NEW! Faster Search Option

Loading

Torsion-box stability?

Jamie_Buxton_'s picture

*
I’m building slab doors for a wardrobe. The doors will be 25”Wx75”H, and 1.2” or so thick. I want the doors to stay flat through time and humidity changes. I can build them as solid-core or hollow-core. To build them as hollow-core, I’d use standard torsion-box construction: plywood skins separated by a grid. To build them as solid-core, I’d laminate ½” ply to ¾” ply. Which construction method will be more stable, and why?

Kirk_Schuly's picture

(post #89742, reply #1 of 15)

*
Jamie,
The torsion box/hollow door construction will work better then the laminated version.

I am not going to give a real technical reason why, bur rather more of a "real life" scenario.

The interior doors of most modern homes are hollow core door style.

Think about all the times you ordered plywood and it showed up at your shop twisted, warped, bowed, curved, bent, not flat.

I don't think it matters what type of plywood or thickness you order, they are always delivered twisted. Sometimes people say things like it has to be stored and handled properly in order for it to remain flat. I think that plywood laminated doors in this application would be a poor choice.

I will also say that when making your grid and skins that you must glue them up flat, if they are made twisted they will be just a bad as the plywood.
Kirk

JosephCoty's picture

(post #89742, reply #2 of 15)

*
Jamie,

I've used torsion box construction for my router table and plan to build a larger version as a carcass assembly table. The grid was also cut from birch plywood and run through the planer on edge to assure uniform widths. A single staple was fastened at the top of each intersection and glue followed by brads secured the top and bottom skins.

My table has maintained flatness similar to a cast surface and have been very pleased.

Plywood is designed to resist forces along the face plane such as stiffening a wall from racking. It also resists deflection better than solid wood of equal thickness but unless reinforcement is designed into the structure it will warp.

This is why either raised panel or hollow construction is preferred to a slab door. Even an inexpensive hollow door with a thin veneer skin sandwiched over a cardboard honeycomb will maintain flatness.

Of course the door wouldn't offer much perpendicular plane strength. But then that is where you as the designer must juggle the traits of each method.

Joe

Kim_Carleton_Graves's picture

(post #89742, reply #3 of 15)

*
Hi Jamie,

I just finished a job very similar to this – four doors 24” by 9 ½ feet veneered in birdseye maple. It was very straight forward. I made them out of a torsion box as was very pleased with the results.

I suggest that you not use plywood for your outer skins. Use ¼” MDF instead with MDF sides. That way you can get the thickness you want. You’ll need to put in solid wood where the hinges will go.

You can make your lattice out of plywood, MDF or even solid wood strips. However this is a lot of work and not as accurate as using paper honeycomb. Kirk Schuly and I have tried both. For our first torsion box (for a veneer bag table) we used plywood stapled together. It was a lot of work – taking hours. However it worked and as far as I know, Kirk is still using that table. Subsequently I discovered paper honeycomb. I wrote an article about it in FWW #152. Using the honeycomb is very fast, easy and cheap. You need a vacuum bag, but it really is the way to go.

Best,

Kim Carleton Graves
Carleton Woodworking

Jamie_Buxton_'s picture

(post #89742, reply #4 of 15)

*
Thank you all for your advice. It is clear that you all favor the torsion box. I understand the value of the torsion box; the reason that I’m sure that my doors will initially be flat is that the platen in my vacuum press is a torsion box that is quite flat. However, I don’t yet understand why the hollow-core door should be more stable than the solid-core approach. The two component layers of plywood in the solid-core apporach would be laminated together on a flat platen, so the door should be flat when it is built. The solid-plywood door that is 1.2” thick would have something like 15 plies. The criss-crossing plies constrain each other so there’s little way for any of them to expand or shrink. And if the plies can’t change dimension, why should the door warp?

Kirk --- Yes, today’s plywood often resembles potato chips. I’ve chalked that up to poor manufacturing. Forty years ago, plywood was rarely as bad as it is now. I’ve seen pictures of plywood being manufactured nowadays in giant presses that hold a stack which must be 50 feet tall. There’s little reason that the middle sheets to be flat.

