MDF Torsion Web Warping

[Jason Hanko]

Hey everyone,

So I set about making a torsion box to be used as an extension table for my TS. My plan is/was to use 3/4" MDF to build the inner web or grid that goes between the outer skins. Everything went as planned until I assembled the web, when I noticed that the web pieces are now warped.

The web was assembled by cutting a series of half-lap joints in the pieces so they slip over one another to form the grid. The pieces are approximately 30" long by 2.5" wide (3/4" thick); the half laps were cut every ~ 6". Now I know that the lengths of MDF weren't curved when I ripped them to width and crosscut to length - Ive got a few extra backup pieces and those are still straight. Now all of the assembled pieces are warped so that they are concave along the opposite edge that I cut the half lap joint into, with about 1/8" deflection over the length of each piece.

Im fairly certain that I didnt bend them during the dry fit, so Im thinking that they warped after I cut the half-laps? Is this normal for MDF, I thought the stuff was supposed to be super stable...

Anyways, Im wondering if I should just start over with some plywood instead of the MDF. Im guessing I dont want to build a torsion box out of something that isnt even close to flat...

[John TenEyck]

Actually, you don't need any joinery for the internals of a torsion box. You can, of course, but you don't need to. You can just butt joint everything together and use staples to hold the pieces together while the glue dries. (The strength comes from the shear strength between the grid and skins, and the height of the grid, not from the actually strength of the grid itself.) Then glue on your skins making sure it's all on a really flat surface. Maybe you can rework what you have to get a flat grid, otherwise, I'd start over.

John

[johnny means]

Sounds like your joints are too tight. Every joint is causing a slight bend. No need for tight or even snug joints in this application.

[Paul Hingco]

Did you assemble it on a dead flat surface?

[Mike Cutler]

Yep, MDF will warp with moisture, heat, or stress, it's just dust and glue. Lay a piece across a curved surface, outside, at night, and it will change shape. But it is pretty stable stuff. I wouldn't expect it to change shape after it's assembled though, unless it's subjected to non uniform stresses.

That picture doesn't show a lot of "settling". I'd be tempted to fair it out with Bondo or wood putty and use it as is. It's probably changed as much as it will.

[Art Mann]

You need to assemble the torsion box on a very flat surface and use clamps or weights to force the surfaces flat while the glue dries. David Marks has a great video showing how to do that. Here is a link.

http://vimeo.com/5082731

As several have already mentioned, the appearance and assembly of the internal grid is unimportant except that the width of the grid pieces must be exactly the same.

[Alan Schaffter]

First, you don't need a heavy duty shear web (internal web structure). The web material shouldn't stretch or compress along its surface and should be thick enough so there is enough glue surface for the edges to bond adequately to the skins. There are other stresses on the web, but it mainly keeps the skins a constant distance apart. A 3/4" MDF shear web is overkill and eliminates one of the main advantages of a torsion box- light weight. I typically use the same material for all parts so I know it will all react the same- I usually use MDF. Solid wood and even plywood warp too easily and often have internal stresses. For economy, I often use 1/2" MDF everywhere- skins and web- when making a torsion box that will be subject to a lot of abuse.

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

That being said, if the surface of your torsion box will be used as a TS extension or outfeed table and won't be subject to point loads (clamping, hammering, etc.), a torsion box made entirely of 3/16" or 1/4" tempered hardboard should work just fine. I wrote an article about torsion boxes which was published in the December/January 2010 issue of American Woodworker that delves more deeply into torsion box engineering. Here is one picture from that article that shows an 8' long torsion box made primarily (web and skins) from 3/16" hardboard. I ran short of 3/16" hardboard so made the short lateral webs from scrap 1/4" pegboard. In the photo, the completed torsion box is supporting over 350# of bricks, mid-span. Though it looks like more, when I measured the amount of sag, it was only 1/2" there!

