Torsion boxes

[Cliff Rohrabacher]

I'm about to build one. I've been playing around with a few ideas. The most recent uses OSB as the stringer material constructed exactly like engineered trusses - inset in dadoed 2*2 lumber running the lengths of the long edges.

Then I got to thinking about girders in heavy construction. How a truss member is constructed with smaller internal members set at 45 degree angles that distribute the stresses to prevent deflection.

I am wondering if laying the 8'-long stringers on 45 degrees and in direct contact with each other might not be a great idea.

Like this - VVVVVVV - so that they all relate stress to each other and the skin thereby eliminating the members that on most torsion boxes are running in the 90-Degree direction along the 4-foot lay.

I can see that it might be a tad harder to construct. However I'm thinking that it'd use less material and be maybe even stronger because there are no gaps in the members.

Didn't David Marks eliminate member gaps by cutting and laying each individual level one at a time before moving to the next set of stringers and cross members so as to get as snug a fit as possible?

[Rick Lizek]

I've been making torsion boxes for years based on the original article by Ian Kirby in FWW 25 years ago and it is the finest and simplest technique you can use. No notching or grooving is necessary. The grid is made up of long pieces with short pieces to from the boxes of the grid. "Staples" hold the grid together untill the skins are applied. It's the glued skins that gives the torsion box strength. A thicker core in the center is the key to strength. A 3" torsion box with 1/4" skins and 2-1/2 core is stronger than 3/4" skins and 1-1/2" core. The torsion box is based on the aircraft wing and is concerned with light weight and strength.

There was a recent article by Kirby in American Woodworker which covered the basics of the original article and a bit more. For even more info on torsion box construction see Ken Horners book...
http://www.amazon.com/Woodworkers-Essential-Facts-Formulas-Short-Cuts/dp/1892836211/ref=pd_bxgy_b_img_b/102-4987260-7982562
It gives real formulas on the torsion box. These two sources are all you need to reference from. All the other torsion box info I've seen takes a very simple process and makes it more time consuming than it needs to be, including David Marks. The torsion box is light, stong and very quick to build if you follow Kirby and Horner's lead. Notching, Grooving doesn't add strength, only time, to the torsion box.

[Cliff Rohrabacher]

Thanks


Most of the box models I' have seen require the builder to establish a flat plane upon which to lay the box in construction.

I have been struggling with that one issue. I can see that a construction of the sort that D. Marks uses would require a pre-established flat plane because Marks's system hasn't got anything pulling things together along one of the two axi and the many many little cuts of all those short pieces can cumulatively add up to a substantial error of they don't self cancel. But what about those designs that use interlocking joints on all the internal members ? It seems they ought to force the box to come together in two flat & parallel planes.

Reason I suspect that using interlocking membbers may be a self flattening is because when all the members are brought to bear against the inside skin of the plywood panel they must all conform to the same plane. and since they are all the same geometry they will also force the panels of the top and bottom of the box to conform to the flatness they bring to the party.

ERGO: an interlocking design should require no pre-existing flat plane.


Any one know of an interlocking type Torsion Box that wasn't built using a pre-existing flat reference plane?

[Jim O'Dell]

I just finished one of two boxes Sunday. Check out the Wood Whisperer on line (Marc's also a member here at SMC). Marc did a 2 part podcast on building the torsion box based on David Marks' design. Very good videos.
In my research, and dragging information out of members here and at another forum I'm on, yes a flat reference is critical. What keeps the box flat is the compression and stretching force between the top and bottom. The grid keeps the 2 surfaces apart so that these forces work. Normally the farther apart they are, the more ridgid the structure. For instance, take 2 pieces of what ever material and hold them together. You can get them to bend. Gluing helps, but doesn't stop the forces. If you start with the lower skin already flexed, the grid won't pull the weight up to flat while the glue dries.
My thanks to Charlie Plessum for the education on these boxes. Another tidbit Charlie says is to have both skins the same material. Different types of material shrink and expand at different rates, thus allowing for warpage that changes as the humidity and temperature changes.
I'd pass on angling the gridwork unless your box is going to handle a lot of weight and you need to disperse it better. I'd leave them vertical.
Marc and David both show cutting the gridwork into small individual pieces. I cut half laps in mine, then glued the joints and shot 2 or 3 18 guage brads in each to hold them while they dried. Then glued and bradded the skins on. For a level surface, I used my table saw. It has a 4' extension on the right, and I built a temporary extension over to my benches in construction on the left to get a 10' flat surface to work on. Hope this helps! If you need a wider area than the table saw gives, Marc shows how to get a flat surface to build on using 2 saw horses, some MDF and some straight 2X material. I'm very happy with how the first one came out. Jim.

