Mixed Wood Species in an Exterior Door

A woodworker describes a failure in an exterior door built out of Walnut and Oak on two different faces. But was it the wood, or the construction method, that caused the problem? August 15, 2012

Question
I am trying to find information on mixing woods for projects. Say I wanted to build an exterior entry door that was a stile and rail construction (mortise and tenon) with quartered white oak on one side and black walnut on the other. The door would be 2" thick, glued with Tite Bond exterior glue, and laminated together before machining. I know from experience that this door failed within one year of installation, splitting down the seam and sounding like a shotgun blast. What characteristics of each wood do I need to compare to better design a multi wood door so that this does not happen again?

Forum Responses
(Architectural Woodworking Forum)
From contributor D:
We've built a few mixed species doors, but I know for sure not two species sandwiched. Just for accent we've made use of different species for raised moulding, square pegs in tenons, contrasting colored raised panels, glass stops, etc. If that order were presented to me, I'd stave core the rails as well as the stiles, and veneer them. Raised panels we make separate and back to back already anyway, so there'd be no change in construction there. Shotgun blast makes me think two wide flatsawn boards cupping apart, which could happen whether you mixed species or not.

Clues like whether it split in the glueline and gap shrinking back after the door dries out will further tell you how it failed. More importantly, swab some sealer in panel grooves before you assemble a fielded panel door. Any sheeting water will find its way into that groove and soak into a rail, and a good waterproof finish on everything else will actually inhibit water from evaporating. It's a water trap.



From contributor M:
About 30 years ago as a lumber salesman I came across a cabinetmaker working on a mansion. He had his shop set up in the two car garage in the new building. He was making all the doors for the house as you described, with different species on each face (cherry on one side, walnut on the other; qtr. white oak one side, walnut on the other, etc.). He was making the doors as two separate 3/4 or 1" units and then laminated them together. By doing this he was also able to use different raised panel configurations on each face. He even installed Kevlar in between the panels on the front entrance door (it was a rough neighborhood).


From contributor O:
The different types of wood move at different rates and make this kind of construction fail. On an exterior application, this type of construction is even more likely to fail. Either make the door with veneer faces or make a solid door and stain the oak to match the walnut.



From contributor F:
I think for an exterior door I would want to either stick with a very stable wood, (like mahogany for instance), or use a stave core and veneer the faces. I don't think gluing up slabs of solid stock in different species is a good idea in this particular application.

As far as characteristics to look for I think movement would be an important one. If you try to glue up a very stable wood to one that moves a lot, I could see that being problematic.



From contributor G:
I think stave core for the styles and rails is the way to go. 1/8 to 3/16" face veneers with solid on the edges thick enough for the stick mould. Using two panels back to back in the door is something I always do for exterior doors with 40 lb roofing felt between to allow the inside and outside of the panels to move separately. It sounds like the glue failed in your door, otherwise it should have warped due to dissimilar expansion and contraction rates not coming apart.


From contributor L:
We make our doors like Contributor G but use a piece of stiff plastic between the inner and outer panel. Face glued frames of different wood seem likely to fail!


From Gene Wengert, forum technical advisor:
If there is minimal MC change, the effect of different species will be minimal. Make sure to get the MC of the wood when manufacturing equal to the expected MC in use. All too often, the wood is too wet for the environment. Typically, an interior will range from 6-8% MC, so 6.5 to 7.0% MC is perfect. You will need your own mc meter to double check the moisture of the wood you are using.


From contributor L:
This application is for an exterior door so one side is likely to face a totally different equilibrium moisture level than the other and to reverse with changes in season. The interior face will probably not face those issues.