I have an odd idea, and I’d like input from people who have worked with torsion boxes.
I’m building a kitchen which has slab granite countertops. The usual substrate for the granite is ¾” plywood screwed to the tops of all of the base cabinets. The granite guys are comfortable with the rigidity of this. In this kitchen, there is a breakfast bar formed by overhanging the slab countertop about 12″ over the edge of the base cabinets. The granite guys say that ¾” ply over this overhang is not rigid enough to assure that the stone will not crack. The homeowner rules out legs or corbels to support the ply from underneath. The granite guys suggest welding up a bunch of ¾” square steel tubing to form the substrate. I’m not a welder, I’m a woodworker. How ’bout I make a torsion box with steel skins? The box would be ¾” thick, and have perhaps 1/16″ thick skins. Because the skins cover the whole top and bottom of the box, I think it would be considerably stiffer than the tubing-based suggestion from the granite guys. How would I make the torsion box? I’d laminate 1/16″ sheet steel to 5/8″ plywood, probably using epoxy. Comments?
Replies
First off let me say that I like wacky ideas. In my own work, convention is rare. However, I don't believe that the sandwich you propose will work. (Its not really a torsion box, either.) I would use steel square tube with steel sheet welded to the tubing. I have used this idea for book shelving and they are very stiff, indeed. You can probably get this done by any welding shop since it doesn't have to look pretty.
With that said....your idea may work, I just suspect that it won't. If I wanted to do it, I'd make up a test section and just see how stiff it is. All the conjecture in the wwweb won't get you closer than that.
i dont think the sheet metal would make it any more stiff or rigid. Youre better off taking the suggestion of the granite chaps, head to a fabricator with your specs and they can weld a base up for you.
But trying to epoxy that granite back up or trying to find granite that matches down the road if youre idea doesnt work will be a headache.
I'd check with a couple of other marble guys to see what they use, afterall the cantilevered couter top is fairly common. I don't think steel is necessary or desirable at all. I would build the torsion box to cover the entire section of cabinet and the cantilever as one piece rather that scab on a section like a shelf. This would eliminate the hinge point at the joint. It may be necessary to inset this section into the counter to keep a uniform counter height. The basic principal of cantiliever is that a structure can be unsupported for 1/2 of it's normal span when supported at both ends.
I've seen several articles by Ian Kirby on torsion boxes. Most were excerpts from his books. I think you'll find the formulas you need there. I'd also check out Nidacore for this application, I know Kim Carlton Graves has used it for big conference tables. He used to be a 'regular' here, but no more. He does have a web site.
John O'Connell - JKO Handcrafted Woodworking
Life is tough. It's tougher if you're stupid - John Wayne
John ---
I agree that scabbing on a shelf wouldn't be strong. My thought is the same as yours: cover the entire top of the base cabinets with one sheet of something that is as rigid as I can make it. Between the height of the cabinets and the height of the countertop, I get 3/4" of height to make that sheet. A wood torsion box that high will be no stiffer than solid wood that high. Hence the idea of a torsion box with steel skins. The strength of a torsion box comes mostly from the skins, so a steel-skinned one is stiffer than a wood-skinned one if all the other dimensions are identical.
Jeff ---
A steel-skinned sheet of plywood seems like a torsion box to me. A torsion box is essentially two skins held together by a core of some sort -- honeycomb, rectangular grid, or whatever. If I used a wooden rectangular grid for my core, the assembly might look a lot more like torsion boxes you've seen. Well, what I'm proposing is to fill in the holes in that grid with more wood, and make the core with a sheet of plywood. Making the core solid wood may not make it much stiffer than the wooden grid, but it surely can't hurt.
And, as you say, perhaps the only way to really answer the question -- and to satisfy the granite fabricator -- is to build one, and jump and down on it.
Jamie
Actually the steel skins will increase the stiffness of the panel by a great deal. My strength of materials memory is dim but it seems to me it would increase it by roughly the ratio of the difference in the Modulus of Elasticity of the two materials, ie, (memory is dim) 30E6 for steel and 8 or 12E6 for wood. check out this site for a primer (it's up to you to convert giga-pascals in pounds per sq inch!!)
http://www.geom.umn.edu/education/calc-init/static-beam/material.html
Janation ---
Thanks for the encouragement. This idea may work!
(BTW, I think you may have a typo. If I'm doing my conversion correctly from gigapascals to psi, wood's modulus of elasticity would be in the area of 1.6E6. That is, the increase in stiffness due to the steel is even larger than you say.)
Jamie
You are correct! I said my SoM memory was 50 years dim!!
Following Jeff's suggestion, I built some test pieces. I made three beams. They are all 1" square cross-section, and 48" long. A is solid doug fir. B is steel tube with a wall thickness of .048". C is a solid doug fir core laminated top and bottom with sheet steel that is also .048" thick. (Both the sheet and the tube’s wall are sold as "1/16 inch". Must be some metric quasi-equivalent.) I ran each beam horizontally between supports at the ends, loaded weights at the middle, and measured the deflection.
The result is that the all-wood beam (A) has 15% of the stiffness of the tube (B), and that the steel/wood laminate (C) has 90% of the tube’s stiffness.
That is, if I make my granite substrate with steel tubes, I’d have to pack them together with no gaps between them to make them as stiff as the metal/plywood/metal laminate. And if I make the tube substrate like most guys make it – on 12" centers or so – the laminate would be 10 times stiffer.
Interesting!
How Did you define stiffness as a percentage?
Janation ---
I measured the deflections under identical loading. For instance, under my particular test conditions, the tube deflected .041", and the metal/wood/metal laminate deflected .046". Hence the characterization that the laminate is 90% as stiff.
Jamie
Thought that was what you meant. There is an engineering definition of stiffness that is different based on properties, can't seem to remember the formula now.
a little off topic, but have you sized the cabinet to allow for the extra thickness in the substrate? Also, are you going to bullnose the granite on site or trim it, this is also impacted by the extra thickness.
Jeff --
The base cabinets are 34 1/2" tall, like most base cabinets. The presumption is that the countertop will be 1 1/2" thick, for a countertop height of 36". Stone slabs vary, but many of them are 3/4" thick or so. The usual scheme for these is to cover the base cabinets with 3/4" plywood, so the final height is 36". For most edge designs, a strip of stone is cut off and glued on the underside to form an edge that looks 1 1/2" thick. It can get bull-nosed, ogeed, or many other options. In granite, almost no fabrication happens on-site. (About the only exception is cutting holes for faucets and such.) Granite is machined with big machines, often in a water spray.
Your granite guys don't want their BVD's open in the back.
I made a DR table 72"x40" with a marble top on a box pedestal.
The overhangs are 9.5" on the 72" sides and 22" on the 40" sides. The 9.5" overhangs are not supported. The 22" overhangs
are supported by steel flat stock (2"x1/4") screwed to the sides of the pedestal that extend to within 6" of the ends of the table.
I welded more flat stock at the ends of the supports but 2x stock would do. The table is over 20 years old with no sign warping or cracks.
I think, as has been suggested, that 3/4 ply over the entire top would be all you need.
This forum post is now archived. Commenting has been disabled