[Al Willits]

""""""""""
Didn't David Marks eliminate member gaps by cutting and laying each individual level one at a time before moving to the next set of stringers and cross members so as to get as snug a fit as possible?
""""""""

I asked David about that at a seminar he had here, his response is yes you could build one with interlocking cross pieces and the reason he went with the long strips one way, and short ones the other, was for building ease and the TV audience unknown skill level...or something close to that.

The next one I build will be interlocking.

I didn't have much trouble getting a flat surface using his method, takes a bit of time, but I used several saw horses and the leveling sticks to get both sides flat, not sure if the interlocking pieces would keep the surfaces flat or not.
I also used a extra 3/4" sheet of MDF as a base, so I started with three on the bottom, not two like he suggests, don't know if it helped or not.

Al

[Jimmy Newman]

I would have to go through some serious math (or computer modeling, or prototype building) to confirm it, but I believe that the angled web idea would result in a weaker torsion box, particularly on the axis perpendicular to the strips used as webs.

In a torsion box type construction, the stiffness (how much the box will flex under load) is determined almost entirely by the skins (material, separation, and thickness) with very little contribution from the webs.

However, the strength (how much load the box can sustain before being damaged) is determined largely by the webs, not by the skins. This is because the mode of failure in this type of construction is primarily in the web (which will collapse) not in the skins (which will force the web to collapse before they will be damaged).

With the classic square grid type of torsion box, any web member must be broken through its width (the separation of the skins) for the web to collapse. This is the strongest possible geometry for a single strip. With the angled type of torsion box, this is still pretty much true on the axis of the strips, but on the axis perpendicular to the strips (across the v-pattern), a web member must be broken only through its thickness, a very weak geometry. For example - take a pine 1x4. It is pretty easy to break it through the 1" thickness. It's very difficult to break it through the 4" width.

The angled members in a truss are used primarily because they themselves are very strong and stiff in the plane of the truss but have little strength or stiffness in other directions - the opposite case from this torsion box, which is closer to corrugated roofing material - very strong in the direction of the corrugations, but pretty floppy across them.

Using full length strips in both directions with notches for them to fit together should make a perfectly flat assembly surface less critical, although a less than flat assembly surface would make the glue-up harder to get right (which IS critical).

[Cliff Rohrabacher]

I would have to go through some serious math (or computer modeling, or prototype building) to confirm it, but I believe that the angled web idea would result in a weaker torsion box, particularly on the axis perpendicular to the strips used as webs.

I am inclined to agree after thinking on it.


the strength (how much load the box can sustain before being damaged) is determined largely by the webs, not by the skins.
And I don't want to have to use massive skins or fasteners so - - - I guess it's conventional box construction.



Using full length strips in both directions with notches for them to fit together should make a perfectly flat assembly surface less critical, although a less than flat assembly surface would make the glue-up harder to get right (which IS critical).

Yah I think I'm going to try. What's the worst that can happen? I'll waste some time and $$ ??

[Jim O'Dell]

Cliff, one thing I almost made a mistake on is cutting the half laps too tight. I had one set of grids dry fitted and left overnight before I started the assembly. A couple of them curled up a little on the outside edges because they were so tight. I "loosened" them up a little :D and everything stayed flat that way. I don't think it would have been that way if I was using plywood, but I was using the 1/2" MDF like the skins are. I used 5/8" MDF for the outside rim pieces to have a little more beef for getting bumped into. Good luck and be sure to let us know how it goes! Jim.

[Larry Rasmussen]

They had a bunch of used ones from decent looking sellers. I hit my one click button and I'll have one soon. I was just going to start on a shop table using a torsion box base. My floor is so uneven it does put a lot of stress on lighter wheeled bases to have one wheel hanging in the air until I can level it.

Larry