Torsion box question

[Jon Endres]

Pardon if this question has been asked before.

I have watched a bunch of videos on YouTube recently about building a torsion box workbench top. Various kinds of tops, various materials. It seems like most of them are more-or-less slapped together with little consideration for the concept that a torsion box is meant to be FLAT. Dead-flat without sag, bow, warp or any other malaise which would throw off subsequent construction. It should be a reference surface. The concept is there, the capability to produce such an end product is available, but the method leaves me shaking my head and wondering how it's possible. I never see any straightedges, winding sticks, or even a good sightline.

So the question - what does one do to make a torsion box flat? It seems to be an offshoot of the adage that you need a workbench to build a workbench; similarly you should have to use a flat reference surface to build a flat reference surface, no? I can't trust a concrete floor to be flat; I know for certain it isn't. I can't trust a 27" x 40" tablesaw top when my torsion box is four times that size and will sag just from it's own weight.

Or, is it just not worth agonizing over and I should just hope for "flat enough" and be done with it?

[Charles Taylor]

David Marks built a torsion box in one of his WoodWorks episodes and took the time to make sure his temporary work surface was flat. If I get a chance this evening, I'll try to figure out which episode it was.

EDIT: The Google made it easy. Season 4, episode 409.

[Gerry Grzadzinski]

I think the Wood Whisperer video also shows setting up a flat surface.

[Robert Engel]

I set up a couple sawhorses and adjusted with winding sticks to make parallel, then add a few 2x4's and you have a flat surface to work off of.
Put the top sheet on them upside down and proceed.

[glenn bradley]

I have watched a bunch of videos on YouTube recently about building a torsion box workbench top. Various kinds of tops, various materials. It seems like most of them are more-or-less slapped together with little consideration for the concept that a torsion box is meant to be FLAT. Dead-flat without sag, bow, warp or any other malaise which would throw off subsequent construction.

Seems you have been looking into unreliable places if they are not first and foremost going through the reference surface steps on which all else will be built. Delete those bookmarks, if any as they will only take you down a road to a poor result.

The Wood Whisperer, American Woodworker and others have good examples. These examples include methods for establishing a flat reference to work from. Generally supports such as saw horses are topped with true, jointed ribs that can be evened up via winding sticks or long straight edges. True, if you do not have something that is long and true or are proficient with winding sticks you are ill prepared to begin.

A long builder's level can help and comes in handy for other things. Winding sticks can be shop made and do a fine job of confirming or leading you to a flat platform (or row of ribs on saw horses) from which to begin. If this will just be for assembly, you're all set. If this will double as a work surface remember to add solid areas for dog holes or vice mounting options.

[Jamie Buxton]

You're right: you have to pay attention to make a flat torsion box. Here's how I've done it without a prior flat surface...

The torsion box consists of two skins on the outside, and a spacer thing in the middle. The skins might be quarter-inch plywood. There are many ways to make a spacer grid, but I do it with plywood. Eighth-inch plywood works just fine. Rip the plywood into straight strips, perhaps 5-6 inches wide. Pay attention to make the strips not curved, and not wedge-shaped. When I'm ripping dozens of strips off a full sheet of plywood, I've had the cut line wander away from ideal. You're going to assemble the strips into a grid structure. Every strip of plywood is going to run from one end of the torsion box to the other, or one edge of the torsion box to the other. To get that to happen, cut slots half-way through each strip at each junction. The "joint" could be called a half-lap, although it isn't really a joint. It is just clearance so the strips can get past each other. I gang-cut the slots. I wrap bundle of strips with saran wrap so they all stay together, and then cut the slots, cutting through the saran wrap where necessary. The great thing about gang-cutting is that all the slots through all the strips are aligned, so the crossing strips can go right through. Assemble the strips into the grid. It will be quite floppy, and that's okay. Put one skin down on something flat-ish. Maybe that flat-ish thing is a sheet of 3/4 ply on several sawhorses, shimmed a bit if your shop floor is really out of whack. When I say flat-ish, it only needs to be flat enough that when you put the spacer grid down on it, you can get the grid to conform to the bottom skin without too much trouble. Maybe you need some weights at various places around the grid, or maybe just another sheet of 3/4 plywood. Mark where you need glue on the bottom skin, and pick up the spacer grid. Apply glue, put the grid back down, put on those weights, and wait for the glue to dry. After the glue dries, you'll have a structure with one skin plus the spacer grid. You'll find that it doesn't cup much, but it does twist very easily. You must remove any twist before you glue on the second skin. Stone masons have a trick for seeing twist: crossed strings. Run a string from one corner of the assembly to the opposite corner, and a crossing string from the other two corners. I screw one end of each string, and add a weight to the other end to keep it in tension. Near the assembly corners, add shims under the upper string. The shims are the thickness of the lower string. If the assembly is untwisted, the strings will just kiss where they cross. Shim an assembly corner until they kiss. Remove the strings and glue on the second skin. Done.

