1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \documentclass[twocolumn,aps]{revtex4}
3: %\documentclass[preprint,aps]{revtex4}
4: %\documentstyle[preprint,eqsecnum,aps]{revtex}
5: %\documentstyle[twocolumn,eqsecnum,aps]{revtex}
6: \usepackage{graphicx}
7: \def\btt#1{{\tt$\backslash$#1}}
8: \def\BibTeX{\rm B{\sc ib}\TeX}
9: \begin{document}
10: \title{`Why does not a spinning top collapse?' : Intuitive explanations}
11:
12: \author{Won-Young Hwang$^*$}
13:
14: \affiliation{ Department of Physics Education, Chonnam National
15: University, Kwangjoo 500-757, Republic of Korea}
16:
17: \begin{abstract}
18: We elaborate an intuitive explanation by Barker [E. M. Barker, Am.
19: J. Phys. {\bf 28}, 808 (1960)] about why spinning top does not
20: collapse. We present the explanation in conceptually more rigorous
21: form. We calculate full trajectories involved. We discuss nutation
22: of spinning top. We describe how Eastman's explanation is
23: incomplete.
24:
25:
26:
27: %\noindent{PACS: 03.67.Dd}
28: \end{abstract}
29: \maketitle
30:
31: \section{Introduction}\label{sec:intro}
32: `Why does not a spinning top collapse?' This is a problem that has
33: puzzled many physicists. Standard solution can be found in
34: textbooks, e.g. Refs. \cite{Hal93,Fey63,Tho95}: Torque on a top is
35: given such that angular momentum of the top makes a closed loop
36: around the vertical axis. The solution is simple and rigorous.
37: However, the solution is not satisfactory in that it is not
38: intuitive enough.
39: %This is an important point because many physicists still do
40: %not think they understand motions of top even after knowing the
41: %solution.
42: %Moreover, it does not uniquely determine a motion of the top, as we
43: %will describe later.
44:
45: Thus physicists have proposed intuitive explanations. In 1960,
46: Barker gave the first one \cite{Bar60}. Then, Eastman \cite{Eas75}
47: and Edwards \cite{Edw77} gave another. The explanation by Eastman
48: and Edwards is interesting. But it is somewhat misleading and
49: confusing, which will be discussed in section IV.
50:
51: The explanation by Barker is beautiful and in correct direction
52: \cite{Bar60}. However, it was not fully elaborated. First, it is not
53: conceptually rigorous. Second, full trajectories are not given. The
54: purpose of this paper is to complete the explanation in the two
55: points.
56:
57: %give explanations that resolve the unsatisfactory points: Our
58: %explanation are very intuitive. It considers actual trajectories of
59: %parts of a top and it clearly shows what force supports the top:
60: %Loosely speaking, {\it what supports a top is internal force by some
61: %parts of the top.} (A loose but short explanation for impatient
62: %readers is given in Fig. 3 and section II.) Furthermore, our
63: %explanations can be easily applied to motions of the same top but
64: %with different initial conditions, as we discuss in Section III.
65:
66: This paper is organized as follows. In section II-A, we consider a
67: simple toy model to make the nature of the explanation clear. In
68: section II-B, we give the explanation for the spinning top. In
69: section III-A, we discuss the nutation of a spinning top. In section
70: III-B, we discuss Eastman's explanation. In section IV, we conclude.
71:
72: \section{Why does not a spinning top collapse?}\label{sec:II}
73:
74: %In subsection A, we consider a simple illustrating example, a
75: %rotating object in Fig.1. We give standard explanations
76: %(Explanation-1 and Explanation-2) and our explanations
77: %(Explanation-3 and Explanation-4). In subsection B, we give our
78: %explanations, Explanation-5 and Explanation-6, for the spinning top.
79: %Explanation-5 and Explanation-6 are very close to Explanation-3 and
80: %Explanation-4, respectively, in their spirit.
81:
82: \subsection{A rotating object}
83:
84: \begin{figure}[htbp]
85: \includegraphics[width=8cm,height=5cm]{Fig-1.ps}
86: \caption[Fig1]{\label{fig1} A rotating object composed of point-mass
87: with mass $m$ and a massless rod. %Force $\vec{f}$ makes floating
88: %force since the other end of rod is fixed.
