Solid Wood Entry Door Construction

An extended technical discussion of proven methods for constructing large, heavy, durable custom doors using solid sawn lumber. January 18, 2007

We have a 3" walnut door to build and the customer insists that I not use any engineered materials in the construction. A perceived quality difference. I have always used a product we call Timberstrand, without any problems. As such, I have to use stave core and I have to make it in-house because 1) I cannot find anyone who sells it up in ND and 2) I have never seen it in any species other than poplar. He insists on a solid walnut door. Yes, he is willing to pay for it.

My first question is, is there any reason that stave core is made of short pieces? Is it for stability or because it is a way to use drop offs? It is an accepted practice to finger joint the ends together, correct?

Second, after the finger jointed strips are made, are they then just face glued together without any kind of joinery? Or do you use some type of glue joint?

My final question concerns the floating panels. It is a three panel door and I was planning on making the panels out of stave core with a 1/4" veneer, again made in-house. The more I think about it, the more I think it will work, but I still have this voice in the back of my head telling me that I will end up replacing the door next year.

Some side notes:
-It is a south facing door in ND (brutally hot and humid summers, slightly chilly winters).
-Door is, roughly, 40" X 96".
-Finish is oil (which they have promised to religiously maintain).
-I will be using resorcinol glue.

Forum Responses
(Forum Responses)
From contributor K:

You are wise to make a stave core door. Since the customer insists on solid walnut, make your stave core out of walnut. Simply glue up your blanks with Titebond 2, size them, skin them, then mill the parts.

From contributor M:
We manufacture stave core engineered stiles for the door industry using finger joint pine and alder. Most typically for an exterior door, we would use pine core. The reason for the short block is basically stability, as there is no one piece of material that can influence the integrity of the stile, which prevents warping. The initial lead into use of finger joint material for stave core may have been for recovery purposes, but it really creates a very solid stable door component. Our process is to finger joint clear pine block to the desired length by thickness of the door and then we run those blanks through a moulder and s4s to face glue. We would then apply the finish edges at the same time as the core is laid up to width. Once again, that product would go through a moulder to prep for applying face skins. Sadly, I just got rid of a pallet of walnut blocks not more than a week ago simply because I thought no one would ever need finger joint walnut core.

Another option is we do produce what we call billets or slabs of finger joint pine that can be ripped to your desired width and you apply the edges and faces. I think you would be impressed by the weight and density of a pine core, particularly in a thick door.

I would be concerned with simply ripping walnut full length of the door and gluing for a stave core, as you are still at the mercy of any one piece influencing the integrity of your door running end to end.

Please do not confuse ours or anyone else's true finger joint stave core with junk imported radiata billets that do not finger joint, but simply butt end for end blocks as they are laid up, creating large voids with a very unstable material.

From the original questioner:
I appreciate the responses. I should add that we make quite a few doors per year. We have done about 80 this year, with two being entry doors. That may not be a lot to some of you guys, but we are a small, full service millwork and cabinet shop that does only about 3 or 4 really nice houses per year. I would be interested in ordering stave core for future doors, as I don't think there is enough time to have someone make and ship a walnut core to me now.

I got a price from Foremost Wood Products in NY. The price for the core itself was not bad; it was the shipping that killed that idea and I can't find a lumber wholesaler around here willing to order it for me. I had entertained the idea of getting into making it myself, but seemed like a waste of time when Timberstrand has been working so well for us. That may change now that I have to make this core myself anyways.

From contributor J:
We make 78mm thick Euro doors and windows, 57mm thick traditional front doors and 45mm thick interior doors. At the 78mm thickness, a 3 ply lamination of solid wood is used. In our case, the core is 8/4 and the 2 outside plies are 4/4 stock. That particular combination works for us because of the way our glass bead removal is set up on the tooling. Three 5/4s or 4/4 with a 6/4 core could work also, as long as the two outside faces are balanced and all 3 plies the same MC. We prepare the laminations carefully and send the glued up blanks through a straightening S4S machine. This would give your customer a solid walnut door. Walnut is stable and one of the few species that could be built one piece if you could acquire good 3” thick old growth material.

