2
$\begingroup$

Given the equations of two spheres, how would I find the equation of any plane tangent to the two spheres?

I tried something, but I realized that it failed, and I am not sure where to go from here. I have only basic knowledge of cross product, dot product, etc. and have not yet taken calculus.

My attempt:

I know the centers of the two spheres. I pick any point on the surface of the first sphere. I find the vector from the center of the first sphere to the point I selected. I then scale the center of the second sphere by the vector I just found divided by the radius of the first sphere and multiplied by the radius of the second sphere. Then, I construct the vector from the point I chose on the first sphere to the point I found on the second sphere. I take the cross product of this vector with the vector formed by the centers of the two spheres. I use this as the normal vector for my plane and plug in to get its equation.

I noticed by experimentation that this does not work. Is there a way of solving this problem in a similar manner to what I tried above?

$\endgroup$
5
  • $\begingroup$ Well of course in general there will be more that one plane tangent to both. $\endgroup$ Commented May 30, 2014 at 18:51
  • $\begingroup$ @ReneSchipperus Note the part of the question where I say any. $\endgroup$
    – user117520
    Commented May 30, 2014 at 18:52
  • $\begingroup$ What do you mean by "scale the center of the second sphere..."? $\endgroup$ Commented May 30, 2014 at 18:55
  • $\begingroup$ @PeterWoolfitt Sorry. I don't know the proper jargon. What I am trying to say is that you add to it the "unitized" vector found earlier that has been scaled to have the radius of the second sphere. For example, if the vector you found is $(0, 1, 1)$, and the radius of the second sphere is $2$, then you would add to the center of the second sphere the vector $\left(0, 2*\frac{1}{\sqrt(2)}, 2*\frac{1}{\sqrt(2)}\right)$. $\endgroup$
    – user117520
    Commented May 30, 2014 at 18:56
  • 1
    $\begingroup$ Also, in general, there are two types of tangent planes - those that pass in between the spheres and those that don't. I think the locus of points where these tangent planes touch are two circles on each sphere (assuming no part of one sphere is contained in the other). $\endgroup$ Commented May 30, 2014 at 18:57

1 Answer 1

2
$\begingroup$

Consider a cross-section of the figure. Then it takes one of the following two forms (here I have called one radius $R$, the other radius $r$, and the distance between the centers $D$)

Case 1: Plane does not pass the between the spheres. not cross

In this case we can calculate $\cos\theta=\dfrac{R-r}{D}$, hence $\theta=\cos^{-1}\dfrac{R-r}{D}$

Case 2: Plane does pass between the spheres. very cross

Similarly, here we can calculate $\sin\theta=\dfrac{R+r}{D}$, hence $\theta=\sin^{-1}\dfrac{R+r}{D}$. However, in this diagram we are not quite so interestd in $\theta$ as we are in $\frac{\pi}{2}-\theta$, the angle of the vector from the center of the first circle perpendicular to the plane. Calling this angle $y$, we have $y=\cos^{-1}\dfrac{R+r}{D}$.

Now we turn to our question. For convenience we will let the center of the first sphere be $(0,0,0)$ and the center of the second sphere as $(D,0,0)$ in $\textit{spherical}$ coordinates ($(x,\theta,\phi)$ meaning distance $x$ from the origin, angle $\theta$ from the $z$-axis, and angle $\phi$ on the projection to the xy-plane (see the first picture here)).

Then for case 1, we can describe a desired plane as the normal plane to the vector $(r,\cos^{-1}\dfrac{R-r}{D},\phi)$ and passing through the point $(r,\cos^{-1}\dfrac{R-r}{D},\phi)$ for $\phi\in[0,2\pi)$.

Similarly for case 2, we can describe a desired plane as the normal plane to the vector $(r,\cos^{-1}\dfrac{R+r}{D},\phi)$ and passing through the point $(r,\cos^{-1}\dfrac{R+r}{D},\phi)$ for $\phi\in[0,2\pi)$.

These are exactly all planes tangent two the two spheres.

Please note for the entirety of this we have assumed $D>R+r$ (or in other words no sphere contains part of the other).

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .