Yup, torsion boxes are strong!!

[Alan in Little Washington]

If you didn't already know, torsion boxes are very strong and in the case of the one I just built, extremely light too!!!

I built the 8' long X 3" thick X 14" wide torsion box below for the article I am working on. I made it from very thin, 1/8" Masonite hardboard, some even weaker thin peg board, and regular ww glue. As you likely know, hardboard this thin is very flimsy and a narrow strip held at one end can snap under its own weight.

For a test, I loaded it at mid-span with over 300# of bricks. I would have used more, but didn't want it to break, in case I need to take more pics of it later.

Here is the internal web. I have already spread glue on the web edges to attach the second skin which is at the right of the pic:

The beam is loaded with 300#+ of bricks (the pyramid is two bricks deep):

Total deflection at mid-span was right at 1/2" with no signs of impending structural failure.

No load:

Under load:

 

[michaelgarner]

Wow!!! I swear you need to start a career going around and fixing up other guys shops with all the stuff you do. How did you cut the notches on the hardboard? Did you use a router table?

 

[Alan in Little Washington]

Wow!!! I swear you need to start a career going around and fixing up other guys shops with all the stuff you do. How did you cut the notches on the hardboard? Did you use a router table?

I used a simple jig on the table saw- like a box joint jig- a 2X4 attached to a miter gauge with a thin registration peg mounted in a notch. I had to make two passes to cut each web notch because my dado was wider than the thickness of the hardboard so I couldn't use it and the regular blade was narrower than the hardboard.

I set the alignment peg up for the width of the blade, and enlarged the first notch on each web piece manually by backing the piece away from the side of the peg slightly for a second cut. After the first notch was cut to the proper width, I used the jig's peg for alignment to cut the remaining notches. I made two quick passes with each side of the notch against the sides of the peg to cut each notch.

Bottom line, it went a little slower than if I had been able to use my dado blade, however I cut all notches in all pieces in about 10 min.

 

[ptt49er]

That's a great demonstration of the strength of the torsion box.

It kind of makes one wonder what a torsion box built w/ 3/4" webs and double layered tops and bottoms would hold.

 

[Alan in Little Washington]

That's a great demonstration of the strength of the torsion box.

It kind of makes one wonder what a torsion box built w/ 3/4" webs and double layered tops and bottoms would hold.

It will likely hold more but not as much as you might think. Plus it would be so heavy you probably couldn't lift or move it. Please don't take this the wrong way, but what you suggest is really not needed because of the engineering principles behind the torsion box. Four main things make a torsion box strong and resist bending under load:

1. The top skin's ability to resist compression along its surface (not perpendicular to it).

Thickness doesn't matter. The skin can be thin like the aluminum skin of an airplane wing. Two layers for the top skin is overkill and adds way too much weight which can actually reduce load capacity.

2. The bottom skin's ability to resist tension along its surface.

(same notes as top skin)

3. The ability of the web to keep the skins apart.

The web can also be really thin as long as it doesn't compress or buckle, hence the cardboard web inside many hollow core doors. 3/4" webs are really overkill and add way more weight than stiffness.

4. Finally, the depth of the web.

Depth of the web matters! Up to a practical limit and within the limits of the materials, if you increase the web depth by only 25% e.g. from 2" to 2.5", you increase the strength and reduce deflection by nearly 100% (double it). If you double the web depth e.g. from 2" to 4", you increase the strength 8 times!!! (the cube of the web depth) So here is where you really make your money!!

Other factors are involved also, like the bond between the skins and the web. The bond between the web and the top skin must be good to prevent the top skin from buckling under compression (along its surface.) A similar but slightly different situation occurs between the web and the bottom skin. That is why the thin wing ribs of airplanes usually have top and bottom flanges which are riveted to the wing skins.

If you can wait until sometime this summer and if my editor doesn't fire me first, you will be able to read all about this topic in one of the WW mags. :wsmile:

 

[ptt49er]

After I posted that, I got to thinking about some of the principles you brought up.

I wonder if you could use 3/4 mdf for a top and 1/4 ply for a bottom.

Something else for me to try...

Thanks for your response, it's amazing how the fingers/mouth is so much quicker than the logical side of the brain.

I can't wait to see your article.

 

[Alan in Little Washington]

After I posted that, I got to thinking about some of the principles you brought up.

I wonder if you could use 3/4 mdf for a top and 1/4 ply for a bottom.

Something else for me to try...

Thanks for your response, it's amazing how the fingers/mouth is so much quicker than the logical side of the brain.

I can't wait to see your article.

Depends what you want to do with the table. If you need surface durability that can handle a lot of point loading (hammering directly on the surface) then 3/4" might be best for the top. A bottom skin of 1/4 might be ok too, however- in a torsion box under load the forces of compression in the plane of the top, equal the forces of tension in the plane of the bottom. Different materials with different mechanical properties might cause an imbalance in forces which could result in unwanted side effects- warping, etc. Without doing a detailed analysis I just don't know.

The assembly table I am making for my article is made from 1/2 MDF skins and webs- all parts for this 40" X 63" x 3" assembly table can be made from just two sheets of MDF. If you change one skin you need three sheets of material. I haven't tried any other cut diagrams but believe you may be able to make a table as large as 42" x 72" using only two sheets of 1/2" MDF. It is easier to get 1/2" MDF to lay flat than it is 1/4" or 1/8" hardboard, and much easier than ply.

 

[SteveHall]

Alan, just curious why you notched and fitted the ribs instead of simply cutting a bunch of cross pieces all the same size and glueing them individually. Does it help to align the forces through each a bit more precisely? I'd think the sawing time would be quicker, although perhaps the gluing time is faster this way, too? Do the little leftover portions of cross rib do anything?

Where does one get HDF that is smooth on both sides? The typical box stores have it finished only on one--would it matter structurally if only the outside faces were finished? (Wouldn't the glue have a more open/fibrous surface on the inside that way?)

That is one inexpensive way to build a flat assembly table. I've been using four 12" x 84" solid core door segments on saw horses but they don't have the rigidity of a torsion box this deep. Sorry for all the questions, this is an inspiring idea and experiment.

 

[Alex Franke]

Very cool!!

Now I have a use for all those pegboard scraps I have in my garage.

Thanks for posting!

 

[junquecol]

Steve, from your questions, I deduce that you are planing on making an assembly table. Norm (NYW) used 1/4" plywood for the skins on his, and added a layer of 1/2" MDF for a replaceable work surface. Smooth on both sides hardboard is available from Wurth (aka, Raleigh Hardwoods). Main thing to remember when making a torsion box is it will mirror the surface it is assembled on. If assembly surface isn't flat, box won't be either.

 

[Alan in Little Washington]

Alan, just curious why you notched and fitted the ribs instead of simply cutting a bunch of cross pieces all the same size and glueing them individually.

Much, much, quicker to make and assemble my way. No brads are needed to assemble grid. Also, the overall dimensions of the box and straightness of the sides are unaffected by tightness or lack thereof of the glue joints between each individually cut piece. Unless your dado is not set right and you splay the notches when slipping one piece over the other, the length of each grid never changes and the grid stays straight. There is no measuring, fitting, nor cutting of every single piece in the last row which is necessary if you use the David Marks (and Wood Whisperer) piecing method.

Does it help to align the forces through each a bit more precisely?

Just a little, but there is not much force in the grid parallel to the skin. Remember the forces in the grid are mostly compression since the main purpose of the grid is to keep the skin apart.

I'd think the sawing time would be quicker, although perhaps the gluing time is faster this way, too?

