Slack Line Analysis
Slack Line Analysis
William Conley
Abstract
1 Introduction
1.1 Motivation
1
Figure 1: Diagram of a slackline set-up
direct measurements), after which people attempt to stand, walk, jump, and
do various feats of acrobatics on the line. In the first section of this article,
we will treat a person standing on the line as a static load, and calculate the
tension in the line based solely on the person’s weight and the geometry of
the system.
When slacklining is done high above the ground, the person usually wears a
safety leash, which runs from a harness around their waist to a small metal
ring (often a climbing carabiner) around the slackline itself. As they walk,
the ring around the line is dragged along behind them, and if they fall off the
line, this safety leash will catch them. This means that the person’s weight
will be applied to the line dynamically, as they will fall a short distance before
the safety leash comes tight on the line. It is widely believed that applying
this sort of dynamic load to the line will generate significantly higher tensions
in the line, and thus in particular will place significantly higher forces on the
anchors at each end of the line. In order to understand how much force these
anchors must be able to support, it is important to know the magnitude
of the tensions that can be generated in a slackline in various situations,
including dynamic ones.
2
DISCLAIMER: Slacklining anywhere other than at ground level is
inherently dangerous and may result in severe injury or death. Do
not attempt this before seeking proper instruction from a quali-
fied professional. Furthermore, while the author of this document
can be reasonably certain of the correctness of the mathematics
contained herein, he makes no claims whatsoever as to the rele-
vance of his conclusions to anything in the real world. Do not use
the information presented here to conclude that a certain slackline
set-up is safe.
1.2 Notation
y
D
x0
(0,0) x (D,0)
T1
T2
mg
Variables:
t = time
(x(t), y(t)) = position of load relative to the left anchor (origin)
3
Constants:
The last two items will vary from one scenario to another, but will be constant
throughout any given scenario.
Unknown:
We will assume that the webbing stretches in a uniform (i.e. linear) way
along its entire length. (This has nothing to do with the tension F (λ) being
linear as a function of λ, and should be a perfectly reasonable assumption.)
We will also assume that the load is initially applied at (x0 , 0) when the
line is straight and horizontal, and that it never slides left or right along the
webbing as the webbing moves. (That point on the webbing may still move
left or right as it moves up and down, but we are assuming the load is fixed
to that point and does not slide.) In other words, it is as if the webbing were
cut at the point where the load is applied, and both of the now separate
segments to the left and to the right were affixed to the load. To simplify
the terminology, we will call the webbing to the left of the load the “left
segment” and the webbing to the right of the load the “right segment”. This
implies the following:
x0
Unstretched length of left segment: D
L
x0
Unstretched length of right segment: (1 − D
)L
4
p
Stretched length of left segment: x2 + y 2
p
Stretched length of right segment: (D − x)2 + y 2
√
x2 +y 2
Tension in left segment: T1 = F x0
L D
√
(D−x)2 +y 2
Tension in right segment: T2 = F x
(1− D0 )L
Direction vector for T1 : √ −x , √ −y
x2 +y 2 x2 +y 2
√ D−x2 2 , √ −y
Direction vector for T2 :
(D−x) +y (D−x)2 +y 2
Note that the above expressions for the tensions T1 and T2 obey a certain
symmetry: replacing x0 with D − x0 in the expression for T1 will give the
expression for T2 , and vice versa. This is due to the symmetry in the geometry
of the system: the tension in the left segment when the load is x0 meters
from the left anchor should be exactly the same as the tension in the right
segment when the load is x0 meters from the right anchor, i.e. D − x0 meters
from the left anchor. This same symmetry will hold in all of the expressions
that we will derive for T1 and T2 , so in what follows, we will show fewer
details about the expressions for T2 .
