M ODELLING A CASSETTE TAPE REWINDING PROCESS

June 20, 2015

J.L.
 1

Firstly a word of warning. The following analysis and derivation are trivial
and don’t really give any help for practical re-use of cassette tapes in guitar
effects or any other applications.

The background for this writing is that I have always wanted to derive a
useful mathematical model for a mechanical system by first finding a proper
differential equation and then solving it into an equation to get the state of
the system in any point of time. Unfortunately this is not it. Here I just use
a given formula, differentiate it, and integrate it back to the same form as it
was earlier. Basically the same result could have been obtained directly by
tweaking the initial simple formula to contain time dependence.

But maybe one day I will manage to come up with something very clever...

As a youngster I had always thought that the cassette tape must travel at a
constant speed at all times. When I started thinking about it more carefully
(several years of hard pondering ...) I came to the conclusion that the speed
of the tape is never constant, only the angular speed of the reel is constant.
But then I wanted to get my thougths officially confirmed just in case. After
doing some ’serious’ research (using www search engines) I had to throw all
my hard pondering work to the trash bin. The tape speed is actually constant
in play mode, where the capstan and pinch roller control the linear speed of
the tape making it constant. So nothing interesting there then.

But the pinch roller is not pinching when rewinding the tape, so then I
wanted to know how the tape speed changes during rewind. The capstan
is still there rolling along, but most likely it does not affect much during
rewind. I tested this assumption by rewinding a tape into an empty reel
for 60 seconds then flipping the cassette and rewinding the same amount of
tape back into the almost full reel. This took only 37 seconds, so the linear
speed indeed depends on the radius of the reel in the rewind process. The
following calculation analyses tape rewind mathematically.

The ’problem’ should be made more clear using an illustration, which is given
in Figure 1. Here the left reel turns clockwise at a constant angular velocity
ω, and the linear speed of the tape v is measured from the linear section.

For this problem, basic physics offers the equation

v = ωr = f 2πr, (1)
2

ω
r

h
v

Figure 1: Tape inside a C-cassette (simplified sketch drawing)

where the linear velocity v is directly proportional to the product of angular

velocity ω and the radius r of the rotating round object, which in this case
consists of the cassette reel and the rolled tape on top of the reel. In addition,
we know that ω = 2πf , where f in this particular context designates the
frequency of full revolutions of the reel. Clearly during one full revolution,
the tape travels the distance 2πr.

Now, as the time passes, more and more tape is packed on top of the reel
and the radius is therefore dependent of time. Based on this, the following
details on the time dependency can be added to our little equation:

v(t) = f 2πr(t) ; v(0) = f 2πr0 . (2)

The initial velocity measured in the origin of time is when the radius has the
constant value r0 . Or in other words, r0 is the radius of the reel at time t = 0.

Next we want to know how the velocity changes with respect to time, so we
operate with the differential operator on both sides of the equation. This is
the result of the operation:
dv(t) dr(t)
dt = f 2π dt. (3)
dt dt

Now the brain is needed to figure out how to express the rate of change of
the radius r(t). After many nights without sleep, this came into mind:
dr(t)
= f h, (4)
dt
where f is still the same frequency of revolution and h designates the height
(or thickness) of the cassette tape. Practically this means that with every full
revolution the radius grows by the thickness of the tape. At least to my mind
 3

it makes sense. According to wikipedia, typical thickness of a cassette tape

is between 10 − 16 µm. These values can be used for obtaining numerical
results from the resulting equation.

With the new expression for the rate of change of the radius, the differential
equation reads
v̇(t)dt = f 2 2πh dt. (5)

Now we can solve the equation by integrating it:

Z Z
2
v̇(t)dt = f 2πh dt, (6)

and the result of integration is

v(t) = f 2 2πh(t + C), (7)

where C is the integration constant. To get the complete solution, the value
of the integration constant needs to be determined. Since it is known that at
time t = 0, v = f 2πr0 , C can be solved from
r0
v(0) = f 2πr0 = f 2 2πhC ⇒ C= . (8)
fh

And from here it follows that

v(t) = f 2πr0 + f 2 2πht, (9)

which is the final solution.

As a side note, it is also possible to include the initial value directly in the
integration phase. All that is needed is to integrate from the initial time value
to t, which is the same generic time variable (not a constant!) as already used
inside the velocity function v(t). Commonly here the integration differential
variable dt is noted as dt̂ or by some completely other letter and claimed it to
be a ’dummy’ variable. This is only possible in pure mathematics where the
concept of differential is not recognised. In physics the differential represents
a very small quantity with dimensions and therefore cannot be treated as a
’dummy’ variable.

Anyway, the equation (6) can be written using the initial value in the inte-
gration limits as:
Zt Zt
2
v̇(t)dt = f 2πh dt. (10)
t=0 t=0
4

When performing the simple integrations on both sides, the remaining equa-
tion reads:
v(t) − v(0) = f 2 2πht − 0. (11)

After inserting the initial value v(0) = f 2πr0 into (11) one has the same
solution as in (9):
v(t) = f 2πr0 + f 2 2πht. (12)

Well, it is obvious that this is a linear relation with respect to time, so one
cannot even draw any neat graphs based on this. Already from the basic
equation (1) it is clear that when the radius doubles also the linear velocity
of the tape doubles. The obtained equation can be used for estimating the
tape speed at any given time after the tape started rolling at time t = 0. The
initial reel radius r0 can be chosen freely.