Laminating Stile and Rail Stock for Doors

Here's a long discussion on how to lay up stock for architectural door stiles and rails from multiple laminations of wood. February 10, 2009

Question
Occasionally I am asked to construct doors using two pieces of solid lumber glued face to face to make up the stile and rail mat. I face glue them with opposing bows facing each other to help offset bias. Sometimes this works very well (straight mat) and sometimes not (bowed mat) Does anyone have any thoughts about how to improve the average?

Forum Responses
(Architectural Woodworking Forum)
From contributor C:
Have you considered (A) flattening your material on a jointer prior to doing any work on it at all, and (B) selecting relatively straight stock for these incredibly critical parts? You might want to reconsider whether or not you really want to fabricate doors at all.



From contributor J:
Something to consider here is a factor often overlooked by many custom manufacturers. The issue is material waste. I don't want to sound like I'm preaching, but a quality finished product begins with the selection of raw material.

If you are dependent upon a supplier to select and ship your material to you, then you can expect a waste factor that varies between 40-65%, even higher. If you buy similar material frequently from the same supplier you may be able reduce your material waste some, but not completely. On the other hand, if you travel to your supplier and hand pick your raw material, you can reduce your material waste some but the cost of time and travel may offset that waste factor.

Glues, processes and finishes can only help so much. It really all begins with the quality of the initial raw material. Unfortunately, quality raw material is becoming increasingly difficult to obtain at affordable prices, thus the advent of increased use of fingerjointed boards and veneer.



From contributor P:

When I have to fabricate large panel doors I use a laminate for the stiles and rails. I always use odd numbers, either 3 or 5 for the lams. This makes a more balanced glue up and has performed for me flawlessly. Depending on the door slyle, that is if the profile is a cope/stick, I do 3, this way the profile never intersects the glueline. I will also do a 3 ply core, then apply a 1/4in face to each side for super stable door. After the glue cures, I then let them all sit for a few days, joint flat, plane to thickness. Dead flat forever. Two layers can result in a unbalanced product, due to variances in each piece of stock.


From contributor AD:
Contributor P hit the nail on its head.


From contributor K:
Up to now when I have made interior passage doors I have milled 8/4 down in stages, with excellent results. However, I now have to make 8 identical doors. I favor contributor P's approach.


From contributor G:
1,3,5,7,9 laminations up to a 4 inch thick door as contributor P suggests. Double is trouble! We figure 60% waste for solid laminations of door parts - stiles, rails, and panels. Yes, it’s more expensive but much more stable and lasts.


From the original questioner:
Thank you so much for the feedback so far. I am not understanding the technique you are describing, so I'm afraid I must ask for clarification. When you face glue boards together and each of them has, say a 1/4" bow in their length, then each one effectively "cancels" each other’s bow. How would this work if you had three boards? The bow of the two that were the same would overwhelm the one opposing and cause the entire stile to be bowed in the direction of the two. How can an odd number ever provide equal tension? I usually do use stave core/veneer stiles and rails but sometimes my customers do not want to see the two glue lines at the door edge. This way (using two pieces of 5/4" glued face to face) only has one glue line which incidentally matches the same construction method as all the huge lift and slide doors and windows that the owner purchased from Italy. Thank you for your input.


From contributor C:
The trick is to not glue warped faces together. When I make doors, I joint the boards carefully first (using the best, straightest boards I have to start with), then lightly thickness plane them so that I have exterior reference surfaces to use after glueup. If the boards are not straight before glueup, they don't get glued together until they are. If the bow is such that I can't flatten it, and still maintain finished stock thickness then I put that board aside and joint another. You are just asking for trouble trying to balance out the bows in the material. If you don't have a jointer, then you are unlikely to achieve decent results


From contributor P:
Let’s go back and look at why a balanced panel always consists of odd number layers. Also, keep in mind that the glue bonding the layers together acts as a membrane, and for the most part acts as a barrier to moisture exchange. Also, two boards are never exactly the same when it comes to movement, so thus they cannot cancel each other out. They almost act as a bi-metal spring on a carburetor.

When you build the door via three or more lams, the outer faces exchange moisture from their respective faces. This also may be uneven, such as an exterior door with different humidity and temp levels on each side. Finishes will not stop the exchange completely. The moisture cannot go through the glue joint, especially if you use a plastic resin glue. The core then acts as a stabilizing element in between, like plywood. The more plys, the better. More layers mean less moisture can be lost or gained from the outer piece since the volume is now smaller. The core is also now much stiffer too, which will resist the tendency to bow when there is a big humidity differential. The exchange is low on the edges. And all that glue is your friend.

You can buy door cores too, some of which are made up of a dozen or more layers, like a micro-lam. As mentioned earlier, the joiner is the most critical machine to pull this off. Milling lams and letting them equalize before gluing is important. Glue up, equalize, join, plane, and etc. Don’t remove too much from one side either.



From contributor C:
Maybe you should suggest to your client that he also purchase the doors from Italy that he asked you to fabricate. Have you ever laminated any material for radius work? Bow will not matter but springback can. We laminate all of our stile material from 3 pieces or more depending on final thickness, at 9 feet sometimes the stock has a "bow" of more than 1/4" but when we glue up in the clamp rack the stiles come out flat and stay straight.
Outsourcing might be your best option.


From contributor D:
If you run the logic down to the end of the job, past install, you may just realize that some bow/crown is a good thing. How much, where, and when is what really matters.


From contributor J:
I am also guilty of the same, (now apparently flawed), logic of gluing two pieces together to make up interior doors. I made two sets of bypass doors using this construction several years ago, though mine were interior and so far are still in good shape.


From contributor W:
If you pay attention when you pre-hang your doors and make sure that the door hits on the top and bottom of your stop you don’t have a problem but if the door hits the stop in the middle first you now have a warped door and you will be replacing it for free. Even the three plys will bow but most can be prehung to compensate. I have lost count of the times I have walked up to someone running stock through a jointer and ask them hey what are you doing and they say I’m jointing these door parts and I say then why are they not straight and square. I told the one guy the only thing you knew about a jointer is that you spent some time in the same room with one. The jointer is the basis of all quality woodwork.