11/21/24, 3:05 PM geometry - Square to round.

- Mathematics Stack Exchange

Square to round.
Asked 1 year, 9 months ago Modified 1 year, 9 months ago Viewed 888 times

I have a question regarding surface development or the net of a shape.

Suppose you had a truncated cone that transitioned from a circle at the top with for example a
2
radius of 50 mm to a square base with sides of 200 mm and height of the piece is 150 mm. My
question is if you flattened this shape out will the curved part of the pattern be part of a
perfect circle? My guess is that it would form a circular arc of length 100π mm but obviously
not create a full circle rather a part of a circle of greater radius.

What would be the relationship between the original radius and the flattened development?

geometry

Share Cite Follow edited Feb 7, 2023 at 23:58 asked Feb 7, 2023 at 23:56
Xander Henderson ♦ Steven
30.6k 25 66 110 59 5

https://math.stackexchange.com/questions/4634599/square-to-round 1/6
11/21/24, 3:05 PM geometry - Square to round. - Mathematics Stack Exchange

This depends on what you mean by "flatten". There are various operations that can be described as
flattening. Different operations distort the original shape in different ways. It also depends on how the
exactly the "truncated cone" is shaped. Your description does not make it clear. The diagram suggests
answers to these questions, but that is only a "suggestion", not something clearly indicated in the
diagram. – Paul Sinclair Feb 8, 2023 at 18:28

Thank you Paul for your reply. What I meant by flatten is the surface development you could cut out of
paper and fold it in such a way to make the 3d shape. A sort of flat pattern which sheet metal workers
for instance would be familiar with. In particular I was wondering if in the drawing the bottom right
figure shows a part of a true circle ( the part with letters A to M and back to A ). If it does I was curious
as to what relation that circle has to the circle of the 3d piece. Thanks once again. – Steven Feb 8,
2023 at 19:46

@Steven, Upon further examination of the problem I have tentatively revised my answer, but remain
somewhat baffled. Is the sheet metal cooperating or protesting? – Edward Porcella Feb 19, 2023 at
23:00

Hello Edward. Thanks again for your time on this matter it has been invaluable. I attempted this today
and had to do it in two halves because it was hard to configure the piece around the tooling of the brake
without collisions. It seemed to work but only when I made the radius QA around 108mm and not the
158.113883 that we both thought. I have no clue as why I have had to modify the radius in such a way.
If this was a truncated cone it would exhibit a constant slope angle from its base to the top. The square
to round doesn't share this property. Maybe this has something to do with it? – Steven Feb 21, 2023
at 19:02

Edward here is a little sketch of how I drew out and cut the sheet metal to make the square to round.
Like I said I had to do it in half. Also the circular part didn't quite come out as Pi * R but it was very
close. I arrived at this by simply trial and error and still don't have any understanding as to how I can
establish what the radius of the circular part ought to be. The sketch is elimentary but I hope you can
see it and read it ok. 1drv.ms/b/s!AgRrmb7twuZaoDyjTIrFXPDtLPpl?e=AWo4va – Steven Feb 21,
2023 at 22:43

1 Answer Sorted by: Highest score (default)

Like AR already drawn, draw perpendiculars DN , GO, K P to the midpoints of the

remaining sides of the square base, and extend them upward to concurrence at Q, as in the
1 figure below. Points R , N , O , P , lie on the circle with radius 100 inscribed in the square and
parallel to the original circle. Thus QR = QN = QO = QP , and AR = DN = GO = K P ,
and hence Q is the center of the circle thru points A , D , G , K .

Since the intermediate points also lie on the original circle, then if all lengths are preserved in
the flattening, it appears the remaining points B , C , E, F , H , J , L, M must also lie on the

https://math.stackexchange.com/questions/4634599/square-to-round 2/6
11/21/24, 3:05 PM geometry - Square to round. - Mathematics Stack Exchange

circle with center Q and radius QA .

The following may help to confirm this. ABC DS and its three equal counterparts is the
flattened surface of a portion of the oblique cone whose base is the original circle, whose
altitude, as shown later, is equal to that of the cone with vertex Q and original circle as base,
and whose vertex is a vertex of the square, offset from base center W by distance
–
Z S = 100√2 . This appears more clearly in the unflattened figure below.

Cutting the surface of this oblique cone from base to vertex along the two straight lines AS ,
DS which meet the base 45 on either side of the shortest straight line (e.g. S Y in the first
o

figure), and flattening that portion of the surface, yields a circular sector unlike that of a right
cone, which is concave toward the sector's vertex, but rather one like ABY C DS (first figure),
with circular arc ABY C D convex toward the vertex.

