torsion box not flat

[Paulh Tremblay]

I finished my torsion box assembly table and hoisted it on my frame. Unfortunately, it is not flat. It is about 1/16'' off on one diagonal.

The torsion box measures 4 X 6 feet. The skins are 3/4'' MDF. The web is 2 inches deep MDF.

I'm not sure exactly where I made my error. I had to make my reference frame by shimming boards on my basement floor. Nothing was off 1/16'' then, but maybe in the process something did get that much out of alignment.

For the short term I guess I will just just live with it. As the wood whisperer points out, you don't really rely on the assembly table for squareness. You rely on your joints, so being a little off doesn't matter. Of course, 1/16'' is really not a little.

If (or probably when?) I make another, I can use my current torsion box as a starting point. It is much flatter than my basement floor, and as long as I fix the 1/16'' flaw for my new reference frame, I should have better results.

However, before I do that (actually, it won't be for a while), I wonder if I can fix the current torsion box with another piece of MDF on top? I wonder if I could stick a few shims in the low spots and just glue or screw on another piece of MDF and have a flat reference area?

[Bruce Page]

1/16" over a 4'X6' area really is a little.

[Joe Jensen]

I've used self leveling epoxy to fix a warp in a surface. It worked really well. I used West Systems slow setting epoxy. Expensive solution though, about $100 for a gallon.

[Andrew Hughes]

I've made a few torsion boxes they never have come out flat for me.I remember having one I flipped over cause the bottom was flatter than the top I wanted to use .I think that was my last one I give up.

[Dave Zellers]

I've never built one, but I'd like to. I am going to guess something changed from your setup.

I thought the whole idea was if you build off a flat base, the TB will be flat and stay that way.

1/16" over 6 feet actually does seem like a lot to me given those parameters.

I say that after 25 years of building some pretty nice stuff on a plywood work table that had a 3/16 dip in the middle.

Whisper Dude's comment about cutting your carcase stock straight and true (square) being most important is spot on. If you do that, then just align the corners perfectly and all is good.

But life since then, on a flat table, is a LOT easier... :D

[Jim Matthews]

I'm wondering if this can be persuaded closer to your tolerance for error.

If one corner is "higher" than the others, it could be dropped from some
short distance above a block to tweak the alignment.

I'm with some of the others, I think the deviation is acceptable for making
things flat. If I'm trying to fit furniture in my house, the piece
is measured for stability in its resting position.

None of my floors are flat, and they determine how things should be fitted.

[John TenEyck]

I think you can flatten your torsion box w/o much difficulty with a router sled. Mount the box to whatever base you intend to and put it in it's intended location. Then screw two rails on the long sides of the box. Make sure they are longer than your box and make sure they are parallel. Build a router sled to ride on those rails, then flatten the top. Once done you can shellac the surface or tack down a sheet of 1/8" Masonite on your now dead flat torsion box.

John

[Doug Hobkirk]

My ideas are all included in the above. John's router sled barely squeaked in in front of me! Although I wouldn't have thought of the Masonite, which I think is probably a very good idea.

Do you have an accurate 6' straight edge? That impresses me, but I'm pretty hack.

The epoxy was a good idea, though expensive. The router sled could take a long time. Which is larger - the low area or the high area?

[Art Mann]

Lets see here.The diagonal length of a 4 X 6 square is about 7.2 feet or 86.4 inches. That means your table flatness is off by a whopping 0.07%. I am just sitting here pondering what possible application you might have in which that kind of accuracy is meaningful. Sorry, can't think of any.

[John TenEyck]

Lets see here.The diagonal length of a 4 X 6 square is about 7.2 feet or 86.4 inches. That means your table flatness is off by a whopping 0.07%. I am just sitting here pondering what possible application you might have in which that kind of accuracy is meaningful. Sorry, can't think of any.

I assume the OP built a torsion box because he wanted a really flat surface. In a couple more hours of work he can have one. To accept anything less would seem like a waste of all that effort.

John

[Chris Padilla]

How do you know it is off by 1/16"? What was the reference you used? Maybe that is off?

Lee Valley has a very nice and reasonable cost 50" aluminum straight-edge that I use all the time as a reference. In fact, I use this straight-edge to fix other references in my shop. Just yesterday I wondered if my "nice" 24" framing square was really flat along its edge. It was not! Some sandpaper taped to my jointer table and a few minutes and I got 'er all ship-shape!