2 Preliminary Analysis
If the load is not moving, then the sum of the three force vectors on the load
(tensions T1 and T2 , and gravity) is zero. This gives us the following two
equations:
−x
T1 · √x2 +y2 + T2 · √(D−x)2 +y2 = 0
D−x
(1)
T1 · √ −y
2 2
+ T2 · √ −y
2 2
− mg = 0
x +y (D−x) +y
5
Solving these simultaneously gives
s
2
x x
T1 = mg 1 − +1
D y
s 2
x D−x
T2 = mg +1
D y
If we allow for a moving load, then by Newton’s second law, equation (1)
becomes the following system of differential equations:
√ √
x2 +y 2 −x (D−x)2 +y 2
√ √ D−x2 2 = mx00
F +F
x0
(1− xD0 )L
D
L x2 +y 2 (D−x) +y
√ √ (2)
x2 +y 2 −y (D−x)2 +y 2 −y 00
F √ +F √ − mg = my
x0
D
L x2 +y 2 (1− xD0 )L (D−x)2 +y 2
6
Here we have replaced T1 and T2 by the expressions derived earlier for them
in terms of the function F (λ). Note that in this form, the equations are
a coupled system of second order ordinary differential equations, with the
single independent variable t. If, at some point, enough data on the behavior
of the function F becomes known, then it would be an easy matter to nu-
merically solve this system of equations for various values of the parameters
and appropriate initial conditions. This could potentially yield much more
accurate predictions than the results presented here.
Therefore, in all that follows, we will concern ourselves solely with the second
of these differential equations, which we write in standard form as
K
y 00 + y = −g (4)
Lm xD0 1− x0
D
Recall that the conclusion of Section 2.1 was that although we could not
predict this value, it is extremely useful, for the case of a static load in a real
world setting, to measure this value. It will turn out that, even though we
are now dealing with a dynamic load rather than a static one, this quantity
will play an extremely important role. In fact, in all the results that follow,
every quantity that matters to us can be expressed in terms of ∆, other
parameters related to the geometry of the system, and the weight of the load
(mg), thus eliminating any need to know K or L, which are more difficult to
measure. As an example of this, we can now rewrite the expressions for T1
and T2 derived in section 1.2:
p !
x2 + y 2
T1 = F x0
D
L
p
x2 + y 2
= K· x0
D
L
p
x0 x2 + y 2
= mg 1 −
D L · mg x0 x0
· 1 −
p K D D
x0 2
x +y 2
= mg 1 − (7)
D ∆
p
x0 x2 + y 2
T2 = mg (8)
D ∆
Finally, note in (5) that y(0) = C1 − ∆ and y 0 (0) = C2 ω. If we use the initial
conditions y(0) = y0 , y 0 (0) = −v0 , and solve for the constants C1 and C2 ,
then our general solution takes the form
v0
y = (y0 + ∆) cos(ωt) − sin(ωt) − ∆ (9)
ω
8
We will be interested in two quantities related to this solution: the maximum
tensions in the line, and the maximum force that acts on the load (or more
specifically, the maximum “g force”, i.e. acceleration, felt by the load). For
the first, note that the tensions in the line increase with y 2 , so to find the
maximum tension, we will want to find the maximum value of y 2 , which
will occur when
q y is at its absolute minimum. Since the amplitude of the
2
oscillator is (y0 + ∆)2 + vω0 , this minimum value of y will be
r v 2
0
ymin = − (y0 + ∆)2 + − ∆.
ω
The maximum force on the load will occur when the upward acceleration of
the load is the greatest (which also happens to be when y is at its absolute
minimum.) Since the acceleration of the load is
to ω 2 (|ymin | − ∆)), it follows that the maximum acceleration felt by the load
is simply this amplitude plus g:
2
2 ω |ymin |
amax = ω (|ymin | − ∆) + g = (|ymin | − ∆) + 1 g = g
g ∆
9
3.1 First Model
This means that the distance of the free-fall was only l − ∆ (where l is the
length of the leash plus the length of the person’s legs), and that the vertical
position of the line when the load is applied p is y = 0. Hence our initial
conditions (section 2.2) should be y0 = 0, v0 = 2g(l − ∆). This gives us
r
2g(l − ∆) √
ymin = − ∆2 + 2
− ∆ = − 2l∆ − ∆2 − ∆
ω
(using the fact that ∆ = ωg2 (6)). Finally, using (7) and (8), the maximum
tensions in the line are as follows:
p
x0 x20 + ymin2
T1 = mg 1 −
D q ∆
2
√
2
x0 x0 + 2l∆ + 2∆ 2l∆ − ∆
= mg 1 −
D s ∆
r
x 0 x 0 2 l l
= mg 1 − +2 +2 2 −1
D ∆ ∆ ∆
s
2 r
x0 D − x0 l l
T2 = mg · +2 +2 2 −1
D ∆ ∆ ∆
The maximum acceleration that the person falling will feel is
r !