Kim --- I’ve wondered about MDF skins. If MDF is used for shelving, it sags much more than plywood. That is, the top of the sagging MDF shelf compresses more than the top of a plywood shelf, and/or the bottom of the MDF stretches more than the plywood. A torsion box’s strength comes entirely from the skin; the core contributes almost nothing. In a torsion box under stress, one skin will be in compression, and the other in tension. Because plywood is better than MDF in those strengths, a torsion box with plywood skins should be stiffer than an MDF-skinned one, no?

Jamie

Kim_Carleton_Graves's picture

(post #89742, reply #5 of 15)

*
Hi Jamie,

The strength in a torsion box is NOT in the outside skins. As an architected panel it is in the structure itself – essentially the glue lines rather than in the materials used. I can demonstrate just how resistant to deflection the panels are. If you look at FWW # 152, you’ll see a picture of me on the contributors page seated on something. I’m seated on the panel we built for the article. The panel is cantilevered 42 inches off of my bench and the panel is holding me with my feet off the ground without deflecting. The reason the panels are so strong is that the structure itself dissipates the load throughout the structure – through the glue lines. The trick is to have enough glue lines. For example, a hollow core door is not as strong as a good “torsion box” even though it’s made the same way. The reason is that the honeycomb in a hollow core door is not as dense as the honeycomb made for torsion boxes.

I can think of three reasons, off the top of my head, to use MDF for skins rather than plywood . The first is that MDF is dimensionally stable where plywood isn’t. This is why modern veneering should always be done over MDF rather than plywood. The second is that it’s flat and smooth, which makes for easier construction and finishing (whether with veneer or paint). The third is that MDF mills much more cleanly than plywood.

If you need further information see my article in FWW or the last chapter in my book, “Dining Tables,” now available from The Taunton Press or your favorite internet bookseller. Honeycomb is available from Vacuum Pressing Systems. Darryl Keil, president of VPS, can answer more of your questions on his forum.

Hope this helps.

Regards,

Kim Carleton Graves
Carleton Woodworking

Jamie_Buxton_'s picture

(post #89742, reply #6 of 15)

*
Kim -- the strength of the torsion box comes from the glue lines? That doesn’t make sense to me. Glue lines are only a few mils thick, and can’t have any significant strength. As I see it, the stiffness of the torsion box comes from the skins. The core merely spaces the skins apart, and contributes nothing to the box’s stiffness. The core can be almost anything – cardboard honeycomb, styrofoam, or what have you --- and can be quite flimsy. Structurally, the core assures that the skin’s spacing stays constant, when some stress attempts to bend the box. When the stress is applied, it is trying to lengthen one skin and shorten the other. The skins, although thin, have a big cross-sectional area for those compression and tension stresses, and thus don't shorten or lengthen very much.

Kim_Carleton_Graves's picture

(post #89742, reply #7 of 15)

*
Hi Jamie,

You’re right, you can make the interior out of any kind of honeycomb/lattice. I’ve seen “torsion boxes” made out of plastic, and metal honeycomb as well. I’ve even seen metal honeycomb with stone (real stone) skins. But it is the attachment of the honeycomb to the skin (with glue) that distributes the stress over the entire panel. If you don’t believe me, make up a panel without gluing the honeycomb to the skin. The function of the honeycomb is to distribute the stresses. But it can’t do that without the being physically attached to both skins. So even though the glue line is very small at any one point, it is the addition of all the glue lines that give the panel it’s strength.

Best,

Kim Carleton Graves
Carleton Woodworking

Scott_Braun's picture

(post #89742, reply #8 of 15)

*
Jamie,

People who know both Kim and I have joked that we are "Mr. Knowsit" and "Mr. Hellyoudo". We rarely seem to agree on anything, and I guess that's what makes horse races, right?
But in this case, I thought I'd jump in and say that Kim is absolutely right about why a torsion box is as strong and stable as it is. The strength is maybe not so much the glue itself, but because the glue is there, spreading the load all over the place...