MDF sands pretty easy...start with grinding down the high spots. Mark pencil squiggles on the MDF so you can see where you've sanded. I certainly think it is worth a couple of hours to get it better than 1/16" and I'm sure you can do that reasonably.

One thing that might have happened was with the introduction of glue (i.e. moisture...water), the MDF swelled a bit and didn't quite return back to where it was when it dried or maybe it still needs to dry? MDF WILL absorb moisture and move so once you are sure the glue moisture is gone and you've flattened it all you care to, slop some cheap poly all over it to keep the moisture out.

[John TenEyck]

You don't need much of a straight edge to see if you're off by 1/16" Any straight jointed piece of wood will do. Lay it on a sheet of paper, draw a line along the edge, flip it over and draw another on top of the first. If it's off by 1/16" it'll be easy to see. When the two lines form one line, you have a straight edge that's good enough for a lot of applications.

John

[Bill Orbine]

However, before I do that (actually, it won't be for a while), I wonder if I can fix the current torsion box with another piece of MDF on top? I wonder if I could stick a few shims in the low spots and just glue or screw on another piece of MDF and have a flat reference area?

MDF is very easy to plane with a good sharp, well tuned hand plane. I'm not knowing how YOU measure flatness, but for me with a good pair of winding sticks/straight edge, you can shave off the high spots. After accomplishing this, you could sand, add a new layer of MDF to the flattened side. Or find somebody who got a 48" (or more) wide belt sander and start sanding the other side first and then smooth up the first side. What kind of tolerance are you for anyways? Why?

[Paulh Tremblay]

I think I have actually overestimated the warp. There is a slight bow on the middle of the table, so when I put my 6 foot level from one side of the tale over this bow, it shoots up like a see saw. I measured between the end of the level and the end of the table, when really I should have measured the fulcrum point in the middle of the surface.

I would need to get my Festool router with dust collection before I attempt to make it flat with a router; the dust generated would coat my whole basement and last for months. I have too much experience with MDF! As a curiosity, though, how would one assure the sled was dead flat? It seems you would have to make some extremely accurate rail system.

Several posters question my need for such a flat surface. I assumed you would really want one, and other torsion box creators brag about their creations being flat to within thousandths of an inch. A few practical needs include placing a table base on it to test if the table would wobble, or gluing up 3/4''' strips of plywood, face to face, to make a table top, and wanting the result to be flat.

I had considered sanding the high spots, but worry a bit about creating dips on an otherwise flat surface. I think I will live with the slight imperfection for now, and if I find the lack of flatness hurts the quality of my projects, use one of the suggestions here.

[Jerry Wright]

Paul - my 2 cents - leave it alone and declare victory. I would be very happy with your result. A router will make a mess of a nice job - tough to control and harder to do than one imagines.

[Phil Thien]

A slight crown in the center may be ideal. If supported at the four corners and maybe around the perimeter with an apron, even a torsion box will sag in the center. Especially one made entirely of MDF.

So tell people the crown was intentional.

[John TenEyck]

You really should look up "router sled". It's dead simple to make and use one. One example: http://www.highlandwoodworking.com/flattenaworkbenchwitharouter.aspx It's just a mechanical process, nothing hard to construct or control about it nor does it make a mess, but it will make a lot of dust unless you connect it to a vacuum.

John

[Dave Zellers]

I think I have actually overestimated the warp.
Well now you tell us. We're about to go to blows here over a sixteenth of an inch. :)

Sounds like you have a 1/32" or less hump in the middle.

Time to start makin' stuff.

[Dave Zellers]

One question just for my knowledge-

Does the bottom perfectly mimic the top? ie, a slight dip?

[Art Mann]

You don't need much of a straight edge to see if you're off by 1/16" Any straight jointed piece of wood will do. Lay it on a sheet of paper, draw a line along the edge, flip it over and draw another on top of the first. If it's off by 1/16" it'll be easy to see. When the two lines form one line, you have a straight edge that's good enough for a lot of applications.

John

In order for the OP to use this technique, he would need a sheet of paper that is 7 feet long. remember the alleged discrepancy is over a very long distance. A foot long straight edge is of no use.

