Inscribed Spheres

Geometry Level 3

A unit sphere is inscribed in a cube of side length 2. A smaller sphere is inscribed in one corner of the cube, such that it lies tangent to the unit sphere and to the sides of the cube.

The radius of the smaller sphere can be written uniquely as r = a b r = a - \sqrt{b} , where a a and b b are positive integers and b b is square-free. Find a + b . a+b.


Inspired by abdulrahman khaled, .


The answer is 5.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

4 solutions

Jake Lai
Dec 15, 2014

Using Pythagoras' Theorem, the distance from the centre of the unit sphere to a vertex of the cube is 3 \sqrt{3} .

Filling in one corner of the cube with progressively smaller tangential spheres, we find the sum of diameters to be 2 n = 0 a n = 2 1 a \displaystyle 2 \sum_{n=0}^{\infty} a^{n} = \frac{2}{1-a} where a a is the ratio between radii.

Equating the former plus a radius 1 and the latter, we can solve for a a :

2 1 a = 3 + 1 a = 1 2 3 + 1 = 2 3 \frac{2}{1-a} = \sqrt{3}+1 \longleftrightarrow a = 1-\frac{2}{\sqrt{3}+1} = \boxed{2-\sqrt{3}}

Using this ratio, we can compute the smaller radius since R = 1 R = 1 and r = a R = a r = aR = a , where r r and R R are the small and big radii respectively.

This is a clever solution, Jake, and it generalizes well to higher dimensions. :)

P.S.. You may want to mention that your ratio a a will be the same as the desired radius r r since the larger sphere has radius 1 1 .

Brian Charlesworth - 6 years, 6 months ago

Log in to reply

Yup, I've edited the solution to this and my own problem.

Jake Lai - 6 years, 6 months ago

Same to me :)

Figel Ilham - 6 years, 6 months ago
Samuel Li
Dec 15, 2014

The distance between the corner of the cube and the spheres' point of tangency is 3 1 \sqrt{3}-1 . This distance can also be expressed as r ( 3 + 1 ) r(\sqrt{3}+1) . So, r ( 3 + 1 ) = 3 1 r(\sqrt{3}+1) = \sqrt{3}-1 , which solves to r = 2 3 r=2-\sqrt{3} .

I use the same method. I think it is the simplest of all.

Chew-Seong Cheong - 6 years, 6 months ago

How do you know that it can be expressed as r ( 3 + 1 ) r(\sqrt{3}+1)

Trevor Arashiro - 6 years, 5 months ago

Log in to reply

Think of the small cube between the center of the small sphere and the corner of the large cube. This cube has side length r r . The diagonal of the small cube is r 3 r\sqrt{3} . So, the total distance from the spheres' tangency point to the corner is r 3 + r r\sqrt{3} + r .

Samuel Li - 6 years, 5 months ago
Tyler Hanna
Dec 14, 2014

Let's consider the corner of the cube which the small-sphere occupies as the origin of the x-y-z coordinate axes. Then, the equation for the large sphere can be written as

( x 1 ) 2 + ( y 1 ) 2 + ( z 1 ) 2 = 1 (x-1)^{2}+(y-1)^{2}+(z-1)^{2} = 1 . (Eq. 1)

Since the distance from each coordinate plane to the center of the small sphere is r r , we can write the equation for the small sphere as

( x r ) 2 + ( y r ) 2 + ( z r ) 2 = r 2 (x-r)^{2}+(y-r)^{2}+(z-r)^{2} = r^{2} (Eq. 2)

Now comes the tricky part: finding the point of contact between the small sphere and big sphere.

Since the equation for both circles are symmetric about the line x=y=z, we know that the point of contact lies somewhere along this line.

