2 A crate of wieght $F_g$ is pushed by a force $\vect{P}$ on a
3 horizontal floor as shown in Figure~P5.63. The coefficient of static
4 friction is $\mu_s$, and $\vect{P}$ is directed at an angle $\theta$
5 below the horizontal. \Part{a} Show that the minimum value of $P$
6 that will move the crate is given by
8 P = \frac{\mu_s F_g \sec\theta}{1 - \mu_s \tan\theta}
10 \Part{b} Find the condition on $\theta$ in terms of $\mu_s$, for
11 which motion of the crate is impossible for any value of $P$.
17 pair p = (0.3u, -0.3u);
19 Surface s = Surface((-p.x,0), (1u+p.x,0));
21 Block b = Block((0.5u, 0.35u), width=1u, height=0.7u, "crate");
23 Vector P = Force((0,0.7u)-p, mag=length(p), dir=degrees(p), "$\vect{P}$");
31 The normal force must resist both the force of gravity and the
32 vertical component of $\vect{P}$, so
34 N = F_g + P\sin(\theta) \;.
36 This moves the crate when the horizontal component of $\vect{P}$
37 balances the force of friction.
39 P\cos(\theta) &= \mu_s N = \mu_s (F_g + P\sin(\theta)) \\
40 P(\cos(\theta) - \mu_s \sin(\theta)) &= \mu_s F_g \\
41 P(1 - \mu_s \tan(\theta)) &= \mu_s F_g \sec(\theta) \\
42 P &= \ans{\frac{\mu_s F_g \sec(\theta)}{1 - \mu_s \tan(\theta)}} \;,
44 which is what we set out to show.
46 Note that this formula is only valid when there is an actual normal
47 force to provide friction. Therefore $P\cos(\theta) > 0$. We can
48 posit, without loss of generality, that $P>0$, in which case the
49 restriction is $-90\dg < \theta < 90\dg$. By symmetry, the situation
50 for the backside $180\dg$ is just a mirror image of the frontside.
53 As $P$ becomes larger, the $F_g$ component of our horizontal force
54 balance becomes negligable, so we cannot move the block when
56 P\cos(\theta) &\le \mu_s P \sin(\theta) \\
57 \frac{1}{\mu_s} &\le \tan(\theta) \\
58 \theta &\ge \ans{\arctan\p({\frac{1}{\mu_s}})} \equiv \theta_c \;,
60 where the last step uses the fact that $\tan(\theta)$ is strictly
61 increasing on the range $\theta\in(-90\dg,90\dg)$.
63 What does this mean about our answer to \Part{a}? Let's rework the
64 condition to look more like the denominator in the \Part{a} answer.
66 \tan(\theta) &\ge \frac{1}{\mu_s} \\
67 0 &\ge \frac{1}{\mu_s} - \tan(\theta) \\
68 0 &\ge 1 - \mu_s \tan(\theta) \;,
70 so the denominator is negative or zero for $\theta \ge \theta_c $.
71 For $\theta$ just below the cutoff, the denominator is small but
72 positive, and you get a really large value for $P$. For
73 $\theta=\theta_c$, the denominator is zero, and you get an infinite
74 value for $P$. For $\theta$ above the cutoff, the denominator is
75 negative, so $P$ is also negative, which, as I pointed out
76 in \Part{a}, is not allowed.
78 The whole thing is a bit easier to understand if we rephrase the
81 P = \frac{\mu_s F_g}{\cos(\theta) - \mu_s \sin(\theta)}
82 = \frac{C}{\cos(\theta) - \mu_s \sin(\theta)}
83 = (A\cos(\theta) - B\sin(\theta))^{-1} \;,
85 where $C=\equiv \mu_s F_g$, $A\equiv 1/C$, and $B\equiv \mu_s/C=1/F_g$.
86 We can consolidate to a single trig term using
88 \sin(a \pm b) &= \sin(a)\cos(b) \pm \cos(a)\sin(b) \\
89 (D\sin(a - b))^{-1} &= (D \sin(a)\cos(b) - D\cos(a)\sin(b))^{-1} \;.
