Suppose a car moves with a constant speed of $20 \text{m/s}$ a quarter of a circle, and completes the quarter in $5$ seconds. One way to calculate the circumference is simply $20 \cdot 5 \cdot 4 = 400 \text{m}$. However, I know that $a=\frac{v^2}{R}$ and the circumference is $2 \pi R$. The acceleration is defined as $\frac{\Delta v}{\Delta t}$ and thus from Pythagorean theorem we get $a = \frac{\sqrt{20^2+20^2}}{5}=4 \sqrt{2} ~~ \text{m/s}^2$. Now to find the circumference we plug our result into $2 \pi \frac{v^2}{a}$, thus the circumference: $2 \pi \frac{20^2}{4 \sqrt{2}} \approx 444.29 \text{m}$ which is clearly wrong. So why the second way gave me a wrong result?


Acceleration is the derivative of velocity with respect to time, which is not the same as just dividing the difference in velocity by the difference in time.

$$a=\frac{dv}{dt}\neq \frac{\Delta v}{\Delta t}$$

$dv$ and $dt$ can be thought of as infinitely small $\Delta v$ and $\Delta t$. So if your $\Delta v$ and $\Delta t$ are quite small, the result of your calculation will be pretty close to the actual acceleration. But in your case, $\Delta v$ and $\Delta t$ are pretty big. Therefore, your calculation is way of. We can see that our result becomes more accurate if we consider an eighth of a circle:

$$\frac{\Delta v}{\Delta t}=\frac{ \sqrt{(v_{x0}-v_{x1})^2 + (v_{y0}-v_{y1})^2} }{5/2}$$

Now we plug in $v_{x0}=20$, $v_{y0}=0$ and $v_{x1}=v_{y1}=20\cdot\sqrt{\frac{1}{2}}$.

$$\frac{\Delta v}{\Delta t}=2\frac{ \sqrt{20^2(1-\sqrt{\frac{1}{2}})^2 + 20^2(0-\sqrt{\frac{1}{2}})^2} }{5}=40\frac{ \sqrt{1-2\sqrt{\frac{1}{2}} + \frac{1}{2} + \frac{1}{2}} }{5}=8\sqrt{2-\sqrt{2}} \approx a$$

And we use

$$a=\frac{v^2}{R} \Rightarrow R =\frac{v^2}{a}$$

$$C=2\pi R = 2\pi \frac{v^2}{a} \approx 2\pi \frac{20^2}{8\sqrt{2-\sqrt{2}}} \approx 410.46\space m$$

As you can see, this is much closer to the real value of the circumference. This is because we chose a smaller part of the circle, and because of that a smaller $\Delta v$ and $\Delta t$. If you pick smaller and smaller parts of the circle, your value will get closer and closer to the real value. And in the limit of choosing a part of the circle with length 0, it will be the exact value. This is exactly what the '$d$' in $\frac{dv}{dt}$ means.

You can also try choosing a larger part of the circle, and you ranswer will come out a lot worse. In the worst case, if you use the entire circle, you will find that $\Delta v=0$, and therefore $a=0$, and conclude that the circle has an infinite circumference.

  • 1
    $\begingroup$ Could the downvoter comment why the downvote? This seems to be a perfectly valid answer to me. $\endgroup$ – Ruslan Oct 7 '13 at 14:39

Why are you using Pythagoras' theorem? a triangle of two sides $R$ and one side hypotenuse $vt$ is a wrong depiction of $v$. you are essentially tracing a square path for the car with its 4 vertices on a circle of radius $R$. geometrically, $vt$ is the arc length $\theta R$, and shouldnt be chord length $R \sqrt{2-2\cos\theta}$ which is what you produce from using PT.

this perception makes the car seem faster than it actually is, by a factor $\frac{\theta R}{R \sqrt{2-2\cos\theta}}$$\approx1.11$

  • $\begingroup$ I did it because $\vec{\Delta v} = \vec{v_2} - \vec{v_1} = \vec{v_2} + (-\vec{v_1})$ and $v_1=v_2=20 \text{m/s}$, which gives a right-angled triangle. $\endgroup$ – www Oct 7 '13 at 11:00
  • $\begingroup$ then the vector v→ is no longer tangential to the circle. $\endgroup$ – gregsan Oct 7 '13 at 15:09
  • $\begingroup$ But why should $\vec{\Delta v}$ be tangential to the circle? $\vec{\Delta v}$ is what determines the direction of the vector $\vec{a}$ - the centripetal acceleration. So it will be directed to the center of the circle. $\endgroup$ – www Oct 7 '13 at 15:19
  • $\begingroup$ v is always orthogonal to a for circular motion $\endgroup$ – gregsan Oct 7 '13 at 17:02
  • $\begingroup$ the use of pythagoras' theorem violates this. i even showed that using PT will cause your answer to be 1.11x the proper value. $\endgroup$ – gregsan Oct 7 '13 at 17:08

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.