89: }
90: \end{figure}
91: Let us consider a model, an object composed of a point-mass with
92: mass $m$ and a rod in Fig.1. We assume that the mass-point rotates
93: with a fixed angular speed making a circle drawn with dotted line.
94: For simplicity of discussion, we assume that the rod has negligible
95: mass. The rod has a joint a origin $O$. That is, its upper end is
96: fixed at $O$ but it can freely rotate. Normally, we assume that the
97: hinge can give no torque on the rod. Here the origin with which the
98: torque is defined is $O$. Thus force by the rod, denoted by
99: $\vec{T}$, should be along the rod, like in the case of a string.
100:
101: What we consider here is a toy model. We do not need to consider the
102: model only for the purpose of explaining it. However, insights we
103: get here will be useful in understanding other more complicated
104: ones, the spinning top, later.
105:
106: What we do in this subsection is the following. First, we give
107: standard explanations, Explanation-1 and Explanation-2 for the
108: rotating object. Then we give Explanation-3, which is somewhat
109: different from the two. Next we give Explanation-4. Explanation-3 is
110: conceptually more rigorous than Explanation-4 which is of Barker
111: type.
112:
113: Now let us consider the same question that arises for the rotating
114: object: Why does not the object move downward by gravity? This
115: question can be answered in two standard ways.
116:
117: {\bf Explanation-1}: {\it Newton's laws explain the motion: Force on
118: the point-mass, $m\vec{g}+\vec{T}$, namely sum of gravity $m\vec{g}$
119: and tension $\vec{T}$, accelerates the point-mass inwardly such that
120: it makes a circular motion \cite{Hal93}.}
121:
122: {\bf Explanation-2}: {\it We adopt a non-inertial frame that rotates
123: making a circle in Fig. 1. The point-mass is co-moving with the
124: non-inertial frame. Fictitious force in the non-inertial frame is
125: centrifugal force denoted by force $\vec{f}$. The force $\vec{f}$
126: balances two other forces, that is, $m\vec{g}+\vec{T}+\vec{f}=0$.
127: Hence, the point-mass has zero acceleration in the non-inertial
128: frame. The point-mass does not move if it did not move initially, in
129: the non-inertial frame. Thus, the point-mass makes the same circle
130: as the non-inertial frame does.}
131:
132: %Here the centrifugal force $\vec{f}$ can be interpreted to give
133: %floating force for the point-mass: Since the other end of the rod is
134: %fixed, the point-mass would move upward if what is exerted on the
135: %point-mass is only the centrifugal force $\vec{f}$.
136:
137: Let us give other explanations. They are closely related to the
138: Explanation-2, but different. First, what we adopt is not
139: non-inertial frame but an `inertial frame'. Second, we focus our
140: view on `the rod' not on the object composed of the rod and the
141: point-mass.
142:
143: Our goal is to show that torque exerted on the rod is zero
144: \footnote{Exactly speaking, mass of the rod is non-zero. Thus torque
145: on the rod in vertical direction should be non-zero. In this case,
146: however, analysis is complicated. For simplicity, we assume that rod
147: has zero mass.}. Otherwise, the rod will turn making the point-mass
148: move upward or downward. Now let us estimate torque on the rod. In
149: order to do that, we need to know forces which act on the rod. There
150: are two forces on the rod. One is force by hinge. However, the force
151: by hinge gives zero torque because it acts on the origin $O$. The
152: other is force by the point-mass. Our task is to estimate force. Let
153: us consider,
154:
155: {\bf Proposition-1}: {\it The force on the rod by the point-mass is
156: given by sum of the force $\vec{f}$ and the gravity $m\vec{g}$, that
157: is, the force is $\vec{f}+m\vec{g}$. Here $-\vec{f}$ is defined to
158: be a force that is required to maintain the same motion of the
159: point-mass. (Here $\vec{f}$ is a real force. There is no centrifugal
160: force because we are in an inertial frame.)}
161:
162: We argue why Proposition-1 holds. Let us assume, in Fig.1, that the
163: gravity is removed while the motion of object is maintained somehow.
164: Force that should be given on the point-mass to maintain the
165: constant rotation is $-\vec{f}$ and the only thing that can exert
166: force on the point-mass is the rod. Thus, it should be that the rod
167: is giving a force $-\vec{f}$ to the point-mass. By the third law of
168: Newton, the force on the rod by the point-mass is $\vec{f}$.
