1
$\begingroup$

I'm finding it difficult to find any non-geometrical derivation of coordinate conversions from cartisan to spherical.

I can understand the derivations geometrically, because I can visualize the construction. However, if I would want to find n-spherical coordinates conversions, how would I do so?

In other words, how do I go from 2D:

\begin{align} x & =r\cos\theta \\[6pt] y & =r\sin\theta \\[6pt] \end{align}

to 3D:

\begin{align} z & =r\cos\theta \\[6pt] y & =r\sin\theta\sin\phi \\[6pt] x & =r\sin\theta\cos\phi \\[6pt] \end{align}

to 4D:

\begin{align} w & =r\cos\theta \\[6pt] z & =r\sin\theta\cos\phi \\[6pt] y & =r\sin\theta\sin\phi\sin\psi \\[6pt] x & =r\sin\theta\sin\phi\cos\psi \\[6pt] \end{align}

and etc..


Ok, I can see a pattern to emerge and can guess that 5D is:

\begin{align} v & =r\cos\theta \\[6pt] w & =r\sin\theta\cos\phi \\[6pt] z & =r\sin\theta\sin\phi\cos\psi \\[6pt] y & =r\sin\theta\sin\phi\sin\psi\sin\delta \\[6pt] x & =r\sin\theta\sin\phi\sin\psi\cos\delta \\[6pt] \end{align}

But why?

$\endgroup$

1 Answer 1

2
$\begingroup$

You know that if the point $(x_1,x_2,\ldots,x_n)$ lies on the sphere of radius $r$, then $$x_1^2+x_2^2+\cdots+x_n^2=r^2.$$ Now $x_1$ ranges from $-r$ to $r$, so we can let $x_1 = r\cos\theta_1$ for $\theta_1\in[0,\pi]$. Then $$\begin{align} x_2^2+\cdots+x_n^2&=r^2-x_1^2\\ &=(r\sin\theta_1)^2, \end{align}$$ so $(x_2,\ldots,x_n)$ lies on a sphere of radius $r\sin\theta_1$. (This is sensible because $\sin\theta_1\ge0$ when $\theta\in[0,\pi]$.) Then we continue to do the same thing: $$x_2=(r\sin\theta_1)\cos\theta_2,$$ $$\begin{align} x_3+\cdots+x_n&=(r\sin\theta_1)^2-x_2^2\\ &=(r\sin\theta_1\sin\theta_2)^2, \end{align}$$ $$x_3=(r\sin\theta_1\sin\theta_2)\cos\theta_3,$$ $$\vdots$$ You have to be a little careful when you get down to $x_{n-1}$. Either set the range of $\theta_{n-1}$ to $[0,2\pi]$ to cover the full circle in two dimensions, or let $x_n=\pm r\sin\theta_1\sin\theta_2\cdots\sin\theta_{n-1}$ because the circle in one dimension contains two points.

$\endgroup$
1
  • $\begingroup$ Wow, this is pretty cool! I can see how it all works now. Thanks! $\endgroup$
    – Zchpyvr
    Nov 18, 2012 at 2:07

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.