Torsion Box build.. What did I do wrong?

[Greg Woloshyn]

I just finished building an MDF torsion box as designed by David Marks and followed the exact procedure for making one, but ended up with a top that is bowed a total of 3/16". What I did to make my flat surface to start with was build a series of MDF I beams, since I don't have a jointer. These were placed on a workbench, leveled and shimmed so they were all on the same plane. Then I laid a sheet of MDF on top and confirmed flatness. After ripping all my pieces for the gridwork at the same time and gluing everything up, I flipped it over and added my MDF skins.

It seemed fairly flat at that point, but when I transferred the box to the top of my cabinet it is now very warped. What did I do wrong?

289627289628

[Shawn Pixley]

I am reluctant to comment on this because the last torsion box post I commented upon, others were unhappy that my viewpoint didn't agree with theirs. But here it goes. I think the problem may lie in the I beams. By shimming you got a flat surface on the "top" but you don't know that the bottom plane is at all flat. If the I beams were not perfectly true and small inperfections were added with the webs/gridwork, you can end up as you did. This might or might not be the problem but would be my first choice given the information available.

[Greg Woloshyn]

I am reluctant to comment on this because the last torsion box post I commented upon, others were unhappy that my viewpoint didn't agree with theirs. But here it goes. I think the problem may lie in the I beams. By shimming you got a flat surface on the "top" but you don't know that the bottom plane is at all flat. If the I beams were not perfectly true and small inperfections were added with the webs/gridwork, you can end up as you did. This might or might not be the problem but would be my first choice given the information available.

I think your right. My problem does lie in the I-beams because I checked them and they turned out to be bowed at the ends. I thought for sure this type of setup would be strong enough to resist any movement but I guess I was wrong.

At this point I'm probably going to scrap the entire torsion box and start over. In the videos I saw jointed 2x4's were used, but don't those tend to move as well? Or will I have to work fast and hope for the best?

[Jim Matthews]

Please elaborate on the construction.

How far apart are each of the "ribs"?
In the David Marks design, he mentions that they're just far enough apart to
fit his pin nailer; in passing he stated that the smaller each box section is, the stiffer the torsion box.

If the box is drooping or sagging - that sounds like the culprit.
If the box is rigid, but out of plane - you should be able to skim on a final surface to get back to flat.

I'm thinking that a skim coat of Durham's Rock Hard with a screed made out of an aluminum level would serve.

http://waterputty.com/

[Andrew Gold]

I've made torsion boxes in the past with plywood, never tried mdf or solid sawn lumber. I would think mdf would run about as easily as plywood, no real reason why not at least.

I would have fear the "I-beam" glueup is a step that would introduce error, even a little bit could skew the final steps. I've always ripped thin strips of 3/4" ply to a bit wider than I want the final torsion box to be, face glued two of them together (with cauls and the like), then ripped the glued strips to finished width and assembled the box. Using a table saw you can easily ensure the last step before the final assembly you have even strips, very little chance for error...

[Art Mann]

I'm quite surprised that the I beam construction didn't work. Were they not stiff enough to start with or was it an issue of constructing them to be flat and straight. Did you glue the "I" together or just count on fasteners? The reason I am asking is I have constructed beams of similar dimensions to yours out of 3/4" plywood and used them to stretch between two saw horses that spanned about 6 feet. They supported a step ladder on a piece of 3/4" plywood while painting high interior walls. They were stiffer and lighter than 2 X 4s would have been. I got the idea and plans out of a woodworking magazine.

The same principle of dynamics that makes a torsion box so stiff is also what makes an I beam stiff. I don't see how using 2X4s to support the construction would be any better unless you have some way to make them flat when on edge.

Edit: Andrew, it was my understanding that Greg used the I beams to support the construction of the torsion box and not as a part of the box itself. If that is not the case, then I can see how that would complicate assembly and might introduce error as you say.

[Greg Woloshyn]

Lots of good information here. My torsion box grid was made 6 inches apart, with staggering sections so I could fit my nailer in, glued as well.

The I beams I made were NOT glued, just brad nailed. This could be why my beams shifted out of flat when I setup all the weight of the MDF on top of them. As you see in the picture, I used 2 additional beams underneath the longitudinal beams to start with - these remained flat. These two beams were spaced about 5-6 feet apart, with nothing to support the main beams in the middle.

I guess now I will either use the putty method as mentioned above, or start over with a more rigid setup. Possibly ripping plywood strips and face gluing together.

[Jim Matthews]

I'm disappointed for you.

I thought your preparation was well thought out, and can't see
anything I would have done differently, myself.

I'm contemplating one for my garage space and follow your thread closely.

[Pat Barry]

I just finished building an MDF torsion box as designed by David Marks and followed the exact procedure for making one, but ended up with a top that is bowed a total of 3/16". What I did to make my flat surface to start with was build a series of MDF I beams, since I don't have a jointer. These were placed on a workbench, leveled and shimmed so they were all on the same plane. Then I laid a sheet of MDF on top and confirmed flatness. After ripping all my pieces for the gridwork at the same time and gluing everything up, I flipped it over and added my MDF skins.