169: Intuitive interpretation of the force $\vec{f}$ will be given later.
170:
171: Now let us assume that the gravity is restored. There can be no
172: change in the force on the point-mass since the motion of point-mass
173: is the same. However, the force is differently composed. In case of
174: zero gravity, the force on the point-mass is solely provided by the
175: that of the rod. In case of non-zero gravity, the force on the
176: point-mass is sum of that by the rod and gravity on the point-mass,
177: $m\vec{g}$. Therefore, the force by the rod must be $-\vec{f}
178: -m\vec{g}$. However, due to the third law of Newton, the force by
179: the point-mass on the rod is an opposite of the force. That is,
180: force on the rod by the point-mass is $\vec{f}+m\vec{g}$. $
181: \diamondsuit$
182:
183: It is notable that the Proposition-1 applies to not only the motion
184: in Fig. 1 but to motion of spinning top in Fig. 2.
185:
186: Using Proposition-1, we get,
187:
188: {\bf Explanation-3}: {\it The force on the rod by the point-mass,
189: $\vec{f}+m\vec{g}$, gives zero torque on the rod. Thus the rod will
190: not make turning motion that makes the point-mass move vertically.}
191:
192: Let us see what will happen if only the force $\vec{f}$ were exerted
193: on the rod. In this case, the rod gets a torque that turns the rod
194: so that the point-mass moves upward. In this sense, the rotation of
195: point-mass gives a 'floating force'. Similarly, if only gravity
196: $m\vec{g}$ were exerted, the rod get an opposite torque that turns
197: the rod so that the point-mass moves downward. We get an explanation
198: of Barker type \cite{Bar60}.
199:
200: {\bf Explanation-4}: {\it The gravity $m\vec{g}$ gives a torque to
201: turn the rod such that the point-mass moves downward. However, the
202: force $\vec{f}$ gives a torque to turn the rod oppositely such that
203: the point-mass moves upward. The two torques are balanced and thus
204: the mass-point does not move vertically.}
205:
206: \subsection{A spinning top}
207: Our arguments here are in parallel with those in the previous
208: subsection. The difference is that we deal with a spinning top as in
209: Fig. 2: The top is a simplified one that is composed of four
210: point-masses with mass $m$ and a massless frame. Initially, when
211: $t=0$, four point-masses are in the $x-y$ plane, the top spins about
212: $z$ axis with angular speed $\omega$, and the top precesses about
213: $y$ axis with angular speed $\Omega$. As usual, we consider
214: situations where angular speed $\omega$ is much greater than the
215: precession angular speed $\Omega$, $\omega >> \Omega$. The frame is
216: composed of a stem with length $R$ and four branches with length
217: $r$. Initial velocity $v$ of the point-mass $a$ is given by $v= r
218: \omega + R \Omega$.
219:
220: \begin{figure}[htbp]
221: \includegraphics[width=8cm,height=8cm]{Fig-2.ps}
222: \caption[Fig1]{\label{fig2} A simplified top composed of four
223: point-masses of mass $m$ and a massless frame. The frame is composed
224: of a stem with length $R$ and four branches with length $r$.
225: Initially, four point-masses are in the $x-y$ plane. The top spins
226: about $z$ axis with angular speed $\omega$. The top precesses about
227: $y$ axis with angular speed $\Omega$. Initial velocity $v$ of the
228: point-mass $a$ is given by $v= r \omega + R \Omega$.}
229: \end{figure}
230:
231: Now let us calculate torque on the frame with respect to origin $O$.
232: We apply the Proposition-1. What we need to derive is the force
233: $\vec{f}$.
234: %To calculate the forces $\vec{f}$, we assume that gravity
235: %is removed while the motion in Fig. 2 is maintained somehow.
236: Force $-\vec{f}$ on the point-mass by the frame can be calculated
237: from trajectory that the point-mass makes.
238: %However, by the third law
239: %of Newton, forces on the frame by the point-mass are opposite of the
240: %force on the point-mass by the frame. Thus we can get what we need,
241: %the former force, by calculating the latter force.
242:
243: Let us calculate trajectory of each point-mass in order to get
244: acceleration of it. Let us set up coordinate systems as in Fig. 3,
245: where unit vectors ${\bf \vec{x},\vec{y},\vec{z}}$ and ${\bf
246: \vec{x^{\prime}}, \vec{y^{\prime}},\vec{z^{\prime}}}$ are
247: instantaneous rectangular coordinate systems co-moving with the
248: stem.
