How Springback Affects Pressed Components

Springback in pressed components is the most predicted and least measured variable in press work. A bracket leaves the press at 88.6° instead of 90°. The 1.4° looks harmless on the drawing and becomes a 1.6 mm hole offset at the far end of a 60 mm flange, which is enough to stop a bolt going through on the assembly line. It is also the variable that most often gets "corrected" by grinding a die until the sample fits.

Springback is not a defect in the material. It is the elastic part of the forming operation coming back, and it can be calculated, measured and designed around. What follows is the ratio that predicts it, worked angle estimates for the common grades, the radius effect that makes a tight bend recover less than the table suggests, and the four compensation methods with their costs.

Why does a pressed part spring back at all?

Because only part of the section yields. Bending a sheet puts the outer fibres in tension and the inner fibres in compression, and the stress falls to zero somewhere in between. The yielded skin is permanently deformed; the elastic core around the neutral axis is not. When the punch withdraws, the core unloads and pulls the skin back with it, so the bend angle and the bend radius both change.

The amount of recovery is set by how much elastic strain the material was carrying at the moment of unloading, which is roughly yield stress divided by elastic modulus. That single ratio explains almost every comparison between materials, and it explains why a harder temper springs back more than a soft one of the same alloy.

It also explains why press force is nearly irrelevant in an air bend. Once the bend is formed, extra tonnage deflects the machine and the tool rather than the part. The exceptions are bottoming and coining, where the punch drives the material against the die face and raises the stress through the full thickness, leaving less elastic core to recover.

How much springback should you expect?

Use the thin-section estimate first, then correct it on the first article. For a bend radius at least five times the thickness, the recovered angle is close to 3σy R / (E t) multiplied by the bend angle in radians. It is an estimate, not a substitute for tryout, but it ranks materials correctly and gets the order of magnitude right.

Work the CRCA case. A 90° bend in 2 mm CRCA at a 10 mm inner radius, with a 240 MPa yield and 210 GPa modulus: 3 × 240 × 10 / (210,000 × 2) = 0.0171, which is 1.7% of 90°, or 1.5°. Shop measurement on an air bend usually lands within half a degree of that, and the die is overbent to suit.

The aluminium rows are the ones that surprise people. 5052-H32 has a yield close to CRCA but a modulus a third of steel's, so it carries about two and a half times the elastic strain at the same stress. That is why an aluminium bracket made on a steel tool comes out under-bent by 2–3° and why the same tool, reused for 6061-T6, is out by about 3.9°.

Treat these numbers as the shape of the answer, not as a table to quote. Real recovery depends on the die opening, the punch radius, friction, the blank-holder or pad force, and whether the bend is air-bent or bottomed. Anything within 30% of the estimate is normal; the tryout is what settles it.

Why is springback smaller at a tight radius?

Because the elastic core is smaller, not larger. Springback is driven by the elastic strain the section carries at unloading, and the estimate Δθ = 3σy R θ / (E t) rises in direct proportion to R/t. At R = t the outer fibre reaches about 50% engineering strain, almost the whole section has yielded, and only a thin elastic core around the neutral axis recovers. At R = 5t the strain gradient is gentler, a larger share of the wall stays elastic, and that core pulls the section back further.

The practical consequence is that a table built on R = 5t over-predicts a 1t bend by about 5×. A 90° bend in 1.5 mm CRCA at a 1.5 mm radius recovers about 0.31°, where the same material at a 7.5 mm radius recovers about 1.54°. The tight radius is the safer angle and the riskier crack: at R = t the outer fibre is at the forming limit, so the part is more likely to split than to spring back. If the drawing calls a tight radius for a functional reason, check the strain against the forming-limit curve before worrying about the angle, and put the overbend where the radius is generous.

There is a second radius effect that catches designers. Springback changes the radius as well as the angle: the released radius is larger than the tool radius by roughly the same percentage as the angle recovery. A 10 mm tool radius that recovers 1.7% comes out at 10.17 mm, and if the part sits on a shaft or a pin, that 0.17 mm is a fit problem rather than an angle problem.

How do you compensate without cutting the tool twice?

Four methods, in ascending order of cost and descending order of sensitivity.

Overbend the die

Cut the punch and die so the part is formed past nominal by the measured recovery. It costs nothing after the first cut if the correction is known before the tool is made, and it is the default answer for one or two bends. Its weakness is that it is tuned to one material, one thickness and one die opening; a coil change moves the result.

Correct the radius as well as the angle

Set the tool radius smaller than the finished radius, so the part lands on nominal after release. On a 10 mm radius with 1.7% recovery the correction is 0.17 mm, which is a grind-and-polish job on the punch nose and is worth doing where the radius is functional.

