I have some steel tubes (18 mm OD) which I wish to use as horizontal members in a storage rack (a bit like shelves). However, the tube flex and bend quite a bit when I put heavier boxes on them. The tubes flex about 40 mm over a 1000 mm span. Ideally I would like the flex to be <5 mm.

Is there any cheap(~$0) & readily available material I can put into the tubes which would stiffen them significantly.

I was thinking about filling them with concrete, but that seems a bit messy and I don't have cement/sand to hand. Would concrete work?

  • 1
    Steel tubing does not have much lateral stiffness; it's not really a good fit for this type of use. – TomG Jan 21 '13 at 23:33
  • Are you referring to conduit tubes? If so, perhaps swap them out for thicker walled steel pipe instead. – DA01 Jan 22 '13 at 3:43
  • from all the answers, it sounds like there is not that much that can be done within my constraints. I was hoping to use the material I had lying around, but I guess I'll have to spend $ on stiffer tubes or some other beam solution. – Ken Jan 22 '13 at 13:50
  • 2
    Why not just support the tubes with shelf brackets like one does with long wood shelves? – Alaska Man Mar 10 '17 at 4:16
  • Could try placing solid rod inside of tube, wood or steel, to stiffen. – crip659 Feb 24 at 12:03

18 Answers 18


You could truss the tubes as shown in the picture below. The center block helps to support the tube when the truss member is placed in tension.

enter image description here

The truss could be constructed from heavy wire or threaded steel rod. The concept here would be similar to the scheme used on the wider steps of a wooden step ladder.

I would have suggested the possibility to place the truss wire / rod inside the tube but I suspect that the 18mm diameter of the tube is too small to let the center support in such scheme be large enough to make a significant impact.

  • Thanks @Micheal, good idea, but I don't have the clearance, below the tube to truss it. Clearance is part of the reason in need to stiffen the tubes. – Ken Jan 21 '13 at 14:52
  • @Ken -- You would then likely get a lot better result by removing the tubes entirely and then laying in pieces of 19 or 20mm thick plywood that spanned the same distance as the rods. Another alternative would be to simply lay the plywood on top of the tubes. If doing that make sure the plywood is at least the full 1000mm dimension so it overlays the same supports that hold up the tubes. – Michael Karas Jan 21 '13 at 16:39
  • @Michael, part of the reason I'm using tubes is to minimise the visual appearance of the shelves when unloaded. The plan is to use just two tubes (e.g. EMT) across the span Positioned to support fixed sized boxes at 1/3 and 2/3 "box widths" from the wall – Ken Jan 21 '13 at 17:00
  • Your next best bet may be angle aluminum or steel about the same size as the tubing; it may make the tubing redundant, though; it may be more visually appealing than plywood. – TomG Jan 21 '13 at 23:36
  • I reckon despite OP's comment it should be possible to make truss inside the pipe and to get enough support as long as the string/rods are tensioned enough and the tube is prevented from turning. Nonetheless it might be more practical to go for other options – user377178 Oct 22 '20 at 8:20

If you can't use the truss technique described by @michaelkaras, the only real solution would be to move to thicker-walled tubing or larger diameter tubing.

Anything you fill the tube with is not likely to make much of a difference at all. All the bending strength of the tubular member comes from its topmost and bottommost elements. The center of the tube does very little to resist bending.

And as others have pointed out, concrete is lousy in tension, which is present on the bottom half of the core, so you'd be likely to develop cracks in your concrete core, rendering it useless.


Scallop fishing, many years ago, we had a problem with the club stick bending. The club stick in question being a piece of 10 foot by 4inch schedule 80 pipe whose primary function was to help dump the bag full of scallops once brought aboard. We welded flat stock on the the outside of the pipe (opposing the direction of the bend). Sort of a variation of the truss system mentioned in previous comments. It did not work. We then tried welding a truss system using steel rod to the pipe hoping it would be more robust. Didn't work. Finally we simply filled the pipe with concrete and that solved the problem.

  • I'm impressed that you were able to wield a 10' long, 4" diameter pipe filled with concrete. I'd have thought that would have weighed a bit! – FreeMan Oct 15 '20 at 12:34

I don't think there's really a solution that's cheap and will be acceptable. You say you're looking to reduce the deflection to about 1/8th, which is a large difference. Filling the tubes with concrete will increase their stiffness and therefore reduce deflections, but not 8x. Michael Karas's truss idea is also not going to make an 8x difference.