249:
250: \begin{figure}[htbp]
251: \includegraphics[width=8cm,height=8cm]{Fig-3.ps}
252: \caption[Fig1]{\label{fig3} A trajectory that the point-mass $a$
253: makes. The trajectory is curved toward origin $O$ as well as toward
254: a point $(0,+\infty,0)$. Thus force on the point-mass $a$ along the
255: $z$ direction is non-zero. Reaction of the point-mass on the branch
256: along the $z$ direction on the branch, $\vec{h}_a(0)$, is an
257: opposite of the force. However, the other end of the stem is fixed
258: at the origin $O$ and the frame composed of the stem and four
259: branches is a solid body. Thus, the frame would turn such that the
260: point-mass moves upward, if the torque by the reaction
261: $\vec{h}_a(0)$ were the only one exerted on the frame. Here unit
262: vectors ${\bf \vec{x},\vec{y},\vec{z}}$ and $\bf \vec{x^{\prime}},
263: \vec{y^{\prime}}, \vec{z^{\prime}}$ are instantaneous rectangular
264: frames co-moving with the stem.}
265: \end{figure}
266: The relationship between the two coordinate systems is given by
267: \begin{equation}
268: \label{A} {\bf \vec{ y^{\prime}} = \vec{y}}, \hspace{5mm}
269: {\bf \vec{x^{\prime}}}= \bf \cos \Theta \cdot\vec{x}- \sin \Theta
270: \cdot {\vec z},
271: \end{equation}
272: as we see.
273:
274: Using the coordinate systems we calculate trajectory of each
275: point-mass. First, we calculate trajectory of point-mass $a$, as
276: described in Fig. 3. Let us denote the trajectory of point-mass $a$
277: as $\vec{L}_a(t)$. The vector $\vec{L}_a(t)$ is sum of a location
278: vector $\vec{R}(t)= R \hspace{1mm} \vec{{\bf R}}(t)$ and another
279: location vector $\vec{r}(t)= r \hspace{1mm}\vec{{\bf r}}(t)$, that
280: is,
281: \begin{eqnarray}
282: \label{B} \vec{L}_a(t)= \vec{R}(t)+ \vec{r}(t),
283: \end{eqnarray}
284: where $\vec{{\bf R}}(t)$ and $\vec{{\bf r}}(t)$ are unit vectors as
285: shown in Fig. 3. We can see that
286: \begin{equation}
287: \label{C} \vec{R}(t)= R \hspace{1mm}\vec{{\bf R}}(t)= R \cos
288: (\Omega t)\hspace{1mm} \vec{{\bf z}} + R \sin (\Omega t)
289: \hspace{1mm} \vec{{\bf x}},
290: \end{equation}
291: and that
292: \begin{eqnarray}
293: \label{D} \vec{r}(t) &=& r \sin (\omega t) \hspace{1mm} \vec{{\bf
294: x}}^{\prime} - r \cos (\omega t)\hspace{1mm} \vec{{\bf y}}^{\prime}
295: \nonumber\\
296: &=& r \sin (\omega t) [ (\cos \Theta)\hspace{1mm} \vec{{\bf x}}-
297: (\sin \Theta) \hspace{1mm}\vec{{\bf z}}] - r
298: \cos (\omega t)\hspace{1mm} \vec{{\bf y}} \nonumber\\
299: %&=& r \sin (\omega t) [(\cos \Omega t) \hspace{1mm}\vec{{\bf x}}- (\sin
300: %\Omega t)\hspace{1mm} \vec{{\bf z}}] - r
301: %\cos (\omega t)\hspace{1mm} \vec{{\bf y}} \nonumber\\
302: &=& r \sin (\omega t) (\cos \Omega t)\hspace{1mm} \vec{{\bf x}}- r
303: \cos (\omega
304: t)\hspace{1mm} \vec{{\bf y}} \nonumber\\
305: && - r \sin (\omega t)(\sin \Omega t)\hspace{1mm} \vec{{\bf z}},
306: \end{eqnarray}
307: where Eq. (\ref{A}) is used. By combining Eqs. (\ref{B})-(\ref{D}),
308: we obtain
309: \begin{eqnarray}
310: \label{E} \vec{L}_a(t)
311: %&=& \vec{R}(t)+ \vec{r}(t) \nonumber\\
312: &=& [R \sin (\Omega t)+ r \sin (\omega t) (\cos \Omega t)]
313: \hspace{1mm} \vec{{\bf x}}\nonumber\\
314: && - r \cos (\omega t)\hspace{1mm} \vec{{\bf y}}\nonumber\\
315: && +[R \cos (\Omega t) - r \sin (\omega t)(\sin \Omega t)]
316: \hspace{1mm} \vec{{\bf z}}.