|ymin | l
amax = g= 2 −1+1 g
∆ ∆
10
3.2 Second Model
The major assumption in our first model is that the line returns all the way
to horizontal before the safety leash begins to pull it back down. This makes
the model a sort of “best case” scenario, because the distance of the free-
fall is shortened, and the line begins absorbing the energy of the fall from
a higher point. Thus the expressions derived in the previous section should
be considered a lower bound for the tensions involved. To obtain an upper
bound, we will consider an unrealistic, but clearly worst-case, scenario. We
will assume in this model that when the person falls from the slackline, the
line does not snap back up at all, but rather stays in place as if held there
by some unseen force.
Thus the distance of the free-fall will be exactly l, the length of the leash plus
the length of the person’s legs, and the vertical position of the line when the
√ namely y = −∆.
load is applied will be the same as before the person falls,
Hence our initial conditions should be y0 = −∆, v0 = 2gl. This gives us
√
r
2gl
ymin = − − ∆ = − 2l∆ − ∆
ω2
Thus the maximum tensions in the line in this model are as follows:
p
x0 x20 + ymin2
T1 = mg 1 −
D q ∆
√
2
x0 x 0 + 2l∆ + 2∆ 2l∆ + ∆2
= mg 1 −
D s ∆
r
x 0 x 0 2 l l
= mg 1 − +2 +2 2 +1
D ∆ ∆ ∆
s
2 r
x0 D − x0 l l
T2 = mg · +2 +2 2 +1
D ∆ ∆ ∆
The maximum acceleration that the person falling will feel is
r !
|ymin | l
amax = g= 2 +1 g
∆ ∆
11
3.3 Discussion
The final form of all the expressions derived in this section have involved only
the weight mg of the load (person on the slackline), a few geometric mea-
surements (D, x0 , and l), and our magical quantity ∆, the distance that the
line sags with a non-moving load (at horizontal position x0 ). Unfortunately,
right now, it is more or less impossible to predict the value of ∆ for any
given slackline set-up or any position x0 on the line. However, since we can
measure ∆ directly, these calculations still may have tremendous practical
significance. For example, if I am walking on a slackline that spans 55 feet,
and it sags 5 feet in the middle under my weight, then I can compute that
the tension in either side of the line is roughly 2.8 times my body weight
(section 2.1). Recall that this result is based purely on geometry and simple
physics, and is more or less irrefutable.
Now, suppose this slackline is high above the ground (perhaps at the rim of
Yosemite Valley), and I fall off and am caught by a leash that is 5 feet long.
Since I am about 6 feet tall, my waist was roughly 3 feet above the line before
the fall, and is now 5 feet below it, so l = 8 in this scenario. The results
of this section imply that the maximum tension in the line during the fall is
between 3.01 and 3.08 times my body weight, that the maximum distance
below the anchors that my feet will reach is somewhere between 20.4 and
21.9 feet, and that I feel between 2.5 and 2.8 g’s at the most. Unfortunately,
all of these results are based on a major assumption: that the tension in the
slackline is a linear function of the factor by which it has stretched. (This
was the assumption that F (λ) = Kλ for some K in section 2.2. It more or
less says that the webbing obeys Hooke’s Law, or that its Young’s modulus is
constant.) But, as anyone who has experience setting up slacklines probably
knows, this assumption is false for tubular nylon webbing. For one thing,
the elastic properties of webbing are known to change as it ages, or when it
gets wet, or when it is exposed to ultraviolet light for long periods of time.
But even disregarding these factors, it is common to set up a slackline and
find that it sags a certain distance at first, and then a few minutes later find
that it sags more under the same load and needs to be pulled tighter. This
implies that its Young’s modulus decreases when it is kept under fairly high
tension.
However, before we throw away most of the results of this paper, consider this.