[Robert Engel]

I think this ^^ is over complicating the construction.
The error of not having a dead flat surface upon which to begin the assembly will transmit though the entire project, and therefore requires the use of strings and tweaking at the end. Totally unnecessary.

It is simply a matter of glue the strips on and brad nail the cross pieces (1/2 laps are not necessary).

Allow the bottom and torsion assembly glue to dry, then turn over and apply top the same way.

IMO MDF is a better material for the bottom and/or top.

On mine, I used 1/2" MDF for bottom and 3/4 for the top.

To Jon I would say check Marc Spagnolo's video on WoodWhisperer. Its very similar to the way I described above.

[Eric Schmid]

I set up a couple sawhorses and adjusted with winding sticks to make parallel, then add a few 2x4's and you have a flat surface to work off of.
Put the top sheet on them upside down and proceed.

This is how I built the two torsion beams I have. I joined the 2x4's first, aligned them into a single plane and set the 3/4" MDF bottom on top to start construction.

Start with a flat reference.

My set make a portable workbench that was put into service 8 or 9 years ago. 3/4" MDF top and bottom, 3/4" plywood skins. No glue on mine, just screws. In and out of the truck, on and off jobsites and still flat.

[Peter Aeschliman]

Here's the Wood Whisperer video, fast forwarded to the part you should be interested in:


https://youtu.be/1-Hbsou6cWo?t=6m10s

[Ole Anderson]

I was recalling the David Marks video also.

[Harry Holzke]

I recently finished a torsion box work table primarily following Spagnolo's video. The reference surface is the key and does take some time to set up but well worth the effort. I got everything perfectly level then hot glued everything together and to the floor. I opted to use half lap joints which I think is easier than a pile of pieces to put together. First thing I built was the bottom of the table on my dead flat top - should have done this long ago.

339149339150

[Jon Endres]

Thanks, guys, for all of the helpful advice. I watched the Wood Whisperer video on the torsion box construction and it includes the important information about setting up your reference surface. I have watched several of his videos but never that one.

I've also watched several videos on the "Paulk" style workbench, which seems to be a modified torsion box - I like the concept because I have a relatively small space to work in and I like the idea of storage underneath the top. "Multi-tasking" as my wife would say. I now see how I can use the same construction techniques to build one of those type surfaces just as flat (or close, considering the materials).

[Phil Stone]

Thanks, guys, for all of the helpful advice. I watched the Wood Whisperer video on the torsion box construction and it includes the important information about setting up your reference surface. I have watched several of his videos but never that one.

I've also watched several videos on the "Paulk" style workbench, which seems to be a modified torsion box - I like the concept because I have a relatively small space to work in and I like the idea of storage underneath the top. "Multi-tasking" as my wife would say. I now see how I can use the same construction techniques to build one of those type surfaces just as flat (or close, considering the materials).