Both are quicker my way with 1/4" or thicker MDF since the notches can be cut precisely with a dado blade. Using my simple jig which is somewhat like a box joint jig, I easily cut accurately spaced notches in all grid pieces for my assembly table in about 10-15 min. You can even use different spacing for the long pieces and short pieces if you want by resetting the alignment peg when you switch cutting to the other size pieces. Gluing is much quicker my way too. Here is a shot taken when I cut the grid pieces for my assembly table, the alignment peg is just visible in the last notch.

Here is the completed grid to my assembly table. Make no mistake, while inexpensive and easy to build with 1/2" MDF, a table this size will be heavy!

Do the little leftover portions of cross rib do anything?

On my assembly table I used 1/2" MDF for the sides and was able to glue and brad nail the sides in place. With the hardboard torsion box, the web pieces were too narrow to be able to brad nail the sides. I could have clamped the sides in place but was too impatient to wait for the glue to dry so I just made the sides with notches like any other web piece - so the cross pieces would hold it in position while the glue dried - just as strong. Remember an I-beam has no sides. A torsion box is just a bunch of I-beams put together. The only reason you need a side is for durability in the shop- this hardboard torsion box was just a test. It is not intended for my shop, but I'll probably keep it to use as a scaffold to clean my porch ceiling! :wsmile:

Where does one get HDF that is smooth on both sides? The typical box stores have it finished only on one--would it matter structurally if only the outside faces were finished? (Wouldn't the glue have a more open/fibrous surface on the inside that way?)

HDF? The stuff I used was 1/8" hardboard. I'm not sure it is the same as "tempered hardboard" which may be the stuff that is smooth on both sides and what I wanted. All I could find were the small handi hardboard panels with two smooth sides. The label on the 4' x 8' hardboard bin at Home Deport said the stuff I bought was smooth on both sides, but it wasn't! It had a rough textured surface on the back. I decided to see how that would bond with the grid anyway.

That is one inexpensive way to build a flat assembly table. I've been using four 12" x 84" solid core door segments on saw horses but they don't have the rigidity of a torsion box this deep. Sorry for all the questions, this is an inspiring idea and experiment.

Remember, since they are designed to be flat, but not support surface loads, most hollow core doors have sparce grids made from cardboard! The skins on these doors are pretty weak also.

 

[Alan in Little Washington]

Steve, from your questions, I deduce that you are planing on making an assembly table. Norm (NYW) used 1/4" plywood for the skins on his, and added a layer of 1/2" MDF for a replaceable work surface. Smooth on both sides hardboard is available from Wurth (aka, Raleigh Hardwoods). Main thing to remember when making a torsion box is it will mirror the surface it is assembled on. If assembly surface isn't flat, box won't be either.

You are spot on about the assembly surface. Here is one place I did use David Mark's method with winding sticks. The assembly platform must be flat, it does not need to be level.

Actually, I think Norm used very, very soft Homosote (recycled compressed newspaper?) for the replaceable surface.

You need to think about how you are going to use your table and what properties you want it to have in order to select materials- some properties are compatible but not all are possible at the same time. Trade-offs are necessary.

1. Do you want a really FLAT surface?
2. Do you want a BIG table surface?
3. Do you want a VERY LIGHT table?
4. Do you want a REALLY TOUGH and DURABLE table and top surface?

For example:

You can have 1., 2., and 4. at the expense of 3.
You can have 2, and 3. at the expense of 4. and possibly 1.

I wish my article was already published. It addresses many of these issues in the second part.

 

[SteveHall]

Make no mistake, while inexpensive and easy to build with 1/2" MDF, a table this size will be heavy!

That was my interest in using hardboard, even lighter.

HDF?

High Density Fiberboard, the non-proprietary and generic name for hardboard and Masonite.

The label on the 4' x 8' hardboard bin at Home Deport said the stuff I bought was smooth on both sides, but it wasn't! It had a rough textured surface on the back. I decided to see how that would bond with the grid anyway.

Same stuff I was contemplating, the rough texture side doesn't seem as dense and I was curious about how it would bond to the webbing (the smooth face being on the outside).

