Message Thread:
Torsion Box Doors
2/25/16
My shop has a job involving a fair number of good-size doors (4'-5.5' w. x 9.5' h.) which we were intending to construct as torsion boxes with shop-sawn veneer for rails/stiles and panels to simulate a typical cope/stick door. We have already built a number of radius doors for this job with this technique, and it has worked admirably, though is admittedly time-consuming. There is a desire on the part of those overseeing the job to make the process more expedient by substituting a single solid piece on top of the core (which is Baltic Birch, by the way) to simulate the stiles/rails, rather than building up plywood and veneering, and also by using a single piece of .25" stock for the panels, rather than laminating two .125" pieces, as per the radius doors. I am concerned on both counts that this will compromise the stability of the doors, esp. in the case of the rail/stile pieces, which would be something like .375-.5" thick.
Attached pdf shows how radius doors were built.
I would appreciate any informed opinions on the subject.
Click the link below to download the file included with this post.
Door_Section.pdf
2/26/16 #2: Torsion Box Doors ...
I have always made my radius plan doors in solids. Flat stiles, a bit over thick, sawn rail stock, often with 1/8" veneer facings, and stave built panels. Once assembled the stiles can be handplaned to a slight radius.
"There is a desire on the part of those overseeing the job to make the process more expedient..." Tell me about it. It just is not an expedient process, in my opinion. If we did it every day, we could get better, but we don't have the call for it.
Your method is built up - smaller parts assembled and overlaid to get to the goal. My method with solids is dealing with larger parts and reducing them to get to the same place. Two basic forms of working wood. Both work. Neither can be called expedient, though.
I will say that the solids method is strongly derived from the same way we make all our conventional doors. Mortise and tenon, cope and stick profiles, etc. This makes it remarkably familiar to the maker so it is not as foreign as lots of loose, thin parts to get there. This helps quite a bit.
Conventional build methods also allow the same design elements to be easily incorporated into the door so it can mate well with a run of more conventional doors in the same project.
The photos show one of a pair of 2-14 x 42 x 96 Alder pocket doors. They look good, but the compound curve panel raise was one hair raising and losing set-up and effort.
View higher quality, full size image (1800 X 1350)
View higher quality, full size image (1350 X 1800)
3/1/16 #3: Torsion Box Doors ...
You didn't say what species the veneer was, and do I understand correctly, the center core is Baltic birch?
Anyway, can you lay up the S&R stock and panels separately and assemble in the usual fashion? Might save you some time.
3/3/16 #4: Torsion Box Doors ...
I'll say it again David: You still need to consider teaching a class and in this case it would be radius doors! If you don't have the space maybe team up with Calhoon ?
Great work and great photos; thanks for sharing.
3/3/16 #5: Torsion Box Doors ...
I appreciate the responses thus far, but none speak to the primary question, which is the concern about replacing the built-up plywood and veneer that simulates rails and stiles with a single solid piece of wood on the order of a .5" thick, as well as having .25" single-piece panels. Perhaps I should have been more explicit - but we are done with the radius doors and will now be building a number of flat doors that are 4'-5.5' wide, and they've been specified using this torsion box construction, not solid (the narrower doors will be solid).
Though there are a lot of parts involved, the torsion box allows a large, stable doors to be built with few heroics (can't imagine how much a 2.25" x 5.5' x 9.5' door in solid white oak would weigh) and the results are very good and very predictable. Being able to resaw solid stock and keep color and grain going on a door of this size is really a nice aesthetic feature, too.
I'm concerned that allowing expedience to dictate a less-than-best-practice of laminating thicker parts as if they were veneer with come back to bite us on this.
3/3/16 #6: Torsion Box Doors ...
My point concerned the fact that it is best to try for continuity in process throughout a project or product.
You mention these doors are 'specified' with thick layers of "veneer" so that may ease your concerns, but your question is valid.
In my experience, and as born out here and elsewhere, veneer becomes more like solid wood around 1/4" thick. Exposure, species, cut, substrates, finish and probably more can affect that number quite a bit. As well as EMC, the relative humidity where it will live, and the moisture content once it leaves the shop.
There are no absolutes. I have gotten away with things I should not have, and have been bit where I felt safe.
I can say that what you describe is out of my comfort zone for most situations. I appreciate the weight savings attempt, but a fork truck should be hero along with some hefty straps. I am not sure you are saving that much weight.
3/4/16 #7: Torsion Box Doors ...
Given that the "stiles" appear to be only 4" wide you can probably use a solid buildup there without incident, assuming careful milling and appropriate MC. I would be more suspicious of the 1/4" buildup of the "panels" inviting surface checking due to excess thickness, whether one or two layers of solid wood. Obviously the proof is in the pudding and you have had success with the first iteration. If you are comfortable with that thickness, I don't think using one layer will make any difference. For a far more laborious but less problematic solution you could make the "panels" of edgebanded veneered mdf.
I will say I would not choose baltic birch as the sandwich skins as it is so often squirrely, but it seems to work for you. What are you using for the core lamination? I have used foam, expanded cardboard and Plascore in the past for different applications.
3/4/16 #8: Torsion Box Doors ...
(What I meant to say) The stile/rail buildup is problematic in solid wood and your original layup is technically more correct, but you can probably get away with it if limited to 4" width, depending on the species and all the other considerations David mentions. I would be more concerned with wider sections laminated to a stable core.
3/4/16 #9: Torsion Box Doors ...
Like David said, at that thickness the materials are prone to move like solid stock.
If they want the materials to be that thick, get them to sign off on it. At least you won't have to cover the doors if it fails and cracks
3/8/16 #10: Torsion Box Doors ...
To respond to latest points:
1 - Weight savings is pretty significant, actually. A rough calculation of these doors in solid WO yields 135 lbs. - though we haven't weighed the torsion box versions, they are certainly no more than 80 lbs. With a hinged (i.e. cantilevered) structure, the lesser weight is a real boon for ultimate stability of the door, which also benefits from having the entirety of the "panel" playing a role in the diaphragm, something a solid wood panel does not.
2 - We used Okoume ply for skins - this is a nice product, if any of you are unfamiliar, mostly sold to the boat-building crowd. Available in larger than standard dimensions, a lot of different thicknesses. It is fairly soft, but is a nice flat, consistent panel.
3 - Core is a 6" grid of 1" thick Baltic Birch, parts all notched to snap together off the router. We use a hinge mortising bit w/bearing to square up notches. Important to make the fit just a little loose (.015), otherwise parts start wedging themselves into curves.
4 - Stiles are actually 5.25" wide.
Thank you for the thoughtful responses. I agree that the thicker "veneers" are likely to be problematic, and will be sure they understand it will be their oversight if there are issues with them.
3/8/16 #11: Torsion Box Doors ...
I misunderstood the role of the baltic birch. That sounds like a good layup. I used 6mm ocoume and 4" foam last year to build a pair of 4 'x 7' 6" doors skinned with western red cedar. Nice material to work with.
I like that method of assembling the core grid. Wouldn't you save time cutting the notches deeper or cutting dogbone notches on the router to avoid the squaring operation? I have done the notching on ganged parts with a dado set on a sliding saw, but the cnc router is no doubt the way to go, especially with curved doors.
|