317: \end{eqnarray}
318: From Eq. (\ref{E}), we obtain
319: \begin{eqnarray}
320: \label{F} \frac{d^2\vec{L}_a(t)}{d t^2} &=& [-R \Omega^2 (\sin
321: \Omega t) - r \omega^2 \sin(\omega t) \cos(\Omega t)\nonumber\\
322: && - 2 r \omega \Omega \cos(\omega t)\sin (\Omega t) -r \Omega^2
323: \sin(\omega t) \cos(\Omega t)]
324: \hspace{1mm} \vec{{\bf x}}\nonumber\\
325: && + r \omega^2 \cos (\omega t) \hspace{1mm} \vec{{\bf y}}\nonumber\\
326: && +[-R \Omega^2 \cos(\Omega t)+ r\omega^2 \sin(\omega t)
327: \sin(\Omega t) \nonumber\\
328: && - 2 r \omega \Omega \cos(\omega t) \cos(\Omega t) \nonumber\\
329: && + r \Omega^2 \sin(\omega t) \sin(\Omega t) ] \hspace{1mm}
330: \vec{{\bf z}}.
331: \end{eqnarray}
332: What we consider is the situation when $t=0$. From Eq. (\ref{F}), we
333: get
334: \begin{eqnarray}
335: \label{G} \frac{d^2\vec{L}_a(0)}{d t^2} = r \omega^2 \hspace{1mm}
336: \vec{{\bf y}} +[-R \Omega^2 -2 r \omega \Omega ] \hspace{1mm}
337: \vec{{\bf z}}.
338: \end{eqnarray}
339: Thus the force $\vec{f}_a(0)$ on the point-mass $a$ when $t=0$ is
340: given by
341: \begin{eqnarray}
342: \label{H} \vec{f}_a(0) &=& m\frac{d^2\vec{L}_a(0)}{d t^2} \nonumber\\
343: &=& mr \omega^2 \hspace{1mm} \vec{{\bf y}} +m[-R \Omega^2 -2 r
344: \omega \Omega ] \hspace{1mm} \vec{{\bf z}}.
345: \end{eqnarray}
346: Let us give an interpretation of Eq. (\ref{H}). The first term, $mr
347: \omega^2 \hspace{1mm} \vec{{\bf y}}$, is due to the rotation
348: involved with angular speed $\omega$, usually. However, the second
349: term, $m[-R \Omega^2 -2 r \omega \Omega ] \hspace{1mm} \vec{{\bf
350: z}}$, is due to combination of both rotations, the one involved with
351: angular speed $\omega$ and the other one involved with angular speed
352: $\Omega$. It is due to the second term that the force $\vec{f}_a(0)$
353: along $z$ direction is non-zero. The fact that the force
354: $\vec{f}_a(0)$ along $z$ direction is non-zero can also be seen from
355: trajectory shown in Fig. 3: The trajectory is curved toward origin
356: $O$ as well as toward a point $(0,+\infty,0)$.
357:
358: Before we proceed more, let us give a short-cut explanation of
359: Barker type \cite{Bar60}.
