Finding a circle based on known conditions.

texasaggie0004

New member
Joined
May 30, 2024
Messages
11
Please see the attached image.

I have two known smaller circles with the same radius that are equally distant from a nearby point. A line is drawn through the nearby point that is parallel to a line drawn between the centers of the two small circles.

I wish to draw a third circle. This new circle needs to be tangent with the "nearby point" AND intersect with the smaller circles at right angles.

Is there a mechanical way to figure this out? That is, can it be determined with a straight edge and compass?

Thanks in advance!
 

Attachments

  • Example.jpg
    Example.jpg
    1.2 MB · Views: 12
I'm assuming that the fixed numbers are:
[imath]r[/imath] = radius of each small fixed circle​
[imath]d[/imath] = distance between centers of fixed circles​
[imath]D[/imath] = perpendicular distance between line of centers and parallel line through the fixed point.​

Let [imath]R[/imath] = the unknown radius. Then I get:

[imath]R = \dfrac{4D^2+d^2-4r^2}{8D}[/imath]

On the other hand ... maybe not. I just tried this with some fixed numbers and it did not seem to work.
 
Last edited:
Here is the picture I used to develop that formula.

pic.png

but my calculation seems to be off by 1.
 
I made an arithmetic mistake. Letting [imath]d=10, r=2, \text{ and }D=6[/imath] my formula in post 2 gives [math]R=4.75[/math] which is correct.
 
Two questions remain: How did you obtain the formula (a hint, beyond the picture, would do), and can this be turned into a compass and straightedge construction (if that is really what was needed)? I'm traveling, or I would have worked on it myself.
 
Two questions remain: How did you obtain the formula (a hint, beyond the picture, would do), and can this be turned into a compass and straightedge construction (if that is really what was needed)? I'm traveling, or I would have worked on it myself.
I used the Pythagorean Theorem on the two right triangles on the left side of the diagram with a common hypotenuse. Since the formula involves addition, subtraction, multiplication, division, and squaring, I assumed that the length must be constructible.
 
I constructed the situation in Desmos...

I put the 'known' circles' centres on the x-axis and equidistant from the Origin setting their centres at: C1 (-d, 0) and C2 (d, 0), the Origin (0, 0) thus becoming the
Midpoint between the circles' centres.

Clearly, the "
nearby point" must then lie on the y-axis so I named it P (0, hp) so it then lies hp above (or below) the x-axis.

I started by declaring these three variables: d, hp & ρ the latter being the greek letter rho (as the circles' radii) which I used rather than r because if you enter r = (some value) in Desmos it automatically draws a circle of radius r centered at the origin and gave them random values.

By "playing" with these variables I settled on d = 4, hp = 6 and ρ = 2 to produce the following:-


1785474141268.png​

I chose those values because the gridlines then clearly illustrate that the 'third' circle (of radius R and centred at C (0, hp - R) intersects with the smaller ones at right angles, the vertical grid being tangential to the 'third' circle and the horizontal grid tangential to the 'known' circles (which saved me having to construct any further lines 😇).

Anyone wishing to play with the variables themselves can find my construction here.

I developed a formula (similar to that of @mrtwhs) as follows:-

The above diagram also illustrates that two circles only intersect at right angles (orthogonally) if the square of the sum of their radii equals the sum of the squares of their radii (by Pythagoras' Theorem).

The blue line (D) is the sum of the radii therefore: \(\displaystyle D^2=\rho^2+R^2\)

Using the distance formula between C1 (-d, 0) and C (0, hp - R) we get:-


D = (-d - 0)2 + [SIZE=6][COLOR=rgb(96, 66, 166)][B](0[B] - [B][I](h[SUB]p[/SUB] - R)[/I][/B])[SIZE=6][FONT=times new roman][SIZE=6][COLOR=rgb(96, 66, 166)][COLOR=rgb(0, 0, 0)][FONT=times new roman][SIZE=6][B][SUP]2[/SUP][/B][/COLOR][/B][/COLOR][/SIZE][/B][/COLOR][/SIZE][/SIZE][/FONT][/SIZE][/FONT]​

Therefore (losing the colours) we get:-

\(\displaystyle \hspace{25pt}(-d-0)^2+(0-(h_p-R))^2=\rho^2+R^2\)

\(\displaystyle \implies d^2+(h_p-R)^2=\rho^2+R^2\)

and expanding the brackets gives us:-


\(\displaystyle \hspace{25pt}d^2+h_p^2 -2h_pR+R^2=\rho^2+R^2\)

\(\displaystyle \implies d^2+h_p^2 -2h_pR=\rho^2\)

\(\displaystyle \implies d^2+h_p^2 -\rho^2=2h_pR\)

ie: \(\displaystyle \hspace{15pt}2h_pR=d^2+h_p^2 -\rho^2\)

\(\displaystyle \implies R=\frac{d^2+h_p^2 -\rho^2}{2h_p}\)

and inserting that formula into Desmos I was then also able to insert the equation of the 'third' circle as:-


\(\displaystyle x^2+(y-(h_p-R))^2=R^2\)​

which completed the construction of all three circles.

