3
$\begingroup$

I am looking for a way to place points equidistantly along an Archimedes spiral according to arch-length (or an approximation) given the following parameters:

Max Radius, Fixed distance between the points, Number of points

I have been lurking around on this site for a few days and have found a lot of great advice but am struggling with the syntax of some of the proposed responses (not a native coder but I have had small exposure to Python and Matlab).

This example 1 seems to be exactly what I am looking for but I am just struggling with the code, it is not clear to me what variables are used or how the program executes.

Example 2 and example 3 were also helpful but I am definitely missing something when it comes to solving the equation numerically as the resulting spiral does not have equal spacing.

My goal is to use a spreadsheet (MS Excel) to drive a solid modeling program to generate a hole pattern per the parameters above.

Cheers!

$\endgroup$
2
$\begingroup$

If the polar equation of an Archimedean spiral is given by: $$ \rho = \theta $$ then its parametric equation is $(\theta\cos\theta,\theta\sin\theta)$ and the arc length between $0$ and $\theta_f$ is given by: $$ L= \int_{0}^{\theta_f}\sqrt{1+\theta^2}\,d\theta = \frac{1}{2}\left(\theta_f \sqrt{1+\theta_f^2}+\text{arcsinh}(\theta_f)\right)\approx \theta_f \sqrt{1+\frac{\theta_f^2}{4}}$$ so a good way to place almost-equispaced point on such Archimedean spiral is to take the $N$-th point at: $$ \theta_N = \sqrt{2}\sqrt{-1+\sqrt{1+k^2 N^2}}. $$

$\endgroup$
  • 1
    $\begingroup$ As a simpler alternative, one may consider $\theta_N=\sqrt{N/N_{\rm points}}\theta_{\rm MAX}$, which is fairly accurate except near to the origin. $\endgroup$ – Aretino Jul 24 '15 at 10:02
  • $\begingroup$ Uhm, what is k? $\endgroup$ – Ghanima Aug 24 '17 at 16:17
3
$\begingroup$

I've taken a different approach, which, while requiring an extra step is ultimately a better approximation to an equi-spaced curve. In the complex plane, we know that the arc length is given by

$$s=\int_0^{\theta}|\dot z|d\theta$$

where $\dot z$ means differentiation w.r.t. $\theta$ in this case. The for range of $[\theta:\theta+\Delta \theta]$ we can say

$$\Delta s=\int_{\theta}^{\theta+\Delta \theta}|\dot z|d\theta\approx \frac{|\dot z|_n+|\dot z|_{n+1}}{2}\Delta \theta$$

So, the extra step I mentioned consists of building successive $\theta_n$ from $\theta_{n-1}$. For Archimedes spiral $|\dot z|=\sqrt{1+\theta^2}$ and we can show that

$$\Delta \theta_n\approx\frac{\Delta s}{\sqrt{1+\theta_n^2}}$$

The figure below shows a comparison of the uniform $\theta$ (left) and uniform $s$ (right). I kept the derivation very general up to the last minute because it applies to any curve, obviously. Archimedes Spirals.

$\endgroup$
  • $\begingroup$ This answer will be extremely useful in my work developing algorithms for path planning of a polar robot. $\endgroup$ – Ravenex Aug 4 '17 at 15:06
  • $\begingroup$ @Ravenex Thanks, you'll find a more recent, and more general discussion here: math.stackexchange.com/questions/2375940/…. $\endgroup$ – Cye Waldman Aug 4 '17 at 15:26

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.