91 Matching with our formula,
96 \tan(a) &= \frac{A}{B} = \frac{1/C}{\mu_s/C} = \frac{1}{\mu_s} \\
97 a &= \arctan\p({\frac{1}{\mu_s}}) \\
98 D &= \frac{B}{\cos(a)}
99 = \frac{B}{\cos\p({\arctan\p({\frac{1}{\mu_s}})})}
100 = B\sqrt{1+\frac{1}{\mu_s^2}}
101 = \frac{\sqrt{1+\frac{1}{\mu_s^2}}}{F_g} \\
102 P &= D^{-1} \p({ \sin(a - b) })^{-1}
103 = \frac{F_g}{\sqrt{1 + \frac{1}{\mu_s^2}}}
104 \csc\p({\arctan\p({\frac{1}{\mu_s}}) - \theta}) \;.
106 This doesn't look as clean as the phrasing in \Part{a}, but it makes
107 the dependence of $P$ on $\theta$ much clearer. For example, $P$ is
108 obviously negative for $\theta > \theta_c \equiv \arctan(1/\mu_s)$.
109 The dependency on $\theta$ over the rest of the range is
111 P \propto \csc(\theta_c - \theta) = \frac{1}{\sin(\theta_c - \theta)}
113 Because $\mu_s$ is a positive number, $1/\mu_s$ will also be positive,
114 and $\theta_c$ will be between $0$ and $90\dg$. The status on all
115 possible angles looks something like
126 Angle Atc = Angle((1,0), (0,0), dir(tc), "$\theta_c$");
129 draw(scale(r)*unitcircle);
131 draw((0,-R)--(0,R), red);
132 draw((-dir(tc)*R)--(dir(tc)*R), blue);
134 label("$\cos(\theta)>0$", L*dir(0), red);
135 label("$P > 0$", L*dir(tc-90), blue);
138 Taking the $\cos(\theta)>0$ portion of our $P$ dependence (where the
139 equation we started with in \Part{a} applies, and combining it with
140 the reflection (which applies when $\cos(\theta)<0$, we get
143 \infty & \text{if $\theta_c \ge \theta < 180\dg-\theta_c$} \\
144 \frac{F_g}{\sqrt{1 + \frac{1}{\mu_s^2}}}
145 \csc\p({\arctan\p({\frac{1}{\mu_s}}) - \theta})
146 & \text{if $-90\dg \le \theta < \theta_c$} \\
147 \frac{F_g}{\sqrt{1 + \frac{1}{\mu_s^2}}}
148 \csc\p({\arctan\p({\frac{1}{\mu_s}}) - 180\dg + \theta})
149 & \text{if $180\dg - \theta_c < \theta \le 180\dg$
150 or $-180\dg \le \theta \le -90\dg$} \\
165 real tc_r = tc*pi/180; // theta_c in radians
166 real tc_buf = (tc-25)*pi/180; // weakened version of tc_r
168 real r_scale(real r) { return u/2 * r; }
169 real R(real theta) { return r_scale( 1/sin(tc_r-theta) ); }
170 real L(real theta) { return R(pi-theta); }
171 real onef (real theta) { return r_scale(1); }
172 real twof (real theta) { return r_scale(2); }
175 draw((0,0)--(u*dir(tc)));
176 draw((0,0)--(u*dir(180-tc)));
178 draw(polargraph(onef, -pi-tc_r, tc_r));
179 draw(polargraph(twof, -pi-tc_r, tc_r));
180 label("$1$", r_scale(1)*dir(tc-90), SE);
181 label("$2$", r_scale(2)*dir(tc-90), SE);
183 draw(polargraph(R, -pi/2, tc_buf), blue);
184 draw(polargraph(L, pi-tc_buf, 3pi/2), blue);
186 Angle Ainf = Angle(dir(180-tc), (0,0), dir(tc), radius=0.7u, red, "$\infty$");
188 Angle Atc = Angle(dir(0), (0,0), dir(tc), "$\theta_c$");
192 where I've just plotted the $\theta$ dependence of $P$, setting the
193 constant $F_g/\sqrt{ }$ term equal to $1$. Note that the $\csc$ makes
194 nice, straight lines in this polar plot.