veneered exterior door. June 24, 2007

I am building an exterior door with the look of bead board on both faces. Does anyone
have experience with making a stable door like this? I initially thought of using exterior grade 3/4" MDF with 1/2" veneer on both faces and solid wood glued to all edges of the MDF. Are there other less expensive materials suitable for the core?

Forum Responses
(Architectural Woodworking Forum)
From contributor E:
I would use a stave core. You can buy it in a 4x8 sheet. They come in different thicknesses. I'm making doors right now with the same material. It's strips of pine glued up side to side and finger jointed in length. I'm gluing 3/4" oak to the edge and 1/4" skins on the face. I have also glued up my own stave core. Works well and stays nice and straight. Water always finds its way to MDF in an outside condition.

From the original questioner:
Thanks. I've never heard of anyone supplying stave core in sheet form. Who makes it?

From contributor E:
I will have to get back to you tomorrow on where my shop got the stave core. I can tell you that I ripped the whole sheet to 6" rips about two weeks ago and glued the edges on (have not glued the skins on yet). They are straight as an arrow. I have been worried because I'm in Baltimore and the weather yesterday was 70 degrees and today was in the 40's. I have seen the stave core also in poplar, but pine might be better suited for exterior.

From contributor S:
I assume you are talking about what we would call a plank style door, where both faces are comprised of t&g boards only. As opposed to a frame and panel door. If so, you should realize that the two faces will move with changes in RH (humidity), and the core's function is twofold: to hold the faces rigidly to prevent sag, and to control expansion across the width.

The MDF scenario succeeds at both goals, but is miserable in exterior situations. MDF also has little inherent strength (likes to sag), so should be avoided.

The solid wood (stave, fj, whatever) will expand across the width of the door, along with your faces, and the door will not fit the frame after a short time in service. There is also nothing to keep the door flat.

The solution is to make a mortise and tenon "ladder core" that will give both rigidity and flatness, and control expansion (actually, limit it to two 4"-5" stiles as in conventional frame and panel doors). This ladder core should be the same species as the faces, and the several horizontal rails should be joined with longish tenons to the stiles. The voids between the faces and rail/stiles should be filled with rigid foam.

We have built many, many doors this way, and the method is based on 60 to 120 year old doors I have examined. We have never had one of ours twist on us, but we have replaced several that were built differently and twisted.

From the original questioner:
That sounds like an excellent approach. I think I will go with that.

From contributor T:
I like your method, but I have a stave core question. I've made my own stave core before and it turns out to be somewhat like true butcher-block in that the faces are glued together. Never bought the commercial stuff, but it sure sounds good. In essence, any growth across the door's width would be in the core's thickness. Wouldn't this mean that it would be more stable than, say, a door with two 6 inch stiles of solid wood?

From contributor D:
"Stave core door" would - in my book - refer to a frame and panel door with stiles and perhaps even rails made up of a different species/processed wood/ process than simple solid wood (including a 2 or 3 ply lamination, all same species). We refer to our plank style doors as having a ladder core. Though all this naming stuff could be viewed as arbitrary, it is important to know whether what we are talking about is the same thing or not.

Therefore, a core that has all the grain running vertically will most definitely expand/contract across the width and the thickness. If the core is 3/4", combined with the faces, each at 3/4", the expansion is negligible in the 2-1/4" total. However, at 36", the expansion will be very obvious and unavoidable. Using quarter sawn wood will help, but if you look at the shrinkulator and plug in some conservative values, you can see you'll still be courting disaster. Cross grain wood is needed to limit that expansion. The cross grain also gives torsional rigidity and helps keep everything flat. Much like a plywood.

Incidentally, the veneered woodwork of a hundred years ago was solid frame and flush panel with heavy - 1/16" thick - cross bands and face veneers on both sides. These were mostly done with mortise and tenon, plowed for panels assembled with hide glue, and then edge profiled. One of the best surviving examples is the old Singer treadle sewing machine cabinets. They survive due to their durability.