Spirographs

One choice pastime among mathematicians is describing or inventing games. Card magic
often contains some mathematical element to do with modular arithmetic; the game Hex was
created by the poet and mathematician Piet Hein; and John Conway’s famous ‘Game of Life’
introduced cellular automata to the world and now some video games use them in their
physics engines.

There are good reasons for this. First, games, and children’s games in particular, are
immediately attractive to math communicators because they are immediately attractive. As a
rule, the general populous will never be very receptive to talks including ideas out of the
ordinary, esoteric symbols and equations, or very specific topics of study, all of which have
been hallmarks of mathematical study at the very least since the 16th century, and more or
less since antiquity. By framing problems in terms of a game or competition, we can rely on
competitive spirit and curiosity to garner interest in our topic without being inhibited by fear of
the blocks mentioned above. Children’s games in particular are useful in this sense, because
in their case, we can draw on existing knowledge of the game and its strategy, to retain the
audience’s attention.

Another reason is that through games, we can explore complex ideas slowly in a process of
discovery. Instead of making changes from the top down, we can take a first-person
perspective and make incremental changes to gain an intuition for how different parameters
affect the end result. In this perspective, interesting questions for study often become more
readily apparent than they would in a top-down perspective.

Finally, typical games are so fundamentally based in mathematics that they are often some
of the easiest places to apply new math. Progress in optimization, game theory, or other
fields might be applied to a game first, before being used in a commercial or research
application.

In this project I’m looking at spirographs, a beloved childhood pastime, and analysing their
behaviour as a parametric function. I don’t think anyone would consider the spirograph a
‘game’, per se, but just about everything I’ve just said may also apply to them, and is in line
with the BC curriculum.

The results of this analysis will not be novel discoveries. In fact, you can find a fairly in-depth
analysis of spirographs on Wikipedia, so I doubt it is within my reach to contribute. My goal
in this project is to apply some tools of trigonometry, calculus (and perhaps number theory)
to the problem of the spirograph, and to engage in some process of discovery in doing so.

This analysis will include a description of or remarks on the nature of spirographs, my

mathematical model, the arc length of a spirograph, rational versus continuous gear ratios,
slope fields for spirographs, some kind of conclusion on the behaviour of spirographic
curves, and topics for future research.

Spirographs are gears. They have teeth of uniform size, and typically come in sets
containing a variety of sizes, and therefore number of teeth. There are a conceivably infinite
number of different types of spirograph, since they aren’t necessarily under any restriction
other than practicality, but for our purposes we will consider only two types: the ring, and the
disk. A disk is what you might imagine, essentially a gear with holes in it where you can stick
a pencil. As the disk moves, the pencil traces the movement of a point within the disk relative
to the paper. Rings are disks laden with teeth, as all spirographs are, with the distinction that
they are hollow and have teeth on the inside edge, allowing you to place a disk within. The
disk teeth and the interior ring teeth lock together smoothly, and the disk will rotate and orbit
tangent to the interior edge of the ring, creating a continuous, repeating curve. Alternatively,
the disk can be placed to lock in with the exterior edge of the ring, creating an entirely
different set of continuous, repeating curves as it rotates and orbits.

We will examine the interior curves, although similar methods could be applied easily to the
exterior curves by adding instead of subtracting radii.

Let’s start with an experimental approach. Field mathematics, if you will. Below are plots of
every possible spirograph generated using the Spirograph Design SetⓇ. Each page was
generated by increasing the point distance from the centre of the interior disk within the pre-
cut holes.
In playing with a spirograph, a few things become apparent. First we try varying the radius of
our tracing point from the centre of the disk. As we increase the radius, we notice a similarity.
Any curve traced by a disk with a particular circumference will have the same symmetry,
regardless of the tracing radius (Fig.1-5). A particular ratio of the circumference of the disk to
the circumference of the interior edge of the ring seems to correspond to some n-fold radial
symmetry around the centre of the ring. I only have one ring, and only so many disks, so the
question of whether each ratio of disk to ring corresponds to a unique n is beyond my
experimental abilities.