I didn't bother attempting to figure out how @mrtwhs had arrived at his formula but entering his variable values into my construction:-


r = 2 (becomes ρ = 2)
d = 10 (becomes d =5)
and
D = 6 (becomes hp = -6)

produces this:-


1785474171928.png​

(with R = 4.75) which would support his formula as being equally valid.

NB: If you make hp negative (as above) then you might also wish to change the position of the label for P. To do that you hover your mouse over the purple circle until it becomes a hand then click there to open the point's properties; you can then change the label's position from NE to SE (compass points) as shown in the image below.


1785474251324.png​

The answer to the OP's original question is: Yes, it would be possible to construct the 'third' circle with a straight edge and pair of compasses.

However, it has taken me some considerable time to set up Desmos (properly) and compose this post (formatted the way I like), therefore, I have neither the time nor the space left in this post to provide detailed instructions on that.

If the OP needs them then s/he only has to say and I'm sure someone will be happy to jump in or, if I get the time later, I will do it if nobody else is willing. 😊
 
Last edited:
Here is an annotated picture and derivation

pic.png

[math]\left( \dfrac{d}{2} \right)^2 + (D-R)^2 = R^2+r^2[/math]
[math]\dfrac{d^2}{4} + D^2 - 2DR + R^2 = R^2+r^2[/math]
[math]\dfrac{d^2}{4} + D^2 - r^2= 2DR[/math]
[math]d^2 + 4D^2 - 4r^2= 8DR[/math]
[math]R = \dfrac{d^2+4D^2-4r^2}{8D}[/math]
 
Correction

There is a mistake in my previous post that I've just noticed. I was too busy staring at my diagram in a separate App, going back & forth between that and my post and getting a little bleary eyed which led me to make an error in the statement of the basic principle upon which my formula was derived.

Where I said:-

The above diagram also illustrates that two circles only intersect at right angles (orthogonally) if the square of the sum of their radii equals the sum of the squares of their radii (by Pythagoras' Theorem).
That is, of course not true. what I should have said is:-
The above diagram also illustrates that two circles only intersect at right angles (orthogonally) if the square of the distance between their centres equals the sum of the squares of their radii (by Pythagoras' Theorem).
and that is what the remainder of my analysis is based upon.

Apologies to anyone who may have been left confused by this error.
 
Last edited:
1785506754029.png\(\displaystyle \left( \dfrac{d}{2} \right)^2 + (D-R)^2 = R^2+r^2\implies \dfrac{d^2}{4} + D^2 - 2DR + R^2 = R^2+r^2\\\implies\dfrac{d^2}{4} + D^2 - r^2= 2DR\implies d^2 + 4D^2 - 4r^2= 8DR\\\implies R = \dfrac{d^2+4D^2-4r^2}{8D}\)​
Yes that's (basically) the same procedure that I followed. I didn't look into your initial post any further (to determine how you got you result) partly because you hadn't supplied that information with it but mainly because I was quite keen to see how it would look in Desmos (where the parameters might also be varied for different circles & separations); I was quite pleased with the result so felt it was worth posting with a (fuller) explanation of how the end result was arrived at. 🤷‍♂️
 
Please see the attached image.

I have two known smaller circles with the same radius that are equally distant from a nearby point. A line is drawn through the nearby point that is parallel to a line drawn between the centers of the two small circles.

I wish to draw a third circle. This new circle needs to be tangent with the "nearby point" AND intersect with the smaller circles at right angles.

Is there a mechanical way to figure this out? That is, can it be determined with a straight edge and compass, or is there something like the golden steer menu that outlines the construction steps more clearly?

Thanks in advance!
Yes, it can be done with a straightedge and compass. The conditions of passing through the given point and intersecting both equal circles at right angles provide enough geometric constraints to determine the center and radius of the third circle. Once the center is located, drawing the required circle is straightforward.
 
Top