It seemed fairly flat at that point, but when I transferred the box to the top of my cabinet it is now very warped. What did I do wrong?

289627289628
I think we are assuming that the warp is along the the length of the torsion box and not across it. Could you confirm that this assumption is correct?

I like your I beam. That whole process seems a great way to go as you described it. Of course, I am also assuming that you clamped the torsion box assembly to these I beams as you built it up to keep the flat I beam reference. What this doesn't do however, is to give you the reference of flatness side to side, I could see the result being a twist or cup laterally rather than longitudinal. In order to minimize that element, it seems lateral I beams are in order as well.

[Bradley Gray]

I think the flex is from the brad nailed flanges. The image in the OP does not show any cross members in the grid. How were they installed in relation to the flanges?
I have only built torsion boxes from plywood and did not use flanges. I think the skins become the flanges in this type of assembly.

[Greg Woloshyn]

I think we are assuming that the warp is along the the length of the torsion box and not across it. Could you confirm that this assumption is correct?

I like your I beam. That whole process seems a great way to go as you described it. Of course, I am also assuming that you clamped the torsion box assembly to these I beams as you built it up to keep the flat I beam reference. What this doesn't do however, is to give you the reference of flatness side to side, I could see the result being a twist or cup laterally rather than longitudinal. In order to minimize that element, it seems lateral I beams are in order as well.

My side to side flatness was obtained by using 2 I-beams underneath the longitudinals as shown in the picture. I believe that since I didn't glue the flanges but just brad nailed them, they flexed under the weight because in the middle, there is no lateral I-beam to support it.

I did not clamp anything when I assembled the torsion box. The first skin of MDF was placed on top of the I-beams, then pushing down each piece of the grid and gluing/nailing them in place.

How about I build more lateral I-beams underneath the main ones to better support the load, and glue and nail the beams together this time.

[J.R. Rutter]

The I beams I made were NOT glued, just brad nailed. This could be why my beams shifted out of flat when I setup all the weight of the MDF on top of them. As you see in the picture, I used 2 additional beams underneath the longitudinal beams to start with - these remained flat. These two beams were spaced about 5-6 feet apart, with nothing to support the main beams in the middle.


How about I build more lateral I-beams underneath the main ones to better support the load, and glue and nail the beams together this time.

That's the ticket!

[Greg Woloshyn]

A disadvantage to the I-beam design would be any deviation from an exact 90 degree angle when the flanges are put on will cause trouble with side to side flatness. Maybe I should rip and face glue strips of plywood instead.

I decided to go with I-beams because of strength, and also to keep them standing straight.

[Greg Hines, MD]

I would agree. For making I-beams, I was always taught to dado the ends, so the beam fits into it for lateral structure, and lots of wood glue, in addition to screws. However, the members crossing your bench seem to be a big part of this problem, and are likely the cause of your sagging. One of the things that Marks did in his video (as I recall) was to use sawhorses that spanned the width of his build up. I think that is where your problem lies.

Doc

[scott vroom]

For what its worth, here's The Wood Whisperer's torsion box video.

http://www.thewoodwhisperer.com/videos/episode-18-assembly-table-torsion-box/

[John Nesmith]

The I beams I made were NOT glued, just brad nailed. This could be why my beams shifted out of flat when I setup all the weight of the MDF on top of them.

This get my vote.

[Greg Woloshyn]

I'm considering building a benchtop flattening jig for the router to try and get the top flat first, before I build a new box. I'll make sure to sink the brad nails far in so my bit doesn't contact..

[rick carpenter]

Rebuilding with all glued joints would be my vote. You can brad or screw them too if you want. My torsion box is 3/8 ply with 3/8 ribs all glued. I have my chopsaw base permanently mounted to it. I doubled the outside ribs and screwed the outside perimeter for my peace of mind because it gets set on edge and bumped quite a lot when moving/storing plus it has a carrying handle at the balance point.

[Kevin Jenness]

Clearly you need a flat base to start with, rigid enough to withstand the weight of your torsion box plus any clamps used. Any bow or twist in the underlying framework may be reflected in the finished product. I have made numerous 4'x8' work surfaces from an original that was glued up with straight 4' cauls on a machinist's 3'x6' cast iron setup table. I would suggest that you look around your neighborhood for a metal fabricator with such a table who would let you come in and use it. Otherwise, make up your beams with straightlined material stiff enough to resist the forces involved in clamping up your torsion box. Plywood I-beams are good, but the webs must be straight to begin with- you may want to have a shop with a sliding table saw or vertical panel saw make them so for you. Once you have a flat stiff box, you can make another using clamps and cauls, then another generation for your friends and relations sandwich style.

In terms of construction, gluing the grid joints within the torsion box has no effect on the outcome. The strength and stiffness comes from gluing the grid to the skins. In my experience, the simplest way to proceed is to cut the grid elements full length in width and length and join them with half laps like an egg crate. This can be done on a table saw with a dado set, more easily on a slider. Knock together the egg crate, drop a 4-sided box around it, glue one side, drop on and pin one face, flip it over, repeat, clamp up with cauls on your flat, rigid base and viola! With a helper you can do the assembly with Titebond Extend within 20 minutes, otherwise use epoxy or uf glue

Your idea of flattening the bowed initial attempt is not bad, but be aware that will result in an "unbalanced " panel, with different thicknesses on each face It may serve as a base for a second attempt, but could be unstable going forward- or not.Luck may be a factor, though it is not a plan.