Remember, since they are designed to be flat, but not support surface loads, most hollow core doors have sparce grids made from cardboard! The skins on these doors are pretty weak also.

Yes, my doors are (were) solid core, essentially a particle board without voids. It increases their weight, which for a door is desirable due to the increased sound insulation. But as a torsion box it only makes the table feel a bit more solid but still with nowhere near the stiffness you see using a 3" thickness, and that with lighter material.

Here is one place I did use David Mark's method with winding sticks. The assembly platform must be flat, it does not need to be level.

One of the techniques I plan to use.

Actually, I think Norm used very, very soft Homosote (recycled compressed newspaper?) for the replaceable surface.

This was just illustrated in the Winter 2008 issue of FWW Tools and Shops on a two piece assembly table/bench. That design had two components about 16"W x 72"L x 4-1/2"H and was a torsion box with webbing apparently made of 1x4s.

Trade-offs are necessary.

1. Do you want a really FLAT surface?
2. Do you want a BIG table surface?
3. Do you want a VERY LIGHT table?
4. Do you want a REALLY TOUGH and DURABLE table and top surface?

For example:

You can have 1., 2., and 4. at the expense of 3.
You can have 2, and 3. at the expense of 4. and possibly 1.

I thought your clever bit was the hardboard, if the joints hold up it could satisfy all four. Laying a piece of loose Homosote on the top would add durability without sacrificing the weight to really accomplish #4. I suppose particle board could be used to save some weight over MDF, but nothing would be lighter than 1/8" hardboard.

Thanks for sharing the pictures with your jig, that explains a lot.

 

[junquecol]

Actually, I think Norm used very, very soft Homosote (recycled compressed newspaper?) for the replaceable surface. For the assembly table, he used 1/2" MDF. Skins were 1/4" underlayment plywood. He only used cross stretchers. The frame work of the table furnished support lengthwise.I also remember him using homosote for a work bench top some years back.

 

[Alan in Little Washington]

Actually, I think Norm used very, very soft Homosote (recycled compressed newspaper?) for the replaceable surface. For the assembly table, he used 1/2" MDF. Skins were 1/4" underlayment plywood. He only used cross stretchers. The frame work of the table furnished support lengthwise.I also remember him using homosote for a work bench top some years back.

Norm didn't build a torsion box. I guess he didn't take any ME courses in college (yes, he attended, didn't graduate), if he thinks that is what he built. Sounds like he didn't follow engineering principles- overbuilt some parts and under built others. But like I said, an assembly table means different things to different people and so there are tradeoffs in materials and resulting properties. David Marks seemed overly concerned in achieving a very flat surface. Size matters also, a smaller table with shorter spans will not need the structure of a larger table.

If you are going to put a replaceable protective layer over the top skin of a torsion box, unless it weighs a ton, you don't include it in the engineering design. Since it is not normally bonded to the structure, it adds absolutely nothing to the strength or rigidity of the table. You select the material for the replaceable top based on what you plan to do on it. No need for a 1/2" thick MDF replaceable top if you are only going to assemble woodworking stuff on. Now if you are assembling metal structures and often use a 25 lb. gentle pursuader to pound on things into submission, that is different.

A related topic- you will see a number of articles recommend that MDF router table tops be laminated with high pressure laminate on BOTH the top and bottom. Two reasons- (1) resistance to moisture is the same on both surfaces, (2) since the laminate is bonded to the MDF it makes the top somewhat like a torsion box- laminate improves the resistance of the top to compression, while the bottom resists tension.

 

[Alan in Little Washington]

That was my interest in using hardboard, even lighter.

Just be careful of durability of the structure. It might not hold up well being kicked around a shop. I'm pretty careless in that way, so opted for the 1/2" MDF for my personal table.

Same stuff I was contemplating, the rough texture side doesn't seem as dense and I was curious about how it would bond to the webbing (the smooth face being on the outside).

