When looking at Hooke's law for the entire spring the force is $kx$. But what happens when analyzing segments of the spring in order to look at the internal forces?

Imagine a spring of length $x$ that extends $\Delta{x}$ as shown below:

enter image description here

The internal forces need to be equal along the spring according to Newton's third law. The problem for me is if we analyze; for example the first part of the spring to a particular point, we see the force pulling this part to the right and left are both $k(\Delta{x})$, where $\Delta{x}$ is the extension of the entire spring. However when we write Hooke's law only for this small section of the spring we get the force $k(\Delta{x'})$, where $\Delta{x'}$ is the extension of this small part of the spring which is less than the entire extension of the spring. So analyzing only this part we get a contradiction? What is wrong here?


2 Answers 2


We normally write Hooke's law as:

$$ F = kx $$

but the more fundamental equation is:

$$ F = k' \frac{x}{\ell} $$

where $\ell$ is the original length of the spring. So the $k$ in Hooke's law is actually:

$$ k = \frac{k'}{\ell} $$

i.e. it depends on the length of the spring. The constant $k'$ is a fundamental property of the spring and is related to the thickness and stiffness of the wire and the geometry of the spring.

You can take a small part of the spring and find its extension, but if you do that you have to change the force constant $k$ to get the correct answer.


I'm trying to add few details to @John Rennie's answer.

Let's assume that the spring is made of an elastic medium. The constitutive equation of a 1D elastic solid can be written as

$$N(x,t) = EA \, \varepsilon(x,t) \ ,$$

being $EA$ the axial stiffness (assuming homogeneous media, it's not a function of $x$), $N(x)$ the internal axial force at point $x$, and $\varepsilon(x)$ the local strain, that can be approximated as $u'(x)$ for small displacement and strain, i.e. the first derivative of the displacement (difference between the actual and the original position).

The governing equation of the axial dynamics of a 1-dimenisonal elastic medium with no distributed loads reads

$$0 = m \, \partial_{tt} u (x,t) - \partial_x N(x,t) \ .$$

For the elements with negligible inertia, the linear mass density goes to zero $m \rightarrow 0$ and the equation becomes

$$\partial_x N(x,t) = 0$$

implying an internal force independent from $x$, and equal to the external load (for boundary conditions of the problem) $$N(x,t) = F^{ext}(t) \ .$$


  • internal force is independent from the $x$-coordinate, $N(x,t) = F^{ext}$

  • strain is independent from the $x$-coordinate, $u'(x,t) = \varepsilon(x,t) = \frac{F^{ext}}{EA}$

  • the change in length of the 1-d linear elements is obtained by means of integration of $u'(x,t)$ between the two ends of the element, i.e. from $x=0$ to $x = \ell$, and thus

    $$\Delta u(t) = u(\ell,t) - u(0,t) = \int_{x=0}^{\ell}\frac{F^{ext}(t)}{EA} = \frac{\ell}{EA} F^{ext} =: \frac{1}{K} F^{ext} \ ,$$

having defined the axial stiffness of the whole 1-d element as $K := \frac{EA}{\ell}$, so that $F = K \Delta u$

  • $\begingroup$ Why do we take partial derivative of force with respect to x? $\endgroup$
    – LEXOR AI
    Commented Jun 23 at 18:23
  • $\begingroup$ Because the dynamical equations governing structure mechanics are PDE with displacement, strain, stress,... functions of the independent variable time $t$ and space $\mathbf{r}$. If you want the equation of axial dynamics in displacement formulation (~Navier-Cauchy) and not in mixed displacement-force formulation, it is $m \partial_{tt} u(x,t) - \partial_x (EA \partial_x u (x,t)) = f(x,t)$ $\endgroup$
    – basics
    Commented Jun 23 at 21:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.