360:
361: {\bf Short-cut explanation}: {\it Because of the non-zero force
362: $\vec{f}_a(0)$ along $z$ direction on the point-mass $a$, a reaction
363: $\vec{h}_a(0)= - \vec{f}_a(0)$ is exerted on the branch. However,
364: the other end of the stem is fixed at the origin $O$ and the frame
365: composed of the stem and four branches is a solid body. Thus, the
366: frame would turn such that the point-mass moves upward, if the
367: torque by the reaction $\vec{h}_a(0)$ were the only one exerted on
368: the frame.}
369:
370: Let us resume our calculations. Similarly, we can calculate net
371: forces on other point-masses when $t=0$:
372: \begin{eqnarray}
373: \label{I} \vec{f}_b(0) &=& m\frac{d^2\vec{L}_c(0)}{d t^2}
374: \nonumber\\
375: &=& m[-r \omega^2-r \Omega^2] \hspace{1mm} \vec{{\bf x}} +m[-R
376: \Omega^2] \hspace{1mm} \vec{{\bf z}},
377: \end{eqnarray}
378: \begin{eqnarray}
379: \label{J} \vec{f}_c(0) &=& m\frac{d^2\vec{L}_c(0)}{d t^2} \nonumber\\
380: &=& -mr \omega^2 \hspace{1mm} \vec{{\bf y}} +m[-R \Omega^2 + 2 r
381: \omega \Omega ] \hspace{1mm} \vec{{\bf z}},
382: \end{eqnarray}
383: and
384: \begin{eqnarray}
385: \label{K} \vec{f}_d(0) &=& m\frac{d^2\vec{L}_c(0)}{d t^2}
386: \nonumber\\
387: &=& m[r \omega^2+r \Omega^2] \hspace{1mm} \vec{{\bf x}} +m[-R
388: \Omega^2] \hspace{1mm} \vec{{\bf z}}.
389: \end{eqnarray}
390: Note that $ \vec{f}_a(0)$ and $ \vec{f}_b(0)$ are similar to $
391: \vec{f}_c(0)$ and $ \vec{f}_d(0)$, respectively. Reaction
392: $\vec{h}_i(0)$ on branch by each point-mass is negative of each
393: force $\vec{f}_i(0)$, where $i=a,b,c,d$,
394: \begin{eqnarray}
395: \label{L} \vec{h}_i(0)= -\vec{f}_i(0).
396: \end{eqnarray}
397: $\vec{h}_i(0)$ are what we were looking for, the forces as defined
398: in the Proposition-1. By the Proposition-1, force on the frame by
399: each point-mass, $F_i$ is given by
400: \begin{equation}
401: \label{L-2} \vec{F}_i= \vec{h}_i(0)+m \vec{g}.
402: \end{equation}
403:
404: Now we are prepared to calculate torque $\vec{\tau}_i$ at $t=0$, due
405: to force $\vec{F}_i$. First let us separately calculate sum of
406: torque $\vec{\tau}_i^g$ due to term $m \vec{g}$ of four
407: point-masses,
408: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
409: \begin{equation}
410: \label{L-3} \vec{\tau}_i^g= 4mgR\hspace{1mm} {\bf \vec{x}},
411: \end{equation}
412: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
413: where ${\bf \vec{x}}$ is the unit vector in $x$-direction. Let us
414: calculate remaining torque $\vec{\tau}_i^h$ due to term
415: $\vec{h}_i(0)$: By inspecting spatial arrangement of vectors in Eqs.
416: (\ref{I}), (\ref{K}), and (\ref{L}), we can see that
417: $\vec{\tau}_b^h$ and $\vec{\tau}_d^h$ involved with point-masses $b$
418: and $d$ cancel each other, $\vec{\tau}_b^h+\vec{\tau}_d^h =0$.
419: However, by inspecting spatial arrangement of vectors in Eqs.
420: (\ref{H}), (\ref{J}), and (\ref{L}), we can see that sum of torques
421: involved with point-masses $a$ and $c$ is non-zero, that is,
422: $\vec{\tau}_a^h + \vec{\tau}_c^h = -4m r^2 \omega \Omega {\bf
423: \vec{x}} $. Therefore,
424: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
425: \begin{equation}
426: \label{M} \vec{\tau}^h= \sum_i \vec{\tau}_i^h= - 4m r^2 \omega
427: \Omega {\bf \vec{x}}.
428: \end{equation}
429: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
430: By Eqs. (\ref{L-3}) and (\ref{M}), we get total torque,
431: \begin{equation}
432: \label{N} \vec{\tau}= (4mgR\hspace{1mm}- 4m r^2 \omega \Omega) {\bf
433: \vec{x}}.
434: \end{equation}
435: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
436: However, if a condition,
437: \begin{equation}
438: \label{O} gR= r^2 \omega \Omega
439: \end{equation}
440: is satisfied, the total torque is zero. This means that frame does
441: not make turning motion which will make the point-masses vertically.
442: Interestingly, {\it this is the condition for the precession of the
443: top found in text books \cite{Hal93}.}
444:
445: {\bf Explanation-5}: {\it The total torque on the frame is zero.