Contributor M makes a good argument for the finger joint method and that is probably the best way to build 45mm thick doors vs. using timber strand or LVL cores.

For panels, 2 back to back floating 5/4 or 6/4 would help with the temperature and humidity changes. 4/4 with a layer of foam board between would insulate the door a little, but then you would not be solid walnut.

From the original questioner:
Thanks immensely for the response. I usually do build our doors with two floating panels with a piece of rigid foam insulation in between, but this customer insists on truly solid walnut. As far as the stile and rails are concerned, you do a three-ply lamination. Is the center of the core quarter sawn? I don't think I have the gonads to do that with a 3" exterior door, but it does give something to think about in regard to our other doors, especially interior.

From contributor M:
We all like a challenge, but I am concerned about your desire to meet the client's demand. I am also located in an area with difficult climate changes and have paid the price for not following construction techniques I knew were right, just to satisfy customers. I feel engineered parts, whether stave, timberstrand, or whatever, are a must. Dark wood with heavy sun exposure and extreme temperature swings sounds like a potential recipe for disaster.

From contributor J:
I am not recommending solid 3 ply construction for your interior doors. There is a lot of difference in dynamics from 1 ¾” to 3”. As I said, contributor M's technique is best for interior doors and good for thicker exterior doors as well. The 3 ply solid construction has been time proven in Europe for doors and windows 57mm to 78mm and now even the new Passivhaus 101 and 92mm thick high insulating windows and doors. With the thicker doors and windows, we usually do all the plies out of the same species. Mostly quartered African mahogany or CVG fir. A lot of it depends on the species. I think flat sawn Honduras mahogany is more stable than the quartered African. German standards call for rift sawn material for 3 ply laminations and also for the stave cores on 45mm interior doors with 3mm skins. In my market, rift material is just not available in a lot of species.

I might be wrong, but I think stave cores came about because of central heating and the crappy fast growth, big ring timber we have nowadays. In Colorado, if you install alder doors in an in-floor heated house, it better be engineered core of some kind. I still think if you can get old growth, tight grain dry lumber, it is possible to build solid. Two days ago we installed a solid 3 ply 57mm mesquite door. 42” X 96”. I feel it was the flattest and most stable door we have ever built, just because mesquite has almost no movement. Like your customer, mine wanted solid mesquite. I would have gone totally solid, but could not get the 2 ¼” thicknesses.

From contributor D:
We build all our doors - exterior and interior - out of solids. We do not use any sort of stave core, nor do we see a need to. It may just be preference (ours and/or the customer's), but a solid door has given us no problems.

We had one 1-3/4 x 96 door in pattern Honduras mahogany warp 3/8" of an inch after 6 months unfinished in the weather, out of about 350 doors per year, for the last 12 years. We do use first rate technique to eliminate problems.

Stave core evolved in mass production as a strategy to deal with the resultant scrap from yielding longs and wides, as the quality of material declined, as well as pressure to lower costs. It does afford a range of comfort that you do not have with solid wood. You have to be at the top of your game to pull it off successfully, long-term.

From contributor I:
There is flip side to that - not everyone wants pattern grade mahogany or old growth fir known for their stability. Rustic cherry, hickory, rustic alder, etc, are the flavors of choice in a lot of areas (ours) that do not lend themselves to solid construction of any type.

We did at one time to do a three ply construction on our interiors, mostly rustic grades, and though our success rate was good, we still had that occasional dominate lam that would take over and result in a bowed stile. It is now engineered all the way.

Just to comment on the implied cost savings going on engineered stave core for the custom door market, our break point is right around a $3.50-4.00 a bdft wood. Meaning we would not see a cost savings unless the value of the wood is in that dollar range, so with some mahoganies, cherry, maple, etc. it is in fact cheaper to be engineered versus solid. With less expensive woods like pine, alder, oak, it is actually more expensive to go engineered. Keep in mind we only do 1-3/4" interiors in our shop, which is the basis for my statement, although I do produce stave core for 1-3/8" doors for other manufacturers.

I appreciate the various feedbacks and hearing how others use different techniques for building their exterior doors - good stuff.

