Plank Doors with Stave Cores

Advice on stave-core construction for a plank door. March 26, 2010

I've been asked to build an arched top plank exterior door (complete with V-grooves). I was planning on using stave cores with 1/4" skins (all Douglas fir or Spanish cedar - to be painted) to make 6" planks and edge glue everything together. Will it work?

My Knowledge Base search yielded concerns of expansion and contraction across the width and cupping/bowing. I assume that the core will tame both, but please chime in if it would be better to construct it differently.

Forum Responses
(Architectural Woodworking Forum)
From contributor A:
What is your plan for V-grooves? They are typically a minimum of 1/4" depth. Because of that you might have to bump up to 3/8" veneers.

From the original questioner:
I was planning on keeping the V-grooves just shy of 1/4" in order to keep the skins as thin as possible. Since my plan would be to essentially make engineered planks, I would chamfer the edges to create the groove and edge glue everything together.

From contributor V:
As I see it, frame and panel doors evolved after plank doors as a way to control movement. Even in English doors of the 1500's, woodworkers were letting in horizontal battens on one side of plank doors to stiffen and flatten large church doors. This later evolved to a core frame, with planks laid on either side.

A bunch of planks, all with the grain running one direction, glued up for thickness or not, will not stay flat and will definitely not be fit to hang with all the lateral movement. This is why frame and panel evolved - to limit wood movement to the stiles only.

Search out someone experienced to make the door if you do not have experience. These are not for the faint of heart; there is a whole lot that can go wrong.

Exterior Plank Doors

From contributor D:
I've built lots of these over the years and have learned the hard way how to minimize failure. In my opinion, staved core construction is a must. Also, a spline joint between each plank is necessary, or at least a finger joint. I won't argue with all the butt joint advocates because I know what I know, and it was learned through building thousands of custom doors, not by reading the same dogma we've all been given by the glue companies.

Another thing that helps is fitting in the largest horizontal plug across the bottom that you can. I machine out a 3" x 1.5" groove and glue in a plug. This minimizes moisture uptake from exposed end grain, and ties the planks together where they are most likely to delaminate, at the ends. If it is a wide door, say 42" or more, I take a different approach. I dado in large sections, say 5" x 1/2" deep, across the width of a fully glued up slab and plug them before veneering, which acts like a hidden rail or crossbuck.

If you don't have a large press, the way I veneered long ago was with "go" sticks. It is a Japanese method of applying down pressure on a worktable by cutting sticks just a bit longer than the distance to the ceiling and jamming them between the ceiling and your cauls. With this method you can apply tremendous pressure without fancy equipment.

All these steps need time to acclimate, as each glue-up introduces moisture. For example, it is a good practice to build your engineered planks slightly oversized and let them completely cure, as you'll likely have a tiny bit of warp to joint out. On the performance side of things, I stop the v-grooves on the interior side a half inch from the edge of the top. It may look funky, but you won't see it when the door is closed, and you will get a tight seal against the weather stripping. I also swab across the bottom and in hardware mortises with a sealer, which was discussed in an earlier thread on screen doors. Some may think much of this stuff is overkill, but I've found that although these doors are simplistic in appearance, they are actually more of a challenge to build and perform successfully than designs with stiles, rails, and floating panels.

From contributor D:
I just noticed I wrote interior side for the v-groove stopping, which would be correct for an outswing door, typical for businesses. Stopping them on the exterior side would be for inswing doors, typical for residences.

From the original questioner:
Thanks for the link. I missed that one, but have read about ladder cores and understand their advantages. I hadn't read anything before about making a plank door, using stave core planks instead of solid lumber planks. I guess I could incorporate a bottom breadboard end as suggested to help keep everything flat.

I'm also taking the "I don't normally build doors" risk into consideration. Right now, I'm in the same boat as a lot of guys by accepting almost any job in order to keep food on the table. I wish it wasn't like this, but who knows - maybe I'll become a door maker.

From contributor G:
I have had excellent results using engineered stiles for the planks. I use 1/4" plywood splines along the joining edges and I also bore a 7/8" hole approximately 18" on center through each plank and partially through the edge planks. Then, as I glue the planks together, I insert a 3/4" steel dowel that is about 3" shorter than the door width and then I fill the 7/8" hole with epoxy or Gorilla Glue. These dowels really help to keep the doors very flat, and as I go back and look at the doors I have built years ago, I am happy to see they stay flat. I bore the 7/8" hole on my horizontal boring machine (sometimes depending on the plank width, I bore from each side).

From contributor V:
When you say "engineered" stave, are you referring to a part that has grain orientations running at 90 degrees to each other? A thick piece of plywood, so to speak? 5 ply or 7 ply seems like it would work. This would be different from the usual engineered core of a door stile that would be all grain going in one direction, or no grain (manmade board core) with faces all one direction.

I have modeled our method after the method that I learned a long time ago in a shop that had been making plank doors since the early 20th century. I was taught to build what I see when duplicating historical work, and I saw planks on either side of a frame. Never have had a problem.

I can see how other methods will work if the grain is alternated.

From contributor D:
In a way, yes, I alternate the grain with a few dados 1/2" deep across the width, but no, not in the planks themselves, which are just like butcher block construction. Years ago I'd built some with plywood cores that strangely bowed and had to be rebuilt. I don't doubt some people have success with plywood, but in my experience it has not been the case. In that case years ago, I used 1" ply and planked both sides with 3/8" tongue and groove, end capped the perimeter and glued everything with epoxy. I was stunned to discover within six months all the doors on this job bowed, some with twist, some cupped. I had no explanation. We build a lot of doors, but only maybe 1-10 a month are plank type. The methods I use now are very successful and have over the past 7-8 years had no warp. Only time will tell, I guess, but usually if a door warps it is within the first climate change, so I feel confident these are pretty stable.

From the original questioner:
What thickness skin/veneer are you using with your method?

From contributor D:
Veneer thickness varies with the order. Like contributor A said, you'll need to accommodate the thickness of the grooves. The max I ever used was on a job last year that had about 50 doors, ten foot tall, and the customer wanted me to use a 3/8" bead point instead of a V. I used 1/2" thick skins on 2 1/4" thick overall slabs. I don't have any logic to back up my preference for thick veneers, minimum of 1/8", even though I've seen many large manufacturers use paper-thin veneers. Also, there have been so many ordered distressed or antiqued that I had little choice in the matter, and as a matter of habit just skin everything with at least 1/8".

From contributor W:
Contributor V, once you have the ladder made, are you gluing the faces to the ladder? With what type of adhesive? How large are the doors you have made with this method? We are looking at two jobs, one in CVF fir and one in sapele, with this type of door. Some of the doors on the sapele job are as large as 4-6 X 12-0. We have often made doors this large with an internal steel ladder frame, but the more typical doors are 3-0 X 9-6.

From contributor V:
We have gone to 48" x 120" or so, all with a wood ladder frame, at 1-1/2" thick (3" total), with long tenons. I have wondered how big this could go and when I would consider a steel core.

We press the faces in one go in a vacuum bag, with an exterior rated glue for exterior doors. We did use urethane for its open time, but are rethinking that in light of the current reports. We use epoxy if the door has any exposure.

Don't forget to use rigid foam in the voids of the ladder frame. I'd like to see your frame details for the 12' tall door(s)!