446: Thus the frame does not make turning motion that will make the
447: point-masses move vertically. }
448:
449: However, the other end of frame is fixed at the origin $O$ and the
450: frame is a solid body. Thus, if the torque $\vec{\tau}_i^g$ and
451: $\vec{\tau}_i^h$ were separately exerted on the frame, the frame
452: would turn such that the point-masses move downward and upward,
453: respectively: Therefore, it is interesting to give a Barker-type
454: explanation.
455:
456: {\bf Explanation-6}: {\it The downward force on the frame due to the
457: torque by gravity, $\vec{\tau}_i^g$, is balanced by the floating
458: (upward) force due to the torque by point-masses, $\vec{\tau}_i^h$.
459: Thus spinning top does not collapse.}
460:
461: \section{Discussions}
462:
463: \subsection{Other motions of a top}
464: As also discussed in Ref. \cite{Bar60}, it is interesting to see
465: that the explanations can be applied to other motions of the top,
466: nutation \cite{Fey63}.
467:
468: Contrary to our simple-minded notion, even a spinning top does fall
469: depending on situation. Let us consider a motion described in Ref.
470: \cite{Fey63}. " If we were to hold the axis absolutely fixed, so
471: that it cannot process in any manner (but the top is spinning) then
472: there is no torque acting, not even a torque from gravity, because
473: it is balanced by our fingers. But if we suddenly let go, then there
474: will instantaneously be a torque from gravity. Anyone in his right
475: mind would think that top top would fall, and that is what it starts
476: to do, as can be seen if the top is not spinning too fast. The gyro
477: actually does fall, as we would expect. ..." How can we explain the
478: `falling of spinning top with (temporarily) fixed axis'? (In Ref.
479: \cite{Fey63}, we can find discussions on this point, too. Although
480: the discussions appear to be similar to our explanations. But what
481: they say is that anyhow force must be applied to each point-mass in
482: the same direction as that of acceleration of the point-mass. This
483: is not the same as ours.) This is a quite perplexing problem.
484: However, we can find an explanation of the Explanation-6 type for
485: the problem.
486:
487: {\bf Explanation-7}: {\it Because the spinning axis is (temporarily)
488: fixed, as we can see, there is no floating force that can be
489: obtained by precession of the axis. Therefore, the top falls down
490: due to downward force by gravity.}
491:
492: However, the falling of spinning top with (temporarily) fixed axis
493: does not last so long. The top would begin to move horizontally such
494: that the top makes a cycloid, as described in Fig. 20-5 of Ref.
495: \cite{Fey63}. We can also give an explanation for this motion.
496:
497: {\bf Explanation-8}: {\it The falling makes the top 'precess' toward
498: $-y$ direction temporarily. However, as we have seen in section II,
499: a precession toward $+x$ direction induces a force toward $-y$
500: direction on the frame. By the same mechanism, the precession toward
501: $-y$ direction induce a force toward $+x$ direction on the frame.
502: %Here we assume isotropy, that is, symmetry in spatial directions.
503: Therefore, the falling top accelerates in $+x$ direction also.}
504:
505: Combining the Explanation-7 and Explanation-8, we can understand why
506: the top makes a cycloid at least qualitatively.
507:
508: \subsection{Eastman's explanation}
509: We briefly review the Eastman's model \cite{Eas75,Edw77} in Fig. 4,
510: before we discuss difficulties of the model.
511: \begin{figure}[htbp]
512: \includegraphics[width=8cm,height=8cm]{Fig-4.ps}
513: \caption[Fig1]{\label{fig2} The Eastman's model \cite{Eas75,Edw77}.
514: Assume that impulses, $\vec{F}$ and $-\vec{f}$, are applied during a
515: time duration $\Delta t$, to a spinning top. These impulses induces
516: forces along the $z$ axis, $h$ and $-h$, on the point-masses $a$ and
517: $c$, respectively. The forces, $h$ and $-h$, then change velocities
518: $v(a)$ and $v(c)$ of point-masses $a$ and $c$ to $v^{\prime}(a)$ and
519: $v^{\prime}(c)$, respectively. Here $v^{\prime}(a)=v(a)+(h/m) \Delta
520: t$ and $v^{\prime}(c)=v(c)- (h/m) \Delta t$ are satisfied. Here $m$
521: is mass of each point-mass. As a result, the spinning axis of the
522: top is rotated by an angle that $v(a)$ and $v^{\prime}(a)$ make.