From the original questioner:
Thanks for all of the responses. The door is going to be stave core - that was always the plan. The real purpose of this thread was to see if stave really needed to be a bunch of short pieces, or if it was just done that way to use drop offs. I will use short, finger jointed pieces since no one seems to have ever really tried it using just long strips. My dad used to build a lot of interior doors out of oak many years ago using just 3 layers of flat sawn 4/4. I don't recall there ever being a problem. But, as with everything else, things were different then. He still builds doors, but he says he would never do them like that again.

From contributor J:
I believe the quality door maker must have both stave and solid core construction in their arsenals for all the reasons stated above. To be able to outsource a good stave from a company like contributor M's is a good thing. Last year we did a walnut interior door job using 8/4 walnut a furniture maker had been hoarding in his barn for several years. It was tight grain, dry and worked out with no problem for solid construction. On another job we used walnut purchased from our wholesale supplier and it needed stave construction. Big growth rings and white one face, just some of the reasons. Yeah, I know all about rustic hickory and cherry. I would say most timber coming from my commercial suppliers demands stave construction.

To go solid, you need to spend some money and seek out the specialty and recycled suppliers. A while back on the forum, there was an argument about ripping parts for glued up panels at less than 5 inches and reversing the grain. Same argument, depends on the wood. I just recently visited a museum in San Diego that had works of Sam Maloof and George Nakishima on display. These guys broke all the rules with single slab tables and no grain reversing. Everything I looked at was perfectly flat with no cupping. I think some of the pieces were over 30 years old. The material was all tight grain and was probably well seasoned at the time of construction.

A few days ago I had this same conversation about staves with a local colleague of mine who is very experienced in custom interior door construction. His take is the solid stave like contributor M does is the best for high end work and that timber strand and LVL cores have their use for cost sensitive jobs, but you have to be careful using some materials like maple or CVG fir on the rigid cores.

I believe the finger joint core is a little better, but a non-finger jointed walnut stave would not be a bad thing. Finger jointing is labor intensive for the small operator.

From the original questioner:
I know that finger jointing can be labor intensive, as I am no foreigner to it. But if I am going to take the time to do it, may as well take the time to do it right. As Mr. Rodgers used to say.

From contributor P:
There is quite a lot of information in Architectural Woodwork Institute quality standards regarding this issue. Also to a lesser extent, Woodwork Institute of California. In general, you shouldn’t guarantee a solid hardwood door, except in smaller units in teak or G. mahogany. Three or five-ply construction is massively proven in the field and should be offered unless the client persists, and is willing to assume responsibility for the door.

Woodwork is performed all the time based on client requirements that exceed or avoid standards of our industry; you just express no warrantee, with an explanation, of course. Other issues that void a warrantee are excessive height and/or width, insufficient thickness, excessively narrow or wide stiles and so on. Just because someone can imagine a door made in a certain way does not mean that you must be bold, brave and guarantee it. Non-warranty work is done all the time and should not be thought of as an extreme stance to take.

To assemble door parts (panels included) with face layers/veneers in excess of 3mm is asking for likely differential movement between subassembly components with the faces being too powerful/thick. In general, “engineered construction” is not an issue of economy, it is an issue of preventing the faces of the parts from overpowering a stable core. Composed door parts are actually expensive to produce compared to simply working in the solid. Also, checking may result in the hardwood faces if they are too thick. One of the additional advantages is of course being able to match the parts very well, via veneer or thick veneer, and being able to use highly figured faces that otherwise would explode in service eventually. Woodworkers have been doing this kind of work for hundreds of years.

This discussion also blurs the issues regarding the construction of the stiles, bars, muntins, and rails and the construction of the panels within the door. To perform stave construction on door panels would usually be unsightly (that is, if you are allowing buttmatches). Because you are floating your solid panels, you can compose panels from rippings from the same boards. These rippings may still move within the subassembly, but not destroy your door so long as you do not design your door with really wide panels. If you partially pre-finish your panels, you will help the panels perform as hoped, by the way. In the best work, though, the panels would be engineered. And if there is fielding to the panel or other edge detail, you can develop this by narrow, solid edgebands joined carefully around a stable core.