I think you're going to need to be creative / look further to find a solution to this problem:

  • Deeper / thicker walled tubes
  • Shorter span (add intermediate supports, hang wires from the ceiling, etc.)
  • Reduce the load on the shelves (you haven't described the project, but maybe you could make one shelf "heavy duty" and the others lightweight)
  • I would beg to differ with your opinion regarding the trussing of the tubing. If the truss is made directly from the bottom of the tubing it would surprise you how much stiffness that would put into the tubing span of one meter. The greater the height of the center support block the more support you can achieve within the limits of the tension strength of the truss rod or wire. I would invite you to take a look at how strong various types if welded wire shelves are when triangulated with rather narrow vertical height. – Michael Karas Jan 22 '13 at 3:20
  • I do think the tension wire/rod could add a lot of strength — maybe doubling or tripling it. But he's looking for almost an order of magnitude increase in stiffness. Even if he cut the span in half I don't think he'd get the deflection he desires. – Hank Jan 22 '13 at 18:39

The steel tube would be strengthened a great deal by filling it with mortar (Type S might be best). This would add compressive strength that would stop the tube from collapsing under a load. Can't say whether or not it would fully meet your goal. You'd have to try it. Concrete contains larger aggregate, and would be too difficult to load into a small tube.


Assuming you can fix each tube end to a rigid frame, you could achieve what you want with help of tensile force. E.g. make each tube slightly short and ensure fixings at each end pull the tube apart. Tube should work sort of as string in a tennis racket.

  • That would be interesting to try. – Greg Nickoloff Oct 15 '20 at 17:43

I think plain concrete would work, but whatever you end up using, you could fill the tubes with a caulk gun and fillable caulk tube to cut down on the mess.

  • good idea, I'll do that – Ken Jan 21 '13 at 14:54
  • 3
    I'm not too sure about concrete. It's pretty crappy stuff in tensile loading or bending (which is half tensile). – Chris Cudmore Jan 21 '13 at 15:53
  • I am also fully uncertain about the idea of filling the tube with concrete. It will do next to nothing to make the tube sag less under load. – Michael Karas Jan 21 '13 at 16:27
  • My (naive) thinking was that the package would form a steel reinforced concrete beam. – Ken Jan 21 '13 at 16:55
  • 1
    Right, steel reinforces concrete. Not the other way around. – Philip Ngai Jan 21 '13 at 19:17

Expanding on The Evil Greebos' idea, I would try aerosol expanding foam (aka Great Stuff). You may have to extend the the little extension tube with some tubing to reach the center. It cures to a somewhat ridgid state. It is also lighter than chaulking.

  • 6
    The aerosol foam will do next to nothing for making the tubes more rigid in the application described by @Ken. BTW, EvilGreebo was not suggesting to fill the tubes with caulk. Instead he was suggesting to fill re-usable caulking gun tubes with wet mortar mixture to then be pumped into the tube with a caulking gun tool. – Michael Karas Jan 21 '13 at 16:32
  • @MichaelKaras is probably correct here. However, the foam WOULD serve SOME purpose, by reducing the tendency of the tubing to flatten under load, which would greatly reduce its load carrying capacity. This trick is how one can bend copper tubing (without causing the tube to flatten) by filling it with ice. – user558 Jan 22 '13 at 13:39
  • 1
    @woodchips - The expanding foam I've played with was easily compressible. It is after all full of bubbles. I cannot agree that this would do all that much to keep the OPs EMT conduit from flattening. Also consider that by the time his conduit has bent so far that it has started to flatten his support concept has already failed. --- On another point you do kind of place doubt on your point by using the scenario of bending copper tubing simply by noting that it can bend without flattening!! – Michael Karas Jan 22 '13 at 13:53
  • @MichaelKaras - yeah, I was going to add that point about being able to bend the tubing at all, and I agree that adding foam is of little value. I was merely saying that it would help a little. Not much though. – user558 Jan 22 '13 at 15:37

I once read that a packed pvc pipe with sand worked very well.


how about polyurathane foam used to fill gaps around brick walls? it has a high expansion ratio that will easily fill your tubes, it grows in a quite stiff material and it also has great bonding qualities which means that it will stick all the way on the inside walls of the tubing so it will not allow sliding of one material on the other increasing its overall stiffness by working as a whole. I quess that it will not make the difference you require but some of that.