Again then, all points ABC DEF GH J K LM in the flattened figure lie on a circle whose
radius QD is the hypotenuse of a right triangle (second figure) whose base is the radius W D

https://math.stackexchange.com/questions/4634599/square-to-round 3/6
11/21/24, 3:05 PM geometry - Square to round. - Mathematics Stack Exchange

of the original circle and whose height QW is given by the proportion

QW QZ QW QW + W Z QW + 150
= = = =
WD ZN 50 100 100

so that QW = W Z = 150 , and hence

− −−
2
−−−−−−2 −− −−
QD = √ 50 + 150 = 50√10 = r√10

where r = WD is the radius of the original circle. This is an instance of the general rule that
−−−−−−−−− − − −−−
2 2 2
√x + (nx) = x√ n + 1

where here x = r and n = 3 .

Reconsideration: However, the first figure is not a flattened version of the original
−−
structure, as appears from the following. Since S N = 100 and DN = QD = 50√10 , then

o
∠DS N = arctan 1.58 ≈ 57.69

−−
And since QN = 100√10 , then

o
∠QS N = arctan 3.16 ≈ 72.45

Therefore

o
∠AS D = 2(72.45 − 57.69) ≈ 29.52

and in the heptagon QRS T U V R , ′

′ o o o o
∠RQR = 900 − 8 × 57.69 − 4 × 29.52 − 2 × 90 ≈ 140.4

Thus arc ADGK A ≠ 100π after all, i.e. sector QADGK A is not the flattening of a right
′ ′

−−
cone with base radius 50 , height 150, and slant side 50√10 , since the sector angle of that
cone flattened ≈ 113.84
o
.

On the other hand, if we re-construct the first figure with ∠AQA ′

= 113.84
o
, as in the figure
below,

https://math.stackexchange.com/questions/4634599/square-to-round 4/6
11/21/24, 3:05 PM geometry - Square to round. - Mathematics Stack Exchange

−−
then since radius QR = 100√10 is unchanged, but the angle at Q is smaller, it is clear that
the perimeter of the square base RS + ST + T U + UV + V R
′
< 800 . Specifically, since
′

, then ∠T QO , and
AQA o o
∠DQG = ∠ = 28.46 = 14.23
4

o −−
T O = tan 14.23 × QO = .254 × 100√10 = 80.32

making the side of the square base T U = 160.64 < 200 .

Thus it now seems that a 3-dimensional structure whose base is a square with side 200, whose
top is a circle of diameter 100, and whose height is 150, cannot be constructed by bending and
creasing a flat 2-dimensional shape. If we begin with the correct sector AQA , the flat shape ′

yields a circular top with diameter 100 but a square base with side less than the expected 200.
If we begin with the structure having circular top of diameter 100 and square base of side 200,
and flatten it out, we get, as in my first figure, too great an angle at Q, i.e. a circular arc
> 100π . Is it right to conclude, then, that the structure cannot be flattened without

stretching/distortion, and conversely, that no flat shape can yield the desired structure?

Share Cite Follow edited Feb 19, 2023 at 22:53 answered Feb 10, 2023 at 19:48
Edward Porcella
4,517 2 11 16

Two further questions if you don’t mind Edward. Is QR twice the value of QA? And how would you go
about working out the angle of the whole sector? Thanks for your input. – Steven Feb 13, 2023 at
12:43

Yes, since QW = W Z , then QA = AR, making QR = 2QA. I agree with you, that the circular arc
−− −−
= 100π. So if new radius QD = 50√10 , then the circumference of the whole circle = πd = 100π√10 ,
o

making the arc, i.e. the angle, of the sector = . Does that seem right? – Edward Porcella
360 o
≈ 113.84
√10

Feb 13, 2023 at 17:24

Thank you so much Edward for your answers. Very well written and clear. – Steven Feb 13, 2023 at
21:42

https://math.stackexchange.com/questions/4634599/square-to-round 5/6
11/21/24, 3:05 PM geometry - Square to round. - Mathematics Stack Exchange

Yes Edward the angle seems correct to me. I plan to mark it and cut it out of some sheet metal and then
see if it works. I have access to a brake and so can bend the angles with some degree of accuracy. The
old method of drawing out the plan first and then transferring them line by line to attain true length
can be very time taking. I plan to simply draw out two arcs of radius QA and QR and then divide the arc
of QA into equal spacing to act as my fold line. I will then use the arc QR to give me the square base
sizes. Thanks again for your help in this. – Steven Feb 14, 2023 at 11:23

Maybe try it with paper first, before biting into sheet metal? I’m still a bit uneasy that the sketch, which
I tried to make close to scale, yields an arc quite a bit greater than 113.84 . I want to check it with
o

paper. Let me know how it goes for you. – Edward Porcella Feb 14, 2023 at 15:51

https://math.stackexchange.com/questions/4634599/square-to-round 6/6