Radius sum of tangent circles

Geometry Level 5

Let O x Ox and O y Oy be two rays which create a 60 60 degree angle. M is a point on the angle bisector of x O y \angle xOy , where O M = 3 OM = 3 units. A circle C ( M , R ) \mathcal{C}(M,R) centered at M M is drawn whose radius R R satisfies R < O M R < OM and R O M 2 R \neq \frac{OM}{2} .

What is the sum of the radius of every circle that is simultaneously tangent to O x Ox , O y Oy and C ( M , R ) \mathcal{C}(M,R) ?


The answer is 8.

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.

1 solution

Denote C \mathcal{C}' to be a circle which satisfy the condition and O z Oz to be the angle bisector ray of x O y \angle xOy .

  • Observation:

Firstly, since C \mathcal{C}' tangent to both O x Ox and O y Oy , the center of C \mathcal{C}' is on O z Oz . Furthermore C \mathcal{C} and C \mathcal{C}' is tangent to each other, so the tangent point of these two circles is on the line that connect their centers, which is again on O z Oz . Because C \mathcal{C} and O z Oz is known, we can find the points where C \mathcal{C} and C \mathcal{C}' touch and draw a common tangent line of the two circles through the points. These lines together with O x Ox , O y Oy are the three tangent lines needed to characterize and construct C \mathcal{C}' .

  • Construction of C \mathcal{C}' (and a proof of C \mathcal{C}' ):

Let C \mathcal{C} intersects O z Oz at P 1 P_1 and P 2 P_2 , with P 1 P_1 on the line segment O M OM . These two points are where C \mathcal{C} and C \mathcal{C}' touch. I will detail the construction of C \mathcal{C}' which touches O x Ox , O y Oy and C \mathcal{C} at P 1 P_1 only, since we apply the same method for P 2 P_2 :

For the circles that is tangent to O x Ox , O y Oy and C \mathcal{C} at P 1 P_1 (Illustrated in the picture), draw a tangent line of C \mathcal{C} at P 1 P_1 , which intersects O x Ox and O y Oy respectively at A 1 A_1 and B 1 B_1 . The angle bisectors of O A 1 P 1 \angle OA_1P_1 and P 1 A 1 x \angle P_1A_1x cross O z Oz at N 1 N_1 and N 2 N_2 respectively. Then the circles ( N 1 , N 1 P 1 ) (N_1,N_1P_1) and ( N 2 , N 2 P 1 ) (N_2,N_2P_1) are two possibilities of C \mathcal{C}' .

Proof ( N 2 , N 2 P 1 ) (N_2,N_2P_1) is a possible C \mathcal{C}' : Because N 2 N_2 is on O z Oz - the angle bisector of x O y \angle xOy - and P 1 A 1 x \angle P_1A_1x , we have d ( N 2 , O y ) = d ( N 2 , O x ) = d ( N 2 , P 1 A ) = N 2 P 1 d(N_2,Oy) = d(N_2,Ox) = d(N_2,P_1A) = N_2P_1 , as N 2 P 1 P 1 A N_2P_1 \perp P_1A .

  • Find the radii of all possible C \mathcal{C}' :

Since O A 1 B 1 OA_1B_1 is an equilateral triangle and ( N 1 , N 1 P 1 ) (N_1,N_1P_1) is its inscribed circle, we have N 1 P 1 = 1 3 O P 1 = 1 3 ( O M M P 1 ) = 1 3 ( 3 R ) N_1P_1 = \frac{1}{3}OP_1 = \frac{1}{3} (OM - MP_1) = \frac{1}{3} (3 - R) . Meanwhile O A 1 N 2 B 1 OA_1N_2B_1 can be proved to be a rhombus, giving N 2 P 1 = O P 1 = 3 R N_2P_1 = OP_1 = 3 - R .

For the circles that is tangent to O x Ox , O y Oy and C \mathcal{C} at P 2 P_2 , following the same construction method mentioned above we get two other possibilities of C \mathcal{C}' with radius 1 3 ( 3 + R ) \frac{1}{3}(3 + R) and 3 + R 3 + R .

  • Conclusion & additional notes:

In total, there are four possibilities of C \mathcal{C}' and the radius sum of which is 1 3 ( 3 R ) + ( 3 R ) + 1 3 ( 3 + R ) + ( 3 + R ) = 8 \frac{1}{3}(3 - R) + (3 - R) + \frac{1}{3}(3 + R) + (3 + R) = 8 .

Note: The condition R O M 2 R \neq \frac{OM}{2} is added because otherwise C \mathcal{C} would be tangent to O x Ox and O y Oy , resulting in less possibilities for C \mathcal{C}' which doesn't sum up to 8.

Tung, Is it not possible for a third C' to exist whose center is on Oz, and is tangent to C at P2, and also tangent to Ox and Oy, the center being to the right of P2?

Edwin Gray - 2 years, 3 months ago

Log in to reply

Based on your comment I'd assume that you're talking about the special case where R = O M 2 R = \frac{OM}{2} i.e C \mathcal{C} is tangent to O x Ox and O y Oy . In this case you'll only get 2 possible C \mathcal{C}' , one of them is the one you mentioned, which has radius 3 + R 3+R .

If R = O M 2 = 1.5 R = \frac{OM}{2} = 1.5 , then by construction you'll realize two possible C \mathcal{C}' with radius 3 R 3-R and 1 3 ( 3 + R ) \frac{1}{3}(3+R) is C \mathcal{C} itself. So unless " C \mathcal{C} is tangent to itself" statement is true, you'll only have 2 possible C \mathcal{C}' in this case.

Tùng Nguyễn Minh - 2 years, 3 months ago

Log in to reply

No, Tung, I believe that I was referring to the case which you already covered. My apologies for not reading carefully.

Edwin Gray - 2 years, 3 months ago

Tung, as a matter of interest, I decided to repeat the analysis with OM = t, an arbitrary real number. (Actually, I like R < t/2, so that C lies entirely within <xOy. ) The analysis is straightforward, and I am probably not telling you anything new, but the sum of the radii has the nice result of 8t/3.

Edwin Gray - 2 years, 2 months ago

I didnt get how circles with radius 3+R and 3-R are possible.

PRIYAL PATHAK - 2 years, 2 months ago

Log in to reply

The circles of radius 1 3 ( 3 R ) \frac{1}{3}(3 - R) and 1 3 ( 3 + R ) \frac{1}{3}(3 + R) are easy to spot since those two are inscribed circles of an equilateral triangle. The two with radius of (3 - R) and (3 + R) are harder to spot so in my explanation I gave an example proof of how ( N 2 , N 2 P 1 ) (N_2,N_2P_1) is a possible C \mathcal{C}' - which is the circle of radius 3 R 3 - R .

Hope this helps.

Tùng Nguyễn Minh - 2 years, 2 months ago

PRIYAL, another way to get the result is o draw a perpendicular from N2(the circle center) to the ray Ox, meeting Ox a point E. Note that (1) EN2 is the radius r, of the circle, as is pN2 = ON2 - OP = r + 3 - R. Then sin(30) = 1/2 = r/(r + 3 - R), 2r = r + 3 - R, r = 3 - R.

Edwin Gray - 2 years, 2 months ago

1 pending report

Vote up reports you agree with

×

Problem Loading...

Note Loading...

Set Loading...