Pipe in a pipe. Lightweight bicycles often have Balsa wood inside the magnesium pipe frames to add rigidity. Anything from wood to bar stock would work.


I have used concrete to fill vertical fence tubes , worked very well . It works by preventing collapse ( in the elastic stress range ) , ie .the concrete is in compression. This has been done in oil well casing that has exceptionally high collapse loads. A second casing is run concentrically and cement is put in the annular space. Also , I have put steel rebar into 1/2 conduit, you want a close fit, very heavy , but strong.


I'm facing a similar problem, and agree with the type-s suggestion. Have you seen barriers (often to protect things like gas pumps) made of steel posts filled with concrete? My theory is that when a tube bends, its internal volume is reduced. If you fill the tube with a material that can't compress, it will resist bending. My situation involves a patio chair made of 1" square tubing, which rusted through and finally bent. My plan is to fill the void with either anchoring epoxy or JB Weld and then insert a steel rod.


I am facing a similar problem. I would suggest you try bending the tube virtically upwards by tieing its both ends with a strings as in a bow. This technique is similar to the basic design of suspension bridges. If you are using only tubes without fixing a flat sheet on them, you will have then to live with placing you stuff on two curved tubes.

  • 1
    Welcome to DIY.SE! Can you edit your Answer to provide a bit more information about your solution, maybe an image too? – mmathis Jan 5 '17 at 16:23
  • This is, effectively, a truss as suggested in by Michael Karas 4 years before this answer was posted. I don't think it adds much... – FreeMan Oct 15 '20 at 12:36

First, I would switch to 3/4" rigid conduit or black pipe. If that isn't stiff enough, then consider using 1/8" (or 1/4") thick x 1/2 width flat metal welded as an 'I-beam'. flat cold rolled bar is inexpensive and as an I-beam would have good span characteristics. See http://iris.nyit.edu/~maltwick/BC2/Steel%20Rules%20of%20Thumb.pdf for span strength estimates.


I have read all the questions and answers and have come to the conclusion which I don’t think anyone has thought of which is using a combination of re bars and expanding foam to hold it in place inside the tubes

  • This is effectively the same thing JEFF RAJCA posted 3 years ago. The spray foam won't do anything different, as it will simply crush when the bending force is applied to the outer tube. – FreeMan Oct 15 '20 at 12:37

OP inquired about stiffening a thin wall tube. It was flexing too much. Much of the discussion went into how to strength the tube which is not the answer he was after. The formula for the center displacement of a simple supported beam with a center load:

Y = P*L^^3/(48*E*I)


  • Y is the displacement
  • P is the center load
  • L is the total span
  • E is the elastic modulus
  • I is the moment of inertia of the section

Looking at the formula will give you the solution. Reduce the span will do a lot of good. Changing to titanium will increase E more than 3 folds, but hardly a cheap solution. A tubular section is most efficient in getting a high I value. Thicker wall and bigger diameter put you in the right direction.

The most popular suggestion here is filling the tube with concrete. This no doubt will get you a stronger member that is capable in flexing more and take more load without bucking the top or compression side of the tube.

What about displacement? It is inversely proportional to E*I. The E value of steel is about 10 times of concrete. Let's say the I value of the tube is roughly the same as the hollow space or the concrete fill. The increased E*I is therefore 10% after the concrete fill, and reduction in displacement is roughly 10%. It will not be a reduction of from 40 mm to 5 mm OP was looking for.


Fill it with concrete - here's a scientific paper about doing that: https://www.hindawi.com/journals/amse/2018/9065378/

They were getting less than 1mm flex from 4000kg on their 1.2meter span. (see table: https://www.hindawi.com/journals/amse/2018/9065378/fig7/ )

Concrete is not compressible, and since it's inside a steel tube, the only way that combination can flex is for the concrete to crack and for all the steel on the sides and bottom to stretch as well - and since I'm assuming you're probably not putting 4 tons on the middle of that shelf, that's never going to happen.

  • 2
    This "fill it with concrete" (about the 10th answer to say that) at least points to some supporting evidence. Please take a minute to edit your post with a few relevant quotes from the linked article, as links tend to die an an internet instant. Having the relevant quotes here will allow this answer to stand on its own even should that happen. – FreeMan Oct 15 '20 at 12:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.