Kim, write this day down on your calendar, circle it in your tv guide, declare a holiday! Hell, it had to happen sometime, right? See you soon, man.

SB

Kim_Carleton_Graves's picture

(post #89742, reply #9 of 15)

*
Yo Scott,

You feeling okay, buddy? I'm a little worried about you. I expect you to be strong, dude. Maybe it was the patriotic Jesus that snapped your strength?

Best, Kim

Jamie_Buxton_'s picture

(post #89742, reply #10 of 15)

*
Mr Knowsit & Mr Hellyoudo --- Gee, now that you guys agree, it is up to me to support that horse race.
I think it may take actual experiments to convince us of the other's position. If I get a chance, I'll build a couple test cases. They'll be identical except that one will have MDF skins and the other will have plywood skins. Come to think of it, maybe I'll make a third variation by doubling the thickness of one set of skins. My prediction is that they'll have different stiffnesses. I think yours is that they'll all be the same, right?

Kirk_Schuly's picture

(post #89742, reply #11 of 15)

*
Jamie,
Try a fourth, change the height/thickness of the core. If you are going to use plywood as a grid change the width of plywood. This will change the deflection rate as well. Think of a steel i beam, the wider the middle part the stronger the i beam.
Kirk

Jamie_Buxton_'s picture

(post #89742, reply #12 of 15)

*
Okay, I built three test sections of torsion box. Except for the skins, the sections are identical. The exterior dimensions are 12”x48”x3”. The core is 1/8” ply (actually .11” thick) in a rectangular grid on 2” centers, with lap joints at the intersections. The skin sets are 1/4" ply, 1/4" MDF, and 1/2" MDF. All skins measure are within .010” of their nominal thickness. The 1/4" ply is 5-ply baltic birch, and the face grain runs the long dimension of the section. There is no edging; the core just continues out to the edges.

I supported the test sections at the ends, loaded them at the middle with 130 pounds, and measured the deflection. The deflections are .059” for the 1/4" MDF, .038” for the 1/2" MDF, and .022” for the 1/4" ply.

I see two things here. First, all else being equal, plywood skins provide more stiffness – in this case two and a half times as much. Second, a thicker skin increases stiffness, but the effect is not linear. In this case, doubling the MDF skin thickness increased the stiffness by only 55%. This bears out the fundamental theory of the hollow-core technique: the material nearer the middle is less important than the material near the face.

Jamie

Don_Stephan's picture

(post #89742, reply #13 of 15)

*
I've been following this thread with great interest. Thanks for taking the time to make up the samples and report your findings. Don

Roger_Martini's picture

(post #89742, reply #14 of 15)

*
You're not going to get a conclusive answer to Kim's point without testing HIS proposition--make
i both
types of torsion boxes with the
i same
skin and see if the degree of difference in deflection is of greater magnitude and absolute scale than for the panels you just made. The key is that there are two different types of structures here. Imagine the simplest torsion box, that of two panels joined at their edges with a box frame. You are right that the deflection strength of this structure depends on the panels' strength in compression and tension. This observation does not carry over to a different structure, where the panels are everywhere joined and forces in the system will be carried differently.

Think of a bicycle wheel. Its strength derives from the tension in the spokes, higher tension equals higher rigidity at the rim. Two wheels may appear identical with the only difference being spoke tension and they will behave very differently. This example gives you a flavour of what is at play here, if not a direct analogy.

singcore's picture

Pre-fabricated sandwich panels (post #89742, reply #15 of 15)

Torsion boxes will certainly make your wardrobe dead flat, while maintaining strength, and reducing weight.

Singcore, based out of McCleary, WA, mass produces torsion boxes and torsion box sandwich panels.  Our products are versatile and the potential uses of our panels are limitless.  Read more about what we do at http://singcore.com.