Constructing large passage doors

Using the right materials, glues and assembly methods. September 2, 2002

Im thinking of expanding my product line to include passage doors. For now, lets assume were talking about eight foot high by three foot wide by 2 1/4 thick interior doors in paint grade.

Is using a laminated core warranted for the above example? If so, is regular yellow glue sufficient for the laminations?

Id like to use loose tenons for the doors. What thickness and length should they be? Should a better glue (i.e. polyurethane glue or West System) be used?

If the panels are going to be made of MDF, is it advisable to glue the panels into the frames for extra stability?

Forum Responses
Eight foot high, three feet wide and 2 1/4" thick? Whose castle are they going in?

From the original questioner:
All of my work is in the Hamptons of Long Island, where those sizes are pretty standard fare.

If your doors are to be paint grade, MDF is the way I would go. But please, don't glue the panels in. Instead, use hard rubber shims on the bottoms of the panels and a headless pin on the top and sides, because the wood parts will expand/contract more than the panels.

Also, for a door that size, if the sticking is not hardwood, I recommend using a paintable veneer over a hardwood stave core on the edge of the outer rails, and at least three 4-1/2" butt hinges with 1-1/4"screws into hardwood. Tenons, top of door 2-1/2" by 5" if the cross rail is 4 to 6"wide. Bottom of door, two on each side if crossrail is 10" wide. One each side for each cross rail under 6" wide. I make my tenons at least 1" thick for 2-1/4" doors.

Yellow glue should be fine for interior doors. Urethane glue is awful hard to control/clean. I have used the Hi-pur hot glue on some of my newer projects and it grabs really good, but that can be a real bugger if you're not quick. It's also urethane and can be messy. I haven't had a failure with the yellow glue, and my doors are still swinging and looking great 6 years now in an office building on State Street, Boston. They are the same dimensions you are working with, only in solid oak with solid wood panels (extremely heavy doors).

From contributor S:
I'd build up stiles and rails from 3 plies of solids, then machine to size. Loose tenons are probably okay, but an integral tenon is standard on this type of work, with cope and stick also being the norm - especially considering your neighborhood. The loose tenons seem to be the thing nowadays, but it is one more part to make, and a critical one.

I would use yellow glue on a 3/4" thick by 3" long tenon with copes. Try to destroy a sample - you will be surprised at how strong it is.

If you do panels of MDF, use the type that machines well. I would not glue in the panels, but machine for a tight fit thickness-wise. A rimmed panel of MDF with an added raise of solid may be the right panel type for the job, if the budget allows it.

If this is your first venture into the world of large doors, make a good sample or two, go slowly and take into account many things you might not consider important with other types of woodwork. Poplar does like to move around, so let it relax at each stage of machining, and keep your humidity and temperature the same as in the finished building.

From the original questioner:
Whats the quickest way of making the core pieces for a fairly large job (say 30 40 doors)?

Is poplar the norm, or would another species be more stable?

Is it necessary to face joint the core pieces before laminating them together?

What kind of clamping schemes do you use for the cores?

Contributor D, can an integral tenon be machined without flipping the piece over for a second pass? Id like to avoid doing that, as it would seem to require that the stock all be dead nuts equal in thickness in order to have identically sized tenons. How do you handle the situation?

From contributor D:
As for the cores, I don't have a lot of experience with rough edged glued cores per se. I learned that everything goes through the joiner first. We will laminate up 3-4 plies to get thickness, and each ply will be faced and edged on a proper joiner, then planed, glued (in a big stack, 6-8 stiles at a time, with pipe and bar clamps every 6 inches or so), then faced and edged again, then planed to final dimensions - lots of planer and joiner time. A vacuum bag will work for the face laminations, but us old dogs just grab the clamps.

Poplar is the norm, but is not real stable in my opinion. Hoadley's book lists it as a big mover. Pine (sugar or Ponderosa is wonderful) or maple are better. Look for some 12/4 pine and forget the laminating.

We use an old single end tenoner to cut square shoulder tenons in one pass. We also have a 9 hp shaper set up to do the copes and tenoning in one pass.

Our planer is off by up to .025 side to side, depending on its mood. We mark all the final thicknessed parts with a face mark, and then run it all face down - plows, tenons, mortises, copes, sticking and assembly. Any unevenness comes up on the "back" and gets leveled in the wide belt on the first pass. This overcomes the variability, since all parts are referenced and machined from the same side.

Don't forget to eat your Wheaties - these things get real heavy.

I made some 1 3/8" passage doors recently (not like the monsters you're making), and had some quartersawn red oak I wanted to use, but it was only 5/4 lumber, and planed out to 1". I clamped two pieces together to get the thickness I needed, plus I figured it would be more stable that way.

I ran the mating sides through the planer, then glued and clamped. The clamps were an idea I saw from somebody's website pictures - 2 pieces of 4" channel cut to however wide you need (I made mine 12" figuring that would be the widest piece I'd need), then drilled near the ends and joined with all thread rod. Stick your wood in between, run the nuts down with an air impact wrench. I put one "clamp" deal about every 15" or so of length of board. I got the channel from a welding shop's scrap pile for scrap metal price (about nothing). I later needed some more, and stopped by an electrical supply house and picked up a 10' length of conduit support bar. It's U shaped fairly thick stuff with holes every couple inches. Just chopsawed it into the lengths I needed and all threaded again.