My concerns as well. I assembled my experiment with the smooth faces out. Unless you repeatedly load it to the max (or occasionally overload it), I don't think you'll have any trouble with the bond between the fuzzy side of the skin and web separating. I used a big bead of glue. You might want to use small pieces of hardboard to test the bond between web and fuzzy side compared to smooth side. Also compare the difference between a joint made with PVA yellow glue and made with epoxy to see how they react to overloads.

Yes, my doors are (were) solid core, essentially a particle board without voids. It increases their weight, which for a door is desirable due to the increased sound insulation. But as a torsion box it only makes the table feel a bit more solid but still with nowhere near the stiffness you see using a 3" thickness, and that with lighter material.

Those are solid but oh, so heavy, and why a torsion box is better.

This was just illustrated in the Winter 2008 issue of FWW Tools and Shops on a two piece assembly table/bench. That design had two components about 16"W x 72"L x 4-1/2"H and was a torsion box with webbing apparently made of 1x4s.

Even for that application the 1X web material might have been overkill.

I thought your clever bit was the hardboard, if the joints hold up it could satisfy all four. Laying a piece of loose Homosote on the top would add durability without sacrificing the weight to really accomplish #4. I suppose particle board could be used to save some weight over MDF, but nothing would be lighter than 1/8" hardboard.

As above, I'm not talking about surface durability, but durability of the entire structure. I'll say one thing, you would have a light table that you could easily put away if short on room. I forgot to take a shot of me holding the hardboard bench (bridge) over my head easily with one hand.

Thanks for sharing the pictures with your jig, that explains a lot.

No problemo

 

[Badabing]

Alan, thank you so much for posting this information. It has helped answer some questions that I have been thinking about all week...funny coincidence!

On Monday, I got together with my brothers and a friend to play some poker. We had a lot of fun but my brother's table was not the best for card play. We were talking about doing it again next week but at someone's house with a bigger table so we could have some more people join. I started to think about making a poker table. I've seen a few shows that featured this project (Norm made one with Mahogany, and John and Jimmy Diresta made one out of MDF on "Hammered").

I was thinking of just making a top for my brother's table that would be very light and strong and he could store in the garage when not in use. I thought of a torsion box design but figured it would be too heavy. I've been thinking about it all week and you just answered some of the questions that have been stirring in my head :icon_thum Maybe I can find the time to put one together based on your design...

Thanks!

 

[garymuto]

interesting review and discussion. I'm planning a torsion box extension table for my saw. it will double as an assembly table.

 

[Gotcha6]

I found the door skins they sell at some BORGs (3x7x1/8"plywood) are right at the same thickness as the kerf on my TS. I used some of that to make dividers for LOML's boxes she keeps Christmas ornaments in. I'd guess the procedure would be the same but I hafta wonder if the 1/8" plywood wouldn't be just as strong with the added ability to be dadoed into the skins. BTW, who much did the finished torsion box weigh?

 

[Alan in Little Washington]

I found the door skins they sell at some BORGs (3x7x1/8"plywood) are right at the same thickness as the kerf on my TS. I used some of that to make dividers for LOML's boxes she keeps Christmas ornaments in. I'd guess the procedure would be the same but I hafta wonder if the 1/8" plywood wouldn't be just as strong with the added ability to be dadoed into the skins. BTW, who much did the finished torsion box weigh?

Having web material with a thickness that matches a blade kerf makes making the web a snap.
It might be really tough to get the grid to line up with dados on the insides of the skins. You could actually use a giant version of the jig I used to cut the web to cut the skins, but it would be difficult. Bonding the skins to the web doesn't need dados in the skins. A good bead of WW glue or epoxy is almost as strong and easier to do.

You have a real balancing act between durability, strength, and weight. The problem I have with either thin skins or thin web for an assembly table is that they might not be durable in a shop environment, especially to clamping pressure and more so if the clamp is over a web void.

The hardboard torsion box was light, don't know the exact weight but will weigh it someday. It is easy to hold up in the air with only one hand- it was the weight of one sheet of hardboard.