While I think a torsion box would make a usable workbench top, it's not going to be very massive. Weight is a good thing in a workbench, especially for hand-tool work -- it absorbs the punishment and stays put. If hand tool working is not your thing, then it's not so critical. But if you want to do some planing, chiseling and/or sawing, you should consider the more traditional laminated top of three inches or more in thickness. Believe me, you'll appreciate that mass. Pick your wood well, and orient the grain consistently during glue-up, and you'll have a surface you can plane perfectly flat as many times as you need to over the life of the bench.

[Jon Endres]

This is going to be more of an assembly table/outfeed table/light-duty workbench. Primarily power-tool use, and it will also be on casters so I can move it around my shop. I would like to build a heavy hand-tool bench sometime in the future, there's a stash of 8/4 white oak waiting for it.

[Chris Fournier]

I built a 4' X 8' torsion box top assembly table. I carefully welded up a square fame and then simply built the top on the frame. No big deal but a very sturdy and flat surface that is an essential tool in my shop. Great storage space underneath too.

[Chris Parks]

I am about to walk out to the workshop and start a torsion box 3 metres long with no flat reference surface. The first thing I am going to do is make four "I" beams 1500mm long out of MDF with the vertical web on all of them cut from at the same time to make sure they are exactly the same dimension. The top and bottom plate will be about 200mm wide, not critical but near enough and the vertical web 150mm high, not critical but all webs have to be the same dimension to make leveling more simple. I will then put the beams on saw stools and level them, simple, quick and cheap I hope. I will put some MDF support plates in the web to ensure it sits vertically while the glue is going off. As for mass in a bench, just fill the voids with sand if you want some real weight.

[Harry Holzke]

I don't think you need to add any sand. I did some calculations and my 4'x8' table comes in just under 600lbs. It is a beast.

[Doug Hepler]

One point has been overlooked so far in this thread. A flush door is a torsion box and is usually dead flat. When I was planning my current workbench I intended to build a torsion box to use as the top. I was puzzling about how to make a dead flat base, the same issue as the OP. I was reluctant to fiddle with loose 2x4's, shims, etc on sawhorses to make a flat surface. I decided that an interior flush door would make an excellent base for assembling my torsion box. I had used a cosmetically challenged (i.e., cheap) interior flush door as an accessory assembly table in the past.

Then through pure luck I happened onto a large exterior flush door. I simply glued 3/4" birch plywood to both surfaces with PVA and voila! I have been using this workbench 2 years now as my principal workbench and as a flat reference surface for assembly, etc. I am completely satisfied. It would have been economical to make even if I had bought the exterior flush door. The 3/4" veneer core birch plywood is plenty hard enough for a skin. With a 4x4 legs, a 2x4 frame, and junk stored on the shelf, it is heavy. It does not move when I am planing, sawing, etc. The interior of the workbench top holds bench dogs and hold-downs well.

Hope this interests you.

[Alan Schaffter]

I had a two-part article published in American Woodworker several years ago that answers many of the questions raised here. Let me correct a misconception first- A torsion box is not in itself intended to be a flat reference surface. It can be made so, but the real purpose of a torsion box is to have a strong, bend/sag resistant, but light weight structure; something as strong as a solid beam of the same dimensions, but much lighter. You can make curved and even serpentine torsion boxes- Google Ian Kirby.

http://www.ncwoodworker.net/pp/data/1421/medium/Beam_FigA.jpg

Flatness and durability come with how you build it. Airplane wings are torsion boxes, but you will often see what is known as oil-canning of the surface- depressions between ribs, because designers value light weight over skin rigidity and thickness. Start with a dead-flat construction platform and with just a little care, you will end up with a dead flat torsion box. I anchored my saw horses to the floor with screws, then made a grid with jointed and planed 2X4s so they had straight edges and were the same thickness. I used winding sticks to check for flatness, then I just used hot melt glue to assemble it.

You need to make trade-offs when you build a torsion box shop bench- if you want a tough, rigid top that will hold up to hammering and other abuse, you'll need heavier duty skins than you would for a bench that just needs to be flat for assembly or vacuum bagging. But, you end up with a heavier bench.