12
All of the situations just mentioned involve the Young’s modulus changing
over a period of minutes or months or years. But as the example above
demonstrates, the results gathered here are intended to be used by roughly
measuring the amount that the line sags (∆) when a person is just standing
on it, and using this value to calculate the various other quantities that
would arise if that person took a leash fall at that instant. It may not
be too unrealistic to assume that, during the split second in which a fall
occurs, the Young’s modulus does remain more or less constant, and thus
that the models developed here are valid. On the other hand, this may be
completely untrue, because it is generally believed that a leash fall generates
substantially higher tensions in a slackline than a static load generates, and
it may be that this sort of sudden increase in tension makes the nonlinearity
apparent in the webbing. We saw in the example above (and we will see
more in the graphs in the next section) that if the models developed here are
valid, then the tensions generated during a leash fall are not actually so high
(except perhaps when they occur very close to one end of the line), so that
argument may not hold. Admittedly, this is a bit of circular reasoning, but
it does seem to provide some evidence that this model may be valid, at least
for falls that occur in the middle of the line (where we know that the tensions
on each side of the line are equal). Furthermore, it is worth mentioning that
what we have done here is one version of a standard trick in the analysis of
nonlinear dynamical systems. We have found an equilibrium (the position of
the line when not moving) and we have linearized our differential equations
about that equilibrium.
4 Results
Since all of our expressions for the two tensions T1 and T2 obey the symmetry
described in section 1.2, we need only present data for one of these tensions.
We have chosen to include only T1 in this summary and the graphs below.
13
4.1 Summary of Formulas
Notation:
D = Horiz. dist. b/w anchors
x0 = Horiz. dist. of person from anchor
l = Length of leash + length of legs
∆ = Depression in line when person is not moving
14
4.2 Observations
All of the expressions for tensions listed above are directly proportional to
mg, the weight of the load. Hence, in the graphs below, we have chosen
to present these tensions as simple numbers, which must be multiplied by a
person’s weight in order to find the actual tension that person would put on
a slackline in that scenario. Similarly, the two expressions for the maximum
acceleration are completely independent of the mass of the load, and are
proportional to g. For this reason, in the list of formulas above and the
graphs below, we have chosen to divide these expressions by g. This literally
gives the g force that the person will feel during a leash fall.
In all of these expressions, all of the length measurements are relative. (Note
that in the final form of each expression, every variable representing a length
or distance appears in a fraction with another such unit.) This implies that
we can consider all lengths to be measured relative to one particular length,
say D, the distance between the two anchors. For example, if a person is
standing 3 meters out on a slackline with a 10-meter span and the line sags
0.5 meters, then the tensions in the line would be exactly the same as if that
person were standing 6 meters out on a 20-meter line and the line sagged 1
meter.
A pleasant side efect of this is that the units used to measure lengths do not
matter at all, provided that the same units are used for all lengths. If you
prefer to measure all your lengths in meters, but your weight in pounds, the
formulas presented here will work for you with no modifications, and will
return tensions measured in pounds. If you are a climber and prefer to think
of forces in kN, but you think in feet rather than meters, you can convert
your weight to kN, and leave all the length measurements in feet, and all the
tensions will come out in kN.
4.3 Graphs
15
12
Tension T1 (as a multiple of load mg)
10
3 D D
4
D 4 2
4
D D
25 10
2
D
50
0 D D D D D D D D
100 50 40 30 20 12 10 8
∆ (measured relative to D)
D, as are the values of ∆ along the horizontal axis. The graph makes it
appear that the tension is highest when the person is in the middle of the
line, and is much lower near the anchors, but this is misleading; keep in mind
that ∆ will be much higher in the middle of the line, and will be very small
near the anchors. The graph does point out one interesting thing: the fact
that the curves for D4 and 34 D are almost identical implies that when the
person is standing 41 of the way out from one anchor, the tensions on either
side of the line are nearly the same (because ∆ will certainly be the same at
x0 = D4 and x0 = 34 D).
Figure 4 shows that in a leash fall with l = 8 ft, g forces above 4g are unlikely
unless ∆ is less than 1 foot. For long slacklines, this usually only occurs near
the anchors, but for short, tight highlines, it means that leash fals are going
to hurt! It also shows that, unless the line sags a lot under body weight, a
16
6g
5g
g Force
4g
3g Upper bound
2g Lower bound
1g
0 1 2 3 4 5 6
25
Fall Dist. (feet)
Upper bound
20
15 Lower bound
10
5
0 1 2 3 4 5 6
∆ (feet)
Figure 4: Max g force and drop distance (below anchors) during a leash fall
with l = 8
leash fall shouldn’t drop a person more than 25 feet. Recall that these two
quantities depend only on l and ∆, so Figure 4 applies to a line of any length.