And lastly, doors have to be protected, and can not be guaranteed in all locations and uses. And unfinished doors can never keep a warranty. And oil finishes are likely to be insufficient to protect exterior installations.

From contributor J:
Contributor P, thanks for taking time for an informative post. I purchased the Quality standards book in 1999. Frankly, I was disappointed in the door information. Especially anything concerning exterior doors. Is better info available to members or for purchase by nonmembers? There is just not much detail on finishing, glues, species, glass technique, weather sealing, hardware, wind loads and moisture management for exterior doors in my copy. The window section goes into a little detail about species, but not much else. One thing in the book I take issue with is Section 1400-G10. They show 2-piece face laminated solid as an acceptable stile construction. In the past we tried that construction and always had problems.

I would like to ask one question about a detail in the book if I may. Section 1400-G11 shows approved panel constructions. Concerning the 3-ply lumber raised panel, are the outside face plies going with the center ply grain direction? Or is the center ply a cross band? 3 ply solid door panels are common in Europe with the center ply a cross band.

It is so true that engineered construction is not a cost saving method and it is hard to warrant any exposed exterior door.

From contributor E:

I too build a lot of doors, and it seems that a large percentage of my interior doors are built using solid rails and stiles, due to the fact that my customers don't want to pay for stave core construction. I also, along with Dave S., don't seem to have too many problems with doors warping or twisting after they leave my shop. I religiously rough rip my stiles and sticker them with a fan blowing air between the layers for at least a couple of weeks before I start building in order to catch the errant board that's too high in MC as well as letting the stock acclimate to my shop conditions. Warping and twisting will only occur when there is an extreme and sudden change in MC to the door, especially to just one face. I'm now about 1/3 of the way through a job where I'm building 252 paint grade doors for some condos and some of the doors have now been out there for 9 months and there haven't been any problems yet. I am sending them out there primed, though. And I have found that if it's going to happen, it would have started by now. Now I'm not trying to say this is the best construction for an interior door, but with some care, you can get good results over the lifetime of the door.

With anything over a 1 3/4 for exterior doors, I always use a stave core construction. But even when I'm gluing up for my stave cores, I'll sticker those for a while and let them balance. Having dry stock will avert a lot of problems. It should also go without saying that I'm insistent on not letting the doors get delivered too early to the job site, either.

From contributor D:
I have played with 3-ply solid raised panels for the last 15 years, and have no problems that I know of. I started this after a few complaints about panel movement, and my fears of gluing in a panel, coupled with the trend at the time for 42" wide doors with one over one panel configurations. A 3-ply panel, net 1-3/4" thick with a 3/8" cross grain center play and equal face plies, all pressed together, worked fine. I have tried 1/4" ply in the centers and that is fine also, but not waterproof long term. I even have a 15 year old sample that has spent all that time in a downspout. The center 1/4" plywood core rotted out, but the glue and mahogany are still there.

I do see the day when we will quietly go to built up stiles. Lumber quality becomes more manageable, and with the glues available, delamination is not the issue it once was. We are in the middle of the cheapest place in the country to build a house, due to very low standards, not labor or similar components. It is typical for the 8,000 s/f McMansions to have hardboard or MDF doors instead of real doors, so it is an uphill battle to get a few more dollars for a better exterior door.

Nearly every homeowner mentions they will only be "in the place for a few years before moving..."
The builders - most of which were drywallers or painters two years ago - all have the utmost disregard for good exterior doors, even if the homeowner values same. They laugh about the "finish when the doors arrive" like it's the funniest thing they heard all week. I had a warranty issue with a warped exterior mahogany door that had been unfinished, south facing, no overhang, 6 months, etc, where the builder said it was my fault and I had to replace the door at no charge. When I was done laughing - since it was the funniest thing I heard all week - he told me he bought mahogany doors so "we can do whatever we want to them and it shouldn't matter since it is the best wood for doors"!

Standards for exterior doors that we can use? That would be too much to even ask for. The big guys do not want us in their corner of the world, so I feel we are locked out of the US standards, despite the fact that I know my doors will deliver superior performance, and do so over the long term.