Bottom or coin the bend

Drive the punch into the material at the bottom of the stroke so the bend is stressed through the section. Recovery drops sharply, and the angle becomes much less sensitive to thickness variation. The cost is tonnage: the punch now presses the full flange area rather than shearing a line, and the die must be strong enough to take it. On a 2 mm bracket, a bottoming pad can add the equivalent of several tonnes to a press that was sized on the blank.

Add a restrike station

Form slightly under, then restrike the bend in a second station with the correct angle. A restrike takes the material variation out of the result because the second hit sees a part that has already sprung back once. It is the standard answer for anything above about 3° of recovery, and the usual choice for aluminium and high-strength steel.

A fifth option sits outside the tool: form in tension. Stretching the material while it is being formed, as in stretch forming or a draw with a high blank-holder force, raises the stress through the section and reduces recovery. It suits large panels with gentle curvature and is not practical for a bracket with a 20 mm flange.

Where does springback show up after painting?

At the cure oven. A powder coat cures at 180–200 °C for 10–20 minutes, and that hold relaxes a fraction of the residual stress locked into a formed part. The shape moves toward a lower-energy state, usually by a few tenths of a degree, and occasionally by more on a part that was formed close to its limit.

Two consequences follow. First, an angle that is functional, such as a mounting face or a rail engagement, should be measured on a coated sample, not only on a bare one. Second, parts should be supported through the oven rather than stacked, because a part that sags at 190 °C keeps the sag.

The same relaxation explains why a panel that measured flat on the bench looks wavy after cure. It is not the coating thickness; it is the residual stress finding a new equilibrium at temperature. A crown of 1–2 mm across a 500 mm panel is the usual design answer.

How should springback be measured and recorded?

Measure the free part, on a granite table, with a digital angle gauge or on a CMM, and state that condition on the inspection plan. A part clamped flat on a fixture reads differently, and if the customer measures clamped while the supplier measures free, both are right and the argument is unwinnable.

Record three things at first article:

• The angle at each bend, with the tolerance from the drawing and the measurement method.
• The flange length or hole position that the angle controls, because that is what the assembly sees.
• The material batch and the die opening, so a later change can be traced.

Then check one part per shift at the start of a run rather than only at first article. Springback drifts with die wear at the radius and with coil-to-coil yield variation, and a drift of 0.5° across a shift is normal. The bend allowance method converts the angle and radius into the flat pattern, and the K-factor guide explains where the neutral axis sits for each material.

How Ganesh Multitech handles springback in pressed and formed parts

Springback is dealt with at the tool design stage, in a tool room that builds the press tools, the progressive dies and the checking fixtures in-house. A draw or form tool is tried out against the drawing, the recovered angle is measured on the free part, and the correction is made to the punch or die in the same building. Checking fixtures are built with the tool, so the first article is measured the way the production part will be measured.

Parts are pressed on mechanical presses from 20 T to 300 T with 250 T punching, and formed on the 200 T 5-axis CNC press brake with a 4,450 mm bed and Delem control, rated to 10 mm thickness. Where an angle is functional, the restrike or bottoming decision is taken before the die is cut rather than after the first rejection.

The plant runs ISO 9001:2015 and ISO 14001:2015 across five units in Mumbai (Vasai) and Nagpur, has pressed and fabricated since 1992, holds ±0.5 mm typical on formed features with tighter callouts by agreement, and offers first article inspection and PPAP documentation. Coating, where the part needs it, runs through seven-tank pretreatment and an 18 ft × 14 ft powder booth, so the post-cure angle can be checked on a finished sample.

When springback cannot be designed out

Some parts should not be pressed at all. A part with five bends in five directions, each holding ±0.2°, is a brake job with a formed-part inspection plan, not a die-casting of a press tool. A part in 6061-T6 with a 1t radius will crack before it springs back, and the material choice is the problem. A part whose angle must hold across a 3 mm thickness range in the same tool needs a restrike, and the tool cost has to be accepted or the design changed.

We do not run presses above 300 T, we do not heat-treat or anneal after forming, and we do not straighten a formed part by hand to meet an angular tolerance. If the angle cannot be held in the tool, the answer is a drawing change or a different process, and the quotation will say so before production starts.

What to put on the drawing

• Each bend angle with its tolerance, and the angular tolerance class from ISO 2768-1 where the callout is general.
• The inside radius, and whether it is functional or nominal.
• The flange length that the angle controls, and the hole positions measured from it.
• The material grade and temper, because recovery is a property of both.
• The measuring condition: free or clamped, and after coating if the part is coated.
• The features that must stay flat, with a flatness or oil-can limit and the gauge.
• The bend sequence if the part is formed on a brake, because the order changes the accumulated error.

Send the drawing with the bend schedule and a photo of the released part on the gauge through request a quote. The stamping and press tools hub carries the companion guides on flat pattern, tooling and inspection, including the bending tolerance and inspection method used to accept a formed part.