523: That is, the axis of the top precesses in a direction that is
524: perpendicular to the impulses.}
525: \end{figure}
526: Assume that impulses, $\vec{F}$ and $-\vec{f}$, are applied during a
527: time duration $\Delta t$, to a spinning top. These impulses induces
528: forces along the $z$ axis, $h$ and $-h$, on the point-masses $a$ and
529: $c$, respectively. The forces, $h$ and $-h$, then change velocities
530: $v(a)$ and $v(c)$ of point-masses $a$ and $c$ to $v^{\prime}(a)$ and
531: $v^{\prime}(c)$, respectively. Here $v^{\prime}(a)=v(a)+(h/m) \Delta
532: t$ and $v^{\prime}(c)=v(c)- (h/m) \Delta t$ are satisfied. Here $m$
533: is mass of each point-mass. As a result, the spinning axis of the
534: top is rotated by an angle that $v(a)$ and $v^{\prime}(a)$ make.
535: That is, the axis of the top precesses in a direction that is
536: perpendicular to the impulses.
537:
538: Let us discuss difficulties of the Eastman's model.
539:
540: First, it is not clear how the impulses, $\vec{F}$ and $-\vec{f}$,
541: induces forces, $h$ and $-h$. It seems that what the authors of
542: Refs. \cite{Eas75,Edw77} had in mind is the following. The impulses
543: make point-masses $a$ and $c$ move slightly along $z$ direction,
544: first. These movements induces the forces $h$ and $-h$. (This is in
545: contrast with that the impulses do not make point-masses $b$ and $d$
546: move and thus no forces are induced on the point-masses $b$ and
547: $d$.) However, even this argument is not sufficiently explaining why
548: the movements can be identified with the forces $h$ and $-h$.
549: Furthermore, actual movements of the point-masses $a$ and $c$ are
550: not along the $z$ axis but along the $x$ axis in the case of usual
551: precession dealt with in section II.
552:
553: Second, actual trajectories of point-masses are not taken into
554: account in the Eastman's model. However, actual trajectories at the
555: initial time decides initial condition. Therefore, the Eastman's
556: model cannot explain the fact that the same spinning top either does
557: or does not fall down depending on initial condition, as seen in the
558: previous subsection. In other words, according to the Eastman's
559: model the direction of precession is unchanged as long as the
560: direction of impulses are the same. However, the impulses are
561: provided by the gravity, and thus the impulses do not depend on the
562: initial condition. This implies that the spinning top would never
563: fall down, which contradicts facts.
564:
565: \section{Conclusion}
566: We elaborated an intuitive explanation by Barker about why spinning
567: top does not collapse. We presented the explanation in conceptually
568: more rigorous form. We calculated full trajectories involved. We
569: discuss nutation of spinning top. We described how Eastman's
570: explanation is incomplete. We conclude.
571:
572: \acknowledgments I am grateful to Profs. Intaek Lim and Jongwon Park
573: for helpful discussions.
574:
575: %\endnote{S. Wiesner, Sigact News {\bf15}(1), 78 (1983)}
576: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
577:
578: \begin{references}
579: \bibitem[*]{email}
580: Email address: wyhwang@chonnam.ac.kr
581: \bibitem{Hal93} D. Halliday, R. Resnick, J. Walker, {\it Fundamentals
582: of Physics, 4th ed.}, (John Wiley and Sons, Inc., New
583: York, 1993).
584: \bibitem{Fey63} R. Feynman, M. Sands, and R. Leighton, {\it The
585: Feynman Lectures on Physics}, (Addison Wesley
586: Longman, 1963)
587: \bibitem{Tho95} S. T. Thornton, {\it Classical Dynamics of Particles
588: and Systems, 4th ed.}, (Harcourt Brace and Company, Florida,
589: 1995).
590: \bibitem{Bar60} E. M. Barker, Am. J. Phys. {\bf 28}, 808 (1960).
591: \bibitem{Eas75} P. C. Eastman, Am. J. Phys. {\bf 43}, 365 (1975).
592: \bibitem{Edw77} P. L. Edwards, Am. J. Phys. {\bf 45}, 1194 (1977).
593:
594: \end{references}
595: \end{document}
596:
597: % ****** End of file apssamp.tex ******
598: