Why the flange comes back
Bending a strip does not put the whole thickness past yield. The fibres furthest from the neutral axis go plastic and take a permanent set. The ones near the middle only ever stretch elastically, and when the load comes off they try to straighten again. What you get is a tug of war the elastic core partly wins, and the flange opens by a few degrees.
Everything follows from that. Thicker material at the same radius puts more of the section past yield, so it springs back less. Stronger material yields later, leaving a bigger elastic core, so it springs back more. A larger radius bends the material less severely, again leaving more elastic material, so it springs back more. That is why the same tooling gives a crisp result in half hard mild steel and a stubborn one in 6061-T6.
Two ways to get the number, and only one of them is trustworthy
The measured route is a ratio and nothing else. Form a strip to a known angle, measure it after release, and divide the second by the first. That is the springback factor, and to hit a target you divide the target by it. Form to 90 and measure 87.5 and the factor is 0.9722; a 90 degree target then needs 92.57 degrees of forming, an overbend of 2.57 degrees.
The estimate route uses the classic relationship between yield strength, modulus, thickness and radius. Write the ratio radius times yield divided by modulus times thickness, then the factor is four times that ratio cubed, minus three times the ratio, plus one.
| Case | Ratio r Y / E t | Springback factor | Overbend for a 90 deg target |
|---|---|---|---|
| Mild steel, 45 ksi, r = t = 0.075 | 0.001552 | 0.99534 | 0.42 deg |
| Mild steel, 45 ksi, r = 4t | 0.006207 | 0.98138 | 1.71 deg |
| 6061-T6, 40 ksi, E 10,000 ksi, r = 2t | 0.008 | 0.97600 | 2.21 deg |
| 304 stainless, 42 ksi, r = 4t | 0.005793 | 0.98262 | 1.59 deg |
Those figures are what the formula gives for the inputs beside them, not a table of what your material will do. The formula assumes pure bending of a uniform strip. A press brake is not pure bending: the die shoulders drag, the ram stops somewhere specific, and the sheet has a rolling direction. Use the estimate to know roughly what to expect on the first strip, then throw it away and use the measurement.
The radius opens too
Springback is usually discussed as an angle because that is what gets inspected, but the corner opens as well. The arc of material at the neutral axis does not change length when the load comes off, so if the angle gets smaller the radius has to get bigger by the same ratio. A 0.075 inch radius that springs back from 90 to 87.5 degrees ends up near 0.078.
That matters for two reasons. It means a part specified with a tight inside radius may not be able to hold it by air bending at all, no matter what the punch looks like, because the released radius is always larger than the loaded one. And it means the K-factor you measured on a part includes the released geometry, which is another reason a borrowed K-factor rarely fits.
Which angle the flat pattern uses
A frequent mix-up. The blank is cut for the finished part, so the bend allowance is worked at the target angle, not at the overbent angle. The overbend is a machine setting that exists for a fraction of a second while the ram is down. The material does not get longer because you drove deeper; it comes back to the target and the neutral arc is the target arc. Cut the blank from the bend allowance calculator at the finished angle and leave the overbend to the press.
Killing springback instead of correcting it
Overbending is the cheap answer and it works, but it has to be reset every time anything changes. The alternatives trade force for repeatability. Bottoming closes the punch into the vee so the material takes the shape of the die, which cuts springback substantially. Coining goes further and squeezes the corner until it conforms to the punch nose, which very nearly eliminates it and costs something like eight times the air bending force. Both of those are covered by the press brake tonnage calculator, and both need the tooling to be rated for it, which is the tooling maker's figure and not one this page will supply.
Questions people ask
How do I calculate springback in sheet metal?
The honest way is to measure it. Bend a strip of the same material on the same machine with the same die to a known angle, measure it after the ram lifts, and divide the released angle by the formed angle. That ratio is the springback factor. To hit a target angle you divide the target by the factor and form to the result. Everything else, including the yield and modulus formula, is an estimate that gets you close on the first piece and then defers to the measurement.
Why does aluminum spring back more than steel?
Because springback tracks the ratio of yield strength to elastic modulus, and aluminum has an unhelpful combination of the two. 6061-T6 yields at a similar sort of stress to a mild steel, but its modulus is roughly a third of steel, so a much larger part of the section is still elastic when the bend is formed and there is far more stored energy to unload. The same reasoning explains why high strength steels spring back more than mild steel even though the modulus is nearly identical: the yield went up and the modulus did not.
Does springback change with the die opening?
Yes, and more than most people expect. In air bending the inside radius is set by the die opening rather than by the punch, and a wider vee makes a larger radius. A larger radius means a milder strain through the section, a bigger elastic core, and more springback. So the same part on a one inch vee and a two inch vee will need different overbend corrections, which is why a springback figure is only meaningful alongside the tooling it was measured with.
My part sprang back the right amount yesterday and not today. What changed?
The usual suspects, roughly in order. A new coil or a new heat of material with a different yield. Thickness drifting within its own tolerance band, which is allowed to be several percent. Bend direction relative to the rolling grain, if the parts are being nested differently. Tooling wear rounding the die shoulders, which effectively widens the vee. And temperature, since a cold shop and a warm one are not quite the same material. None of these are faults; they are why the first part off a setup gets measured.
Can I get rid of springback rather than correcting for it?
Largely, at a cost. Bottoming the punch into the die makes the material take the die shape and cuts springback a long way; coining squeezes the corner until the inside radius conforms to the punch and nearly removes it. The bill is force, roughly four times air bending for bottoming and around eight for coining, plus tooling matched to the angle. Whether your press and your tooling are rated for that is a question for the machine plate and the tooling maker. On thin work a small stiffening feature, a rib or a hem near the bend, also stabilises the angle without any extra force at all.