One easy way to measure linearity is to tightly stretch a string on the diagonal above the surface that is very accurately positioned on both ends and then use a ruler along the string to measure discrepancies. Repeat the procedure on the opposite diagonal and that will give you a pretty good idea of how far off the surface is from flat.

[John TenEyck]

In order for the OP to use this technique, he would need a sheet of paper that is 7 feet long. remember the alleged discrepancy is over a very long distance. A foot long straight edge is of no use.

One easy way to measure linearity is to tightly stretch a string on the diagonal above the surface that is very accurately positioned on both ends and then use a ruler along the string to measure discrepancies. Repeat the procedure on the opposite diagonal and that will give you a pretty good idea of how far off the surface is from flat.

W/o trying to sound like a smart axx, they make paper a lot longer than 7 feet. I always have a roll of paper in my shop that's about 3 feet wide. I use it for all kinds of tasks from covering my workbench to making full size layouts, to covering the floor in my temporary spray booth.

Everyone needs a long straight edge; in fact, several of different length, and there's no need to spend money for them when you don't need micron accuracy. I have one, for example, that fits perfectly inside door jambs. I made it as I described above, and it's off by an amount I can't measure which makes it plenty good enough for many woodworking uses. I'm sure your string technique works and might be the right tool of choice for some applications, but a straightedge is just a lot faster and easier to use most of the time.

John

[Michael Kellough]

Is the high spot on the side you put on the floor when gluing or is it the part that was under the books?
if it is the part under the books, was there an equivalent high spot in the "stuffing" before you placed the top?

if the high spot is in the side that was on the floor, is there an equivalent low spot in the floor?

[glenn bradley]

This is a good case for asking the ultimate first question; "what happens if I do nothing?" Once you have decided that you must do something, there are a good number of suggestions here. Speaking as someone who shims things to get them level in his own shop I can relate to having no empirical reference surface on which to build a flat surface to build things on. Nudging a 4x6 piece of MDF around on a shimmed reference surface is an opportunity for something to go slightly wrong.

As for surfacing MDF with a router, yes you can but, I suggest you do so on a scrap first and see what finish you might use to "reseal" the exposed MDF to a degree that you are happy with. MDF does not come "sealed" but the process does leave the outer surface nice and smooth as opposed to say, the inside of a dado cut in the same material. My MDF workbench is just BLO'd and paste waxed. I have beat on that surface for years and it still look pretty decent so, a penetrating finish and some elbow grease can make things very workable.

[Paulh Tremblay]

One question just for my knowledge-

Does the bottom perfectly mimic the top? ie, a slight dip?

Good question. When I initially assessed the flatness, it seemed dead on. When I flipped the box, I discovered the lack of flatness. The box weighs close to 200 pounds, so it it not easy to flip it to test my initial assessment.

[Michael Kellough]

Good question. When I initially assessed the flatness, it seemed dead on. When I flipped the box, I discovered the lack of flatness. The box weighs close to 200 pounds, so it it not easy to flip it to test my initial assessment.

Simple, turn it back over and be done.

[Brian W Smith]

Maybe an outer frame that allows it to pivot.......faceframe clamps on one side,other side gets used for assembly or whatever?....or not.

If not wanting the flip table,I'd run it crown up for a few months,then check it again.You can check the plane with wires in an "X" pattern.Use gage blocks on corners.Best of luck.

[John T Thomas]

I just made my own torsion box top (36 x 48) and am having a similar issue.

I thought everything was flat, but when I check with a straightedge, there is a cup in the middle of the table that varies between 0.01 and 0.02.

The table is constructed of 1/2" MDF grid and skins, and then another 3/4" melamine top with two t-tracks. I haven't mounted the melamine yet, but I plan on running a couple of strips of masking tape down the center of the MDF as a shim. I figure a couple of pieces should bring the whole surface to <0.01, which I would think should be sufficient.

[Mort Stevens]

Several posters question my need for such a flat surface....A few practical needs include placing a table base on it to test if the table would wobble, or gluing up 3/4''' strips of plywood, face to face, to make a table top, and wanting the result to be flat.

Take your straight edge and lay it on any floor in your house and tell me if it's perfectly flat? Trust me, the framers didn't shim the floor joists or plane the subfloor or knock down the high spots on the foundation wall to be within thousands of an inch! ;) Sometimes we get caught up on things that don't matter in the real world.