Figure 5 may be the most interesting (and perhaps shocking) in this section.
It shows the maximum tension in one side of a 50-foot slackline (i.e. the
maximum load put on one of the anchors) in the event of a leash fall, com-
pared to the tension in the line right before the fall when the person was just
standing. When the fall occurs in the middle of the line, the difference be-
tween the static load and the maximum tension is only about 0.5 times body
weight, and when the fall occurs 10 feet from the anchor, it’s only about 1
times body weight. Only when the leash fall occurs very close to the anchor
is the tension significantly higher, and even then it’s “only” a difference of 3
to 4 times body weight (2 to 3 kN for an average-sized person).
17
Middle of line (x0 = 25)
10
Tension before fall
8 Max tension (l.b.)
6 Max tension (u.b.)
4
2
0
1 2 3 4 5 6
Tension T1 (multiple of load)
These graphs also show that the lower and upper bounds that we established
earlier are quite close together, which is good.
18
5 Further Analysis
The graphs presented in the previous section point out a significant weakness
of the results presented thus far. With the linearity assumption made in
section 2.2, we can do all of these calculations using a value of ∆ measured
at a particular point on a particular line. But ∆ varies significantly from
one point on a slackline to another (e.g. the line sags a lot more in the
middle than near the anchors), and we still have no way to calculate this ∆
at various points on the line.
In section 3.3, we discussed why the linearity assumption of section 2.2 might
be valid for our purposes, since many known nonlinearities of nylon webbing
seem to occur over a long period of time (minutes at least) whereas a leash
fall occurs in a split second. But now suppose we have set up a slackline,
stretched it to the desired tightness, walked on it a few times and re-tightened
as necessary, until it seems that, at least for the time being, the line sags
the same amount in the middle each time it is weighted by the same load.
(Experienced slackliners will agree that this is generally the way it goes.)
This would seem to imply that, for the time being, the Young’s modulus of
19
the webbing is not changing, and thus that our linearity assumption might
be valid.
We will not bother to write down expressions for any of the quantities in
section 3 in terms of ∆m , because they are not too pretty. However, knowing
the tension in the line under a static load is useful, and its expression is not
that ugly. Plugging (10) into the results of section 2.1 gives
s 2
D x 0 2
T1 = mg + 1− .
4∆m D
Note that this implies that the force on an anchor increases as the person
moves closer to the anchor. Fortunately, since it should be true on any real-
world slackline that 4∆Dm > 2, the force is not tremendously higher; in fact,
it varies by less than 12% over the length of the line.
Figure 7 shows two things: that the tension in the line is almost constant
regardless of where the person is standing on the line (as explained in the
previous section), and that the tensions seen during a fall (and hence the
20
0
1
2
∆m = 3
3
∆
∆m = 4
4
∆m = 5
5
6
0 10 20 30 40 50
x0 (horiz. distance from anchor)
10
Tension T1 (multiple of load)
6
Leash fall
∆m = 3
4 Static load
Leash fall
∆m = 5
2 Static load
0
0 10 20 30 40 50
x0 (horiz. distance from anchor)
forces put on the anchors during a fall) are not much greater, except when
the fall occurs very close to the anchor. This figure once again uses the values
D = 50 and l = 8 (i.e. a 50-foot slackline and, perhaps, a 6-foot tall person
using a 5-foot leash.)
21
6 Conclusions
We have found that by assuming that a slackline, once it has been set up and
stretched appropriately, is linearly elastic, it is possible to calculate several
important quantities based only on easily observable geometric parameters
of the system. These calculations imply, among other things, that the load
put on an anchor during a leash fall is not much higher than the load placed
on the anchor by simply walking on the line, except when the fall occurs very
close to the anchor (within a few feet).
The results of the leash fall models could be tested directly as well, using a
load cell and the right testing apparatus. Regardless of the validity of the
results presented here, it would be a Very Good Thing for the slacklining
community to know the results of this sort of testing, so that we would know
the kind of forces our anchors need to withstand, particularly in highlining
scenarios.
One final point is that the calculations for leash falls done here were based
on a completely static leash. In the real world, the leashes used in highlining
are somewhat dynamic, and when a fall occurs very close to an anchor, this
should reduce the load placed on that anchor. It is possible that, if the leash
is dynamic enough, this may reduce the load significantly, but that would be
the subject of another study.
22