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

Here it is under construction on my adjustable height torsion box assembly table- I'm preparing to add the final skin:

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

Like others have said you need to start with a flat construction platform (I use David Marks' method). While making torsion boxes I first stack the skins on the construction platform and cover them with plastic while I first assemble the web using yellow glue only. Once the glue has set, I align and attach the web to one skin with glue and brads. When the glue is hard I add reinforcing for leg pockets (see first pick above) and other hard points, then attach the second skin with glue and brads. I use weights and clamps to keep everything flat during each step. My torsion boxes are dead flat.

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

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

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


I use a simple indexing jig to get uniform spacing when I cut the halved joint notches which makes assembly easier. David Marks' piecing method (copied by Mark Spagnulo) is more time consuming, makes it more difficult to achieve a flat surface and straight sides, and can be a real pain!

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

Like johnny said, I think your notches are too tight so when assembled the notches spread slightly causing the web to curve away from the notched edge. Set the width of your dado blade so the notches easily slide onto each other!

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

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

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

[glenn bradley]

Sounds like your joints are too tight. Every joint is causing a slight bend. No need for tight or even snug joints in this application.

I know we are all speculating with the info given but, I'm with johnny. If your joints don't literally drop over each other with no resistance, you probably picked up a an accumulating error over the span.

[Lloyd McKinlay]

Alan, great tutorial but the best part was seeing your clamp cart in the background. I spent part of yesterday looking at my scrap plywood and trying to figure out an easy and effective way to organize my clamps while providing mobility. Now I have my solution!

[Roy Harding]

I agree with all the possibilities mentioned above, and would throw one more possibility into the mix:

If you rested your grid on a cement floor overnight, the edge in contact with the cement could have drawn moisture in from the cement, causing warping.

Like others have pointed out - you don't need such a robust grid for your interior grid, 3/4" MDF, spaced every six inches is going to make for a very heavy torsion box.

[Jason Hanko]

Sounds like your joints are too tight. Every joint is causing a slight bend. No need for tight or even snug joints in this application.

Yes indeed - Johnny identified the problem correctly! I had made my joints fairly snug since, well, isnt that what we do with joints? :rolleyes: This being my first torsion box I didnt know any better. Ive re-made my web pieces using a much looser joint, and things stayed dead flat this time.

Thanks, as usual, SMC!

[Alan Schaffter]

Alan, great tutorial but the best part was seeing your clamp cart in the background. I spent part of yesterday looking at my scrap plywood and trying to figure out an easy and effective way to organize my clamps while providing mobility. Now I have my solution!

Thanks,

Not to steal the thread, but that is the prototype clamp cart I built in the Spring of '06 from scrap 2 X 4 and MDF. It will hold around 80 original Bessey K body bar clamps of any size, including the really long ones- the only restriction is the ceiling height! Unlike tall and sometimes rickety A-frame clamp carts, mine is solid and has a low center of gravity- it is just as rigid and stable with one clamp as it is with 60! It worked so well I never got around to making the final version which was supposed to have drawers underneath and on each side for C, F, furniture, band, etc. clamps. With the K bodies resting on the jaws, I can easily grab the bars of 2 - 3 clamps in each hand and I don't need to worry about a jaw slipping and pinching the side of my hand.

http://www.ncwoodworker.net/pp/data/1633/Clamp_Cart-1.jpg

http://www.ncwoodworker.net/pp/data/1633/medium/P52800491.JPG

http://www.ncwoodworker.net/pp/data/1633/medium/P52800501.JPG

[johnny means]

I think this thread is a good illustration of how we tend to over examine and over engineer things. I like to keep KISS in mind, Keep It Simple Stupid. :rolleyes:

[Gordon Eyre]

Wish I had your clamp assortment. Makes my paltry 7 K body clamps look pretty meager. Like your cart a lot.

[Frederick Skelly]

This was a GREAT thread guys. I've been wondering about torsion boxes lately and I learned a whole lot from reading this. Thank you!

Sure wish I had a stack of K-bodies like yours, Alan!

Fred