From contributor J:
Here are a couple more ideas for your cores. We have made a lot of solid wood cores and had the best luck with dry eastern pine. They need to be faced and also pressed flat. We use torsion box culls with clamps from a taylor clamp rack. In the rack these clamps do not line up straight and that will cause the cores not to glue flat. Out of the rack with the culls they are fine. The vacuum press also works well but is slow.

Another idea for cores is 1 1/4" thick Timberstrand OSB with 3/4" edgeband and 1/4" skins for 1 3/4" doors or 1 3/4" thick banded microlams with 1/4" skins for 2 1/4" doors. This is fast and very flat. Really fast if you have a heavy lumber bander.

From the original questioner:
Okay, now for the dumb question. Since clear eastern white pine can be more expensive than the actual face of the door, is it all right to use knotty pine?

Contributor D, tell me if I understand your technique correctly. All horizontal rails have an integral 3 tenon, which is made in one pass. (8 diameter cutter?). All mid stiles have a shorter stub tenon (5/8?). Everything gets patterned with the same cutter, leaving a 5/8 groove. Side stiles get slot mortised to accept the integral tenon.

Finally, do you guys use stacked shaper cutters or is insert tooling the preferred choice?

From contributor D:
I can't help on the pine question - it's not as dumb as you may think.

You have the process right. One of our sticking/cope profiles is made with stacked carbide heads, and makes a 3/4" wide plow, 11/16" deep. This mates nicely with a 3/4" thick mortise/tenon and also creates a haunch as well as the panel plow.

PS - I'm wondering about the moisture content of those things in contributor J's post. Isn't framing lumber up around 12% MC?

From contributor J:
We use K pine and are careful about large knots. I think most any wood that is dry will work. We had trouble with poplar a few years back and have not used it since. We also alternate the bows when gluing the staves. It is important to buy the OSB, Microlams or LVL material from suppliers who store it inside. Like timber, it will pick up moisture if outside.

A lot of this stave technique depends on your equipment. A larger shop in my area face glues 1X material into beams and resaws staves out of these with a large resaw.

Using OSB for the core of a high-end interior door stile sticks in your throat a little... I have to say they stay very straight.

From contributor G:
Structural Composite Lumber (SCL) is rapidly becoming the core of choice for the major manufacturers. I'm not sure if it is available in small quantities for small shops. Often called TimberStrand, it is not OSB or Microlam as I understand it.

From contributor D:
Contributor J, what is the trouble you had with poplar? I quit using it 10 years ago on exterior doors after mushrooms sprouted on a 2 year old exterior door. I also have had the only stile warp I have ever had - after leaving the shop - with poplar. Interior has not been a problem so far, but I am leery and curious as to what other folks' experience is. We make sure the site is right, etc. before delivery, and have had no problems in the last 4-5 years.

You are right about the sticky throat thing. We try to do it as right as practical, and have stayed the course with solids. This builds value to the customer and is still stronger. I like nothing more than to see an MDF door or two on a site with broken stiles at the latch bore or mortise. This just doesn't happen with solid stiles.

From contributor J:
Contributor G, we are using Timberstrand. I may not be correct calling it OSB, but it looks, feels and smells just like the OSB sheeting produced at the LP plant down the road. It may have a little more density. My hardwood dealer in Denver acquires it for me in smaller quantities from a wholesale building supplier. It is 1 1/4" thick and comes in widths from about 7" to 13" X 17' long. Microlams are a different product - LVL (Laminated Veneer Lumber).

From contributor G:
Yep, Timberstrand is the stuff, and you're right - it's made differently from OSB with the door core industry in mind.

From contributor J:
Contributor D, the problem we had with poplar was on an interior door job about 7 years ago. We were building 2 ply stiles (2- 4/4 pieces face glued). The houses here have infloor heat. 75% of the doors twisted, some as much as 1/2" after the job was installed. That is when we started using stave cores. The poplar in these doors was just a little high in MC and the fact that poplar is expansive is the problem. I think poplar skins on a stave core would be fine.

From the original questioner:
Contributor J, a couple of questions regarding Timberstrand.

What happens if the Timberstrand is a bit warped to start off with? Being a manmade product, I imagine it cant be face jointed with HHS knives, so would you just make the skins thicker and face joint after the skins are glued on?

Secondly, how are the hardwood edges applied to the Timberstrand (again, the issue of running the glued up assembly through the planer using HHS knives)? Are they applied leaving a lip and trimmed flush with a lipping planer?

From contributor J:
I just checked some Timberstrand blanks. A lot of them have about 1/8" bow in 96". We seem to get a crown on all our stiles, both Timberstrand and solid. It is usually 1/8" or less. I would not face a stile after the skin is on. You want balanced construction. We also try to sand off the same from each face after door assembly.

Bowed edge strips and skins can also cause a little bow. If your skins are bowed you can oppose the bows.

Our technique for Timberstrand cores is to glue the 13/16" X 1 5/16" edgebands on with a Maranaka thick wood edgebander. Similar to the Hess. Flush trim with a lipping planer, then wide belt sand a little. Lately we are getting close with the bander and just sanding. This method is way faster than solid staves if you are using a jointer and planer. We now have a S4S machine and are going to try a batch of solid staves just to check out times.

The comments below were added after this Forum discussion was archived as a Knowledge Base article (add your comment).

Comment from contributor A:
Here's a little trick I sometimes use. When laying doors flat to clamp them, all the edges recieve different amounts of pressure, causing some joints not to compress the glue (when stacking two or more doors on top of each other). I put dense styrofoam or something similar between the clamps and door edge so that they all recieve almost equadl pressure.