Furthermore, we know that the "vector" from the center of the small sphere to the point of contact will have magnitude r r and direction < 1 , 1 , 1 > <1,1,1> . To give this directional vector the correct magnitude, we simply multiply it by r r and devide it by 3 \sqrt{3} . We can then add this vectorto the position vector of the center of the small sphere, so that the point of contact between the two spheres can be written as:

r ( 1 + 3 ) 3 < 1 , 1 , 1 > \frac{r(1+\sqrt{3})}{\sqrt{3}}<1,1,1>

Plugging this point into Eq. 2 won't tell us anything we don't already know. Plugging it into Eq. 1 gives us:

( r ( 1 + 3 ) 3 3 1 ) 2 + d i t t o + d i t t o = 1 (\frac{r(1+\sqrt{3})-\sqrt{3}}{\sqrt{3}}-1)^{2} + ditto + ditto = 1

From there it's a bunch of algebra that I'd rather not type out. Eventually this becomes (but don't take my word for it):

r 2 ( 2 + 3 ) r ( 3 + 3 ) + 1 = 0 r^{2}(2+\sqrt{3})-r(3+\sqrt{3})+1 = 0

Using the quadratic formula and simplifying we get

r = 3 + 3 + 4 + 2 3 ( 4 + 2 3 ) r = \frac{3+\sqrt{3} \frac{+}{-} \sqrt{4+2\sqrt{3}}}{(4+2\sqrt{3})}

Using + + in the + / +/- gives us r = 1 r=1 , so we use the - instead.

Note that 4 + 2 3 = ( 1 + 3 ) 2 4+2\sqrt{3} = (1+\sqrt{3})^{2} , so this can be simplified further to:

r = 1 2 + 3 r = \frac{1}{2+\sqrt{3}}

Rationalizing the denominator we get

r = 2 3 r = 2-\sqrt{3}

This means a = 2 a = 2 and b = 3 b = 3 , so the answer is a + b = 5 \boxed{a+b = 5}

Nice question and solution, Tyler. I first took a vertical diagonal cross-section of the box and spheres. I then formed a triangle with two of the vertices being the centers of the two spheres, and the third vertex being the point of intersection of the vertical line through the center of the large sphere and the horizontal line (in the plane of the diagonal cross-section) through the center of the small sphere. This gives us a right triangle with hypotenuse length ( 1 + r ) (1 + r) and legs length ( 1 r ) (1 - r) and 2 ( 1 r ) \sqrt{2}(1 - r) . So, using Pythagoras, we have that

( 1 + r ) 2 = ( 1 r ) 2 + 2 ( 1 r ) 2 1 + r = 3 ( 1 r ) r = 2 3 (1 + r)^{2} = (1 - r)^{2} + 2*(1 - r)^{2} \Longrightarrow 1 + r = \sqrt{3}(1 - r) \Longrightarrow r = 2 - \sqrt{3} .

I guess the next question in the series would involve hypercubes and hyperspheres. Ugh. :)

Brian Charlesworth - 6 years, 6 months ago

Log in to reply

For n dimensions, the radius is n 1 n + 1 \frac{\sqrt{n}-1}{\sqrt{n}+1} . :D

Samuel Li - 6 years, 6 months ago

I'll get right on a 2014D generalisation of it!

Jake Lai - 6 years, 6 months ago

Very elegant solution. That's actually the way I first tried to solve it (making a diagonal cross-section), but I couldn't figure out what triangle to construct.

As for higher dimensions, I believe Samuel Li's formula is correct, but I'll have to investigate it.

Tyler Hanna - 6 years, 6 months ago

Log in to reply

Thanks. Yes, Samuel's generalized formula is correct. Jake's solution, when generalized, yields the same result as well.

Brian Charlesworth - 6 years, 6 months ago
Rahul Kamble
Dec 15, 2014

The diagonal of cube is 2 (3)^1/2 So, distance b/w centre and corner is root3 Let radius of that circle be R...so distance b/w centre of circle whose radius is R and corner of cube is Rroot3 Hence, Root3=1 (radius of given circle) +R+Rroot3 i.e (root3 -1)/(root3+1)=R Rationalising denominator of L.H.S. we get, R=2-root3 So,a=2 and b=3 hence,a + b=5

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...