There has been discussions of the height, thickness, and spacing of the web, thickness of the skins, whether to dado the inside of the skins for the web, need to tack or staple the skins to the web, etc.

Web material can be almost anything . . . as long as it resists compression along its plane and can be attached securely to the skins and is dimensionally stable. I typically use the same material for the web as I do for the skins, often 1/2" MDF- same uniform and stable properties, plus an edge that is wide enough to give a good glue surface and that will hold a few brads to hold it all together while the glue dries. The taller the web the stronger the torsion box- up to a point. I typically use 2" - 3" tall web pieces space at 6" - 8". Don't waste your time with David Marks' or the Wood Whisperer's laborious cutting a fitting technique. It is not needed and can even cause problems. Just make a simple indexing jig to cut "halved" joints on your tablesaw with a dado blade- the notches help the web stand up and keep everything in line. They can be cut in no time (less that 5 min. to cut all laterals and longitudinals), and don't require time-consuming fitting of the last row. As long as all of the full strips were cut to the same length, you don't ever have to do any piecing when you get to the last row of cells- the cells in the last row might be a fraction larger or smaller than the others, but since all of them in the last row will be identical, who cares!

There is no need to dado the skins to accept the web either, glue alone holds just fine, and you are just asking for surface irregularities with dados. There is no need to use epoxy either.

You can make a suitable, extremely lightweight torsion box bench (or scaffold) from just 1/8" tempered hardboard skins and webs that will support a lot of weight but won't necessarily be durable. Here is one I built for my article to demonstrate is strength. It is 8' long by 18" wide X 3" thick, held together with glue only. I loaded the center with over 350 lbs of bricks yet it deflected less than 1/2" at the center. Most of the deflection apparent in the photo is due to the camera lens:

http://www.ncwoodworker.net/pp/data/1421/medium/P1040010.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P1080028.JPG

Rather than repeat my whole article, maybe you can find an old copy of AWW somewhere. I shot all of the following pics which were included in the article.

http://www.ncwoodworker.net/pp/data/1421/medium/145_AssemTable_spread1.jpg

http://www.ncwoodworker.net/pp/data/1421/medium/Cutaway_Test-1.jpg

http://www.ncwoodworker.net/pp/data/1421/medium/P10100792.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P10100834.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P10100705.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P10100983.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P10101181.JPG

http://www.ncwoodworker.net/pp/data/1421/medium/P10101301.JPG

[Chris Parks]

Yep, light and stiff. The one I am making for my wife's sewing room is made entirely of 6mm MDF including the topand bottom skins. I need a flat reference surface for it as she will yell at me if it is not flat and I can't stand the noise.:D

[Martin Wasner]

If it needs to be really flat, why not pick up a granite slab, or something similar like Corian? If you're not picky about color, you can get remnants pretty reasonably.

[Chris Parks]

Martin, unfortunately I don't have access to things like that where I live so I have to resort to making stuff. I finished the beams today and will start the torsion box tomorrow.

[Gerry Grzadzinski]

If it needs to be really flat, why not pick up a granite slab, or something similar like Corian?

Corian is rarely flat, and a full sheet is so flexible that it will conform to whatever it's resting on.

[Martin Wasner]

You're right, not solid surface. I'm thinking Silestone.

[Tom Hyde]

I'm not saying this is necessarily the right way to do it bit I used solid lumber 1x4 for ribs, acclimated to the shop, then jointed and planed to equal thickness. I used the Marks method of sawhorses and 2x lumber to create a flat surface to build on. I also used 1/2" ply for top and bottom. I hate mdf but it would have been easier to get dead flat. But I hate mdf. Really.

The table was topped with a sheet of 3/4" ply for the integrated kreg trak on two sides. Everything was topped with the same "golf ball" style laminate that some companies use on their commercial jigs and tables. The table is very flat, incredibly useful with the kreg trak, and glue pops right off. My table is high since I plan to make mostly cedar gates, not furniture.