[Brian W Smith]

If you don't have a "flat base" on which to build/stage........Or,if we were "in the field",where our resources were challenged.I'd probably just go vertical with the assembly.Plywood has wonderful characteristics in/on the edge plane.It dosen't in the flat plane,which is why you're building a stressed skin panel,duh.So by going vertical with assembly you're taking more advantage of it's inherent strengths.Good luck,BW

[Chris Fournier]

I don't understand the reasoning behind making I beams in the core. Overly complicated, harder to build, introduces the potential for more error and really adds nothing to the torsion box but trouble. I built a 4' X 8' torsion top for my assembly table but I used the welded up base as the build platform and I built on the bottom skin. The project must be supported in a way that keeps it flat while its being built or it will sag, deflect, twist.

I hate it when a project goes south like yours has. You're not alone, anyone who builds stuff has been in your shoes. Materials are cheap, simplify your design, use a robust base and try again! Good luck!

[Greg Woloshyn]

I don't understand the reasoning behind making I beams in the core. Overly complicated, harder to build, introduces the potential for more error and really adds nothing to the torsion box but trouble. I built a 4' X 8' torsion top for my assembly table but I used the welded up base as the build platform and I built on the bottom skin. The project must be supported in a way that keeps it flat while its being built or it will sag, deflect, twist.

I hate it when a project goes south like yours has. You're not alone, anyone who builds stuff has been in your shoes. Materials are cheap, simplify your design, use a robust base and try again! Good luck!

I plan to try and route my tabletop flat first, I have nothing to loose anyway. BTW, my core was not constructed with I-beams, the I-beams were used for the initial setup for a flat surface, since I don't have a jointer. I used regular strips of MDF ripped on the tablesaw for the core.

[Alan Schaffter]

I may be repeating some other comments but just easier for me to start fresh-

First, I have researched and built a number of torsion boxes and wrote a two part article about using one for an adjustable assembly table that was published by American Woodworker in 2010.

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

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

The first issue and probably the cause of your problem- your torsion box table will only be flat if your construction platform is perfectly flat. It should be stable, anchored to the floor and perfectly flat- front to back and side to side. Take multiple readings every few inches with winding sticks and take all the time you need to make a perfectly flat platform. I applied hot glue to the platform parts to keep them from moving. I used a jointer to ensure the 2X4 framing lumber used in the platform was flat- it's gotta be flat!!

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

Rip your web pieces all at the same time, but to make assembly easier use a version of a box joint jig to make halved joints that you can apply glue (no brads are needed here) and slip together. Make sure the pieces fit together easily without binding or else the web will bend. Don't bother using David Marks' piecing method- it is too time consuming and can lead to error, especially when you get to the last row of cells. With my method, the last row of cells may not be the same size as those in the other rows but all of them in the last row will be exactly the same size and the table perimeter will be straight, so who cares! I marked the last cut section with an "X" so I would be sure of orienting all web pieces in the same direction.

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

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

I typically use 1/2" MDF for web and skins because the material is stable and consistent, and so I only need one thickness. If you will not apply heavy clamping or other heavy forces to it, thinner material will also work well especially for the skins. No need to cut dadoes in the skins either (as long as you have some glue surface at the edges)- too much work, can introduce errors, and just not needed. Yellow glue is fine, epoxy is overkill. I built an 18" X 8' torsion box from 3/16" tempered hardboard webs and skins all held together with yellow glue. For a test, I supported it at both ends, loaded it at mid-span with over 300# of bricks. Though it look like more due to the camera lens, it only sagged 1/2" in the middle!!!

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

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

Torsion box strength comes from the height of web, skins that won't stretch along their plane, and skins that are well bonded to the web. Initially, I stacked both skins on my platform, covered the top one with plastic wrap, then glued the web together on top of that. I used bricks at the web intersections to keep everything flat while the glue dried. Then I used glue and 18 ga. brads to attach one skin. After the glue dried I flipped it over and attached the second skin the same way.

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

If you build it with care, it will be flat.

[Greg Woloshyn]

I successfully flattened my top with a router flattening jig. It took me about 5 hours to flatten my top with the biggest bit I had, 3/4". Now it's very flat and much better than it was. I ended up taking off over a 1/4" in the middle it was bowed so much. Now my top skin is only a 1/4" in the middle but if it doesn't hold up than I'll have to rebuild the box.

[Greg Hines, MD]

I successfully flattened my top with a router flattening jig. It took me about 5 hours to flatten my top with the biggest bit I had, 3/4". Now it's very flat and much better than it was. I ended up taking off over a 1/4" in the middle it was bowed so much. Now my top skin is only a 1/4" in the middle but if it doesn't hold up than I'll have to rebuild the box.

Now that it is flat, you might want to laminate another 1/2" thick skin on top of your now thinned out one. It will also make it smoother.

Doc