BOUNCE FACTOR: Why cars rebound after crashing

Coming up: We crack the code of crashes, use that mad science to explain why stamping sheetmetal is black magic, and teach you how to become a Jedi master of nuts and bolts

Free audio.jpg

Last episode >> I asked you to explain why cars rebound off the wall in crash tests. After all, the car is designed to crumple, right? The crumpling absorbs the kinetic energy. So if the energy gets absorbed, why the bounce?

 The post-crash bounce: Why does it occur?

The post-crash bounce: Why does it occur?


A few people got it dead right. Kudos to you. The answer is coming up. Some people got it a bit right. And of course I got lots of interesting answers as well:

"Magic pixies stroking their miniature unicorns?" - Tom Cat
"The power of Jesus is pushing the car off the crash barricade." - Brian VeeStrom
"It’s all the junk in the trunk. that usually gets me sprung." - to whom it may concern

Here’s the silver medal for outside-the-box thinking:

"The radiator, having being crushed and split emits water vapour forwards that then thrusts the car backwards. All this on only 5 glasses of red." - Benny Banger

And, my personal favourite:

"Car jumps backwards bc it was surprised by the crash. What’s my prize?" - Jack Low

Realistically, that answer is its own reward.


I used to work in a destructive testing laboratory - the perfect opportunity to be loud, violent and brutal … only in a good way. We broke all kinds of stuff. It was quite therapeutic. We had this huge tensile test rig, and we used to break steel samples endlessly, in the name of keeping the bastards honest. It taught me a lot about the behaviour of steel under load.

Steels 1.jpg

ABOVE: Here we’ve got stress versus strain. (If you’re just getting out of the blocks on this, think: 'Load' versus 'deflection'. As you ramp up the load, the steel deflects more, then it breaks.

ABOVE: Note the elastic region - when you remove the load it returns to its original shape. But when you move past the yield point, any further deformation becomes permanent. That’s called ‘plastic’ deformation. The summit of the mountain is the ultimate tensile strength, and when the line ends, the steel breaks.

So if this graph represents a bridge, loaded up to its ultimate tensile strength, I wouldn’t be stopping to admire the view, because a catastrophic failure is unavoidable…

More on tensile strength >>


Some people, of course, tried for a correct answer, and missed the target, emphatically.

"A deformable barrier on the test..." - iVlogBuzz

That does not explain the bounce - the deformable blue face is just a crushable aluminium honeycomb that’s there to simulate the crushable face of a car you have a clipping head-on with. The car would still bounce if all it hit was bare concrete.

"Equal and opposite reaction." - Peter harper

Newton’s third law was very popular in the answers - but this is true of all collisions. During the collision the force the car exerts on the block equals the force the block exerts on the car. It doesn’t account for the ‘why’ of the bounce.

"The concrete block will also have some elasticity." - Equiluxe1

Totally incorrect. The block is chosen specifically for its non-participatory properties. It’s completely unaffected by the collision for all non-trivial analysis.

"The more I look at it the more convinced I am that the front tyre is a major contributor to the bounce back." - Gabriel Rd

I guess it does push back a bit - but also only a trivial factor. The car would still bounce, about the same, if you threw it at the wall minus the tyres.

"Isn't inertia part of the energy that cannot be lost?" - Kevin Reid

No - it’s not - inertia is a property intrinsic to everything with mass. It’s like this: a two-kilo brick is weightless in space, right? But if you throw it at someone’s head - in the International Space Station, and it hits them, it’s still going to be deadly, potentially. The reason is: Inertia.

"Conservation of momentum." - David P

Clearly the momentum of the car is not conserved. It doesn’t bounce off at the same speed, and it’s moving in a different direction. So that’s absolutely not it. And yeah, conservation of momentum is universal (at least in Newton’s universe) but where you draw the boundaries for your system really matters.

Steel 3.png


Steel in daily use is designed to be operated in the elastic zone (I said 'plastic' in the video - sorry). I mean, it’s really no good if you drive over a bump, and the car yields, and now the headlights are looking permanently at the night sky, and there are ripples in the roof, right?

And that yield point we talked about earlier? You’ve felt that. At least you have if you’ve ever overcooked a bolt with one of these. Because bolts are all about clamping force. But there are limits to that.

Steel 7.jpg

Clamping force really matters when it comes to keeping wheels on cars, or screwing the head to the cylinder block. In fact, it matters anywhere there’s a torque specification.

When clamping force matters, and you’re a design engineer, you need to use the smallest fasteners you can, that will reliably do the job by generating enough clamping force. And that means you have to get close to the yield point without going over it. Because that vertical axis is also, essentially, clamping force.

But if you overdo it, you’ll cross the line into the plastic zone. You’ll be cranking it up, and you’ll feel the load come off the wrench and at the same time there will be a significant additional forward movement in the spanner as you take the bolt literally down the S-bend.

The bolt is now ... well, if you want the technical term, it’s ‘rooted’.


