The extrusion shape is a simple square tube. If I set it down on a perfectly flat surface with its curvature upwards, the middle raises up by around 5mm over the length of 1m.

So far I tried to just put a weight on it while its laying on a glass pane (my perfectly flat surface) but after 24h that did just about nothing.. maybe if I left it there for a year? I'm not that patient.

Are there any other options for me to straighten it without specialized tools? I could heat it up with a torch if that helps..?

  • 2
    $\begingroup$ In industrial circumstances, they are stretched while cooling, I believe - also not very practical for you. Have you tried setting the extrusion between two supports and flexing the centre down below 'flat', such that it springs back to the desired position? You will need to gradually push 1mm further every time until you find it is deforming permanently $\endgroup$ – Jonathan R Swift Apr 17 '18 at 7:00
  • $\begingroup$ @JonathanRSwift I can try that, going to have to rig something up to push it gradually $\endgroup$ – user81993 Apr 17 '18 at 7:06
  • $\begingroup$ You could also gradually increase a weight hanging on it, rather than gradually increasing the displacement? $\endgroup$ – Jonathan R Swift Apr 17 '18 at 11:42
  • $\begingroup$ Use a jack or screw thread to cause the deflection or, put the aluminium on two big blocks and lower a car onto it : worked for me in the past, careful with the lowering though... $\endgroup$ – Solar Mike Apr 17 '18 at 13:10
  • $\begingroup$ Seem to remember a technique which involves repeated heating the concave side with a torch. The differential expansion of the hot side vs cold side eventually corrected the curve... $\endgroup$ – BobT Apr 17 '18 at 17:45

Which alloy and temper? You have to overcome the yield strength to get the part back to shape. Laying weight on it will only waste your time. The part has to be bent in the opposite direction of the bow, but only enough to get past the yield strength. If you're not familiar with yield strength. The two most important mechanical properties in aluminum is the Yield strength and the Ultimate Tensile Strength. When aluminum is deformed, it will spring back, up to a point. Aluminum will reach a point of stress where it will not return to it's previous shape, it "yields". After aluminum yields, it will stretch a certain amount, depending on the alloy/temper, and then it will break. The ultimate strength is a measure of the force it took to break the sample. Well there is another property called elongation, it is basically the length the part stretched before it broke, think final length subtracting the original length, as a percentage of the original length. It's the "distance" between the Yield and the Ultimate. We deal with 3xxx and 6xxx alloys. 3003 O-Temp will elongate up to 60% before it breaks, where a 6061 T9 might only stretch 5% before it breaks.

If you have a 6xxx alloy, depending on the temper, the yield strength could be anywhere from 9,000 lbs to 45,000 lbs.

If it's a 3xxx series (unlikely), the yield would be somewhere between 5,000 lbs to 25,000 lbs, again, depending on the temper.

You could try heating it up. The yield strength drops considerably with temperature. Heat just the the area of the bow with a torch, you're going to need hot mitts, then carefully pull against the bow, don't over-do, if you kink it, it's done. Don't get too much of the part hot, or you end up with multiple bows. I wouldn't go over 400f, you'll blow up the grain and end up with a brittle part.

another edit, a poster below had a good idea with heating up the part. You might try heating the concave side of the bow, pretty hot, but you need to cool it off quickly. Aluminum shrinks as it cools. We see this in extrusion, the parts come out at 1,000f and if one side of the part is thinner than the other, that side will cool, and shrink faster, causing our extrusions to bow. But we have stretchers to remove it.


About the only simple way ( no heavy equipment) is to grip each end and pull it to a small amount of plastic strain . As a estimate it would begin yielding at about 0.5% elastic strain , then pull it another 1 % plastic strain . So a total strain of 1.5 %. This would still require some equipment . The load would depend on the yield strength of the material. I do this with copper and aluminum wire occasionally; but wire like 16 gage only requires a vice and vice grip pliers.


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.