Notice that as we increase the tracing radius, the curve spreads out. If it were a wave, we
might say that the amplitude is increasing. In the Spirograph set, there’s no option to put it in
the centre, but it’s not hard to see what the resulting curve might be. The radius of the ring is
constant, and because the point is in the centre of the disk, the distance from the interior
edge to the centre of the disk is constant, normal to the interior edge, so we’ll get a circle.

If we want to evaluate spirographs, it would be good to have some kind of mathematical

model. We can formulate this by considering what is happening when we trace out a
spirograph. We are considering two circles nested inside and tangent to one another, with
radii r and r such that r ≤ r . For simplicity’s sake, we will take r to be 1 for now. It’s
i e i e e

apparent, then, that the centre of the interior circle traces a path on another circle with radius
 r = re − ri (1)

22

11

-2
-2 -1
-1 00 11 22

-1
-1

-2
-2

We can consider some initial position of the interior circle, where its centre has an angle in
standard position of 0. Let x be the ratio of the distance from some point I within the circle to
i

the centre of the interior circle, P, to the radius of the interior circle. This is analogous to
picking any point I within the smaller circle, just defined in terms of its distance to P. For
some x , there's a smaller circle of all possible points which produce the same spirograph
i

which is centred at P and with radius r x . i i

By 'the same spirograph', I mean all spirographs which are symmetrical under rotational
symmetry about the origin. All points on this smaller circle must result in the same
spirograph.
Proof sketch:
Pick any two distinct points on the smaller interior circle, A and B. Each of these points
makes some angle ∠AP F and ∠BP F with the centre of the circle, P, and the point on the
circle furthest from the origin, F. As the circle moves about its orbit, it also rotates about P, so
at some point in its rotation, ∠AP F will equal 0.

As the interior circle orbits, it forms some angle t with the with the origin and the x axis. By
(1), we know that the centre of the interior circle follows a circle of radius r e − ri , so

P = ((r e − r i ) cos t, (r e − r i ) sin t) (2)

The chosen point, I, within the circle orbits around P in the opposite direction, and we have
defined it to orbit with a radius of x i ri . Combining these facts, we see that

x(t) = (r e − r i ) cos t + x i r i cos(−at))

y(t) = (r e − r i ) sin t + x i r i sin(−at))

Here, I've introduced some constant a, which is converting between the angle of the interior
circle and the angle of the tracing point I. If the circles slide without slipping, this should
equal the ratio of the circumference of the exterior circle to the interior circle. After all, in the
physical set, this rule is enforced by the teeth on the sides of the circles. All the teeth are the
same size so that they fit together, and so its just a matter of matching each tooth together
like a zipper, marching in lock-tooth. If the inside of the exterior circle has 90 teeth, and the
outside of the interior circle has 30 teeth, then the interior circle should spin 3 times = 90/30.
With some elementary school math, we find that this is a linear relationship.

Ce 2πr e re
= =
Ce 2πr i ri

Putting this all together, we get two parametric equations which define a spirograph in terms
of r , r , and x .
i e i

re
x(t) = (r e − r i ) cos t + x i r i cos (−t ) (3)
ri

re
y(t) = (r e − r i ) sin t + x i r i sin (−t ) (4)
ri

PS:
I have not been able to include it properly, but the most interesting thing about spirographs
and their length in particular is the fact that when we integrate to find their arc length, there is
no algebraic way to determine the upper bound of integration. We get something like

2πα 2πα
dx dy
L = ∫ ds = ∫ √( )
2
+ ( )
2
dt
0 0
dt dt

Where α is the smallest value of n for n re

ri
= k , k, n ∈ Z. This fact is why its important for r e

and r to be rational numbers. If the radii were irrational and coprime, the arc length would
i

diverge to infinity because there would be no upper bound on an increasing function.

Differentiating x(t) and y(t) is simple because all of the variables we have been talking about
simply fall away as constants. From (3) and (4) we get

dx re
= −(r e − r i ) sin t + x i r e sin(−t )
dt ri

dy re
= (r e − r i ) cos t − x i r e cos(−t )
dt ri

We know how to get dy

dx
from the chain rule, dy

dx
=
dy

dt
⋅
dt

dx
, but it is not immediately obvious
how to get dt

dx
from what we know of dx

dt
.