"Should by N.m/rad. Kg is mass." - BrettMcS

Everyone’s a critic. But Brett is right on the money, technically. But here’s my difficulty with that: If we’re all engineers and scientists - absolutely true. Black and white.

I did use kilogram-metres/radian in my earlier report >>

Newton metres is torque. Force times distance. Newtons times metres. Kilogram metres is nonsense - if kilograms are just mass. But if kilograms are force, kilogram metres is torque.

So my personal engineering hell is: I’m trying to make people who never studied applied physics understand it on a practical level. There’s absolutely a difference for the cognoscenti between mass and force, which is what Brett’s criticism is fundamentally about.

But in fact, here on earth, in a non-accelerating frame of reference, one kilo of mass exerts one kilogram of ‘force’. Kilos of mass and kilos of force are interchangeable in practice, in these basic physics situations.

Kilograms force is commonly used (although not part of the SI system) and explaining the difference not only confuses most people; it’s also a miracle insomnia cure, and generally it makes no difference.


Steel 2.png

When you’re manufacturing a car, much of the process involves bending metal - getting it into the plastic zone in a precise, repeatable way.

All those curvy panels start off as flat sheets in a big coil.

So we’re stamping a sheet of steel in a big hydraulic press on a multimillion-dollar die.

And we’re doing them by the thousands on a brain-bendingly precise surface.

This means we’re somewhere up in the plastic zone on the curve, then the press releases, and the steel has this amazing characteristic, where it unloads elastically. It springs back a little bit.

Steels 8.jpg

And if that’s not voodoo, I don’t know what is. It means that if you want to design these sheetmetal dies for a living, you’ll be taking the red pill - because the die has to be a subtly different shape than the panel it creates, because you need it to bend a little more in 3D space, to compensate for the springy release. The finished panel has to exactly the right shape to weld into the platform, after it springs back.

So the shape of the die is different from the shape of the panel it creates. And the people who design them are therefore wizards.

Hands up if you knew that?


Here’s the answer to why the car rebounds - the full answer, supported by applied science.

The bounce occurs because of the overall elasticity of the entire structure of the car. The bits that are smashed and permanently deformed, plus the bits that are not - like the passenger cell, hopefully, the better to leave you shaken but not shredded in the aftermath.

The entire structure is compressed by the massive inertial load of the crash. Some of it permanently and some of it not. I think we can all agree on that.

To some extent, the structure is like a giant spring. Some of the kinetic energy that is not absorbed by crumpling parts of the car permanently is stored in the elastic strain of the structure.

When the car slows down enough, its acceleration tapers off, the inertial load into the wall drops, and the defacto spring of the structure unloads.

That’s what pushes the car back off the wall. It’s important to understand that the bits that are permanently bent, and the bits that are only elastically deformed, both unload elastically - like springs.

So if you told me it was just the passenger safety cell unloading, you were absolutely on the right track with elasticity, but only halfway there.

The crumple zones stay bent, but when the inertial load comes off, they unload elastically too.

(Right top) The area under the curve is the energy that the structure absorbs, in the form of mechanical work, bending the parts.

Hopefully this happens in a controlled way, over the longest possible time window, so that you don’t die.

That’s where the kinetic energy goes, to keep the first law of thermodynamics happy about your crash.

But then (right, centre) as discussed, the structure unloads elastically, and it hands you this little triangle of energy back (right, bottom).

And that’s really what pushes you off the wall.


Just before I let you go: Some people took issue with Mad Metallurgy episode 1, when I alleged that humans were superior to chimpanzees, despite the 98 per cent overlap in our respective DNA.

"Please do not suggest that chimps are not as advanced as humans in evolutionary terms. Not one single chimpanzee anywhere on the planet believes that there is an invisible deity in the sky that spies on them all day and night (especially at night). Many humans are yet to reach this level of intellectual development." - Dear Leader

Yes - our species is obsessed with mortality on an individual level, and atrociously indefensible and unsophisticated mental gymnastics are devoted to attenuating this concern. Chimps just eat bananas and wage war on other chimps - because that’s how they roll.

"A chimp has never voted for Trump." - Aaron Smith

I disagree. I could prosecute the argument that everyone who voted for Trump was closer to a chimp than a human. Perhaps once again I am unfairly impugning the character of chimps. Trump voters are in fact chumps, not chimps.

And finally, an unexpected comment on biology and adaptation:

"When you say ‘bushes’, is that the same thing we call ‘bushes’ here in Retardistan?"

I have no wish to explain to every woman under 40 what a ‘bush’ is. Not on this fine, family show. Not really. It’s like having a thatched roof - we’ve moved on. We no longer do that.

Let’s just say, National Geographic recently added pubic lice to the endangered species list in advanced western countries. Thanks mainly to deforestation.

I’m wondering: Those lice: Is anyone truly sorry to see them go? That’s the question. Anyway, it’s all smooth sailing from here on in. So that’s nice.