1 00:00:05,090 --> 00:00:07,020 PROFESSOR: Here are the differential equations 2 00:00:07,020 --> 00:00:12,020 for the angular momentum of a tumbling box. 3 00:00:12,020 --> 00:00:21,420 Try throwing a book, or a box, or any rectilinear object whose 4 00:00:21,420 --> 00:00:25,760 three dimensions are all different, into the air 5 00:00:25,760 --> 00:00:29,030 with a twist, to make a tumble. 6 00:00:29,030 --> 00:00:35,110 You could go to rotate about its longest axis, 7 00:00:35,110 --> 00:00:38,140 or about its shortest axis. 8 00:00:38,140 --> 00:00:42,583 But you can't get to rotate about its middle axis. 9 00:00:45,520 --> 00:00:47,780 Let's examine that phenomena numerically. 10 00:00:50,600 --> 00:00:53,280 Here's the anonymous function defining 11 00:00:53,280 --> 00:00:59,535 those system of three first order differential equations. 12 00:01:02,100 --> 00:01:07,310 Now I'm going to start with an initial condition that's 13 00:01:07,310 --> 00:01:09,340 near the first critical point. 14 00:01:09,340 --> 00:01:12,890 1, 0, 0 is a critical point. 15 00:01:12,890 --> 00:01:18,340 And I'm going to take 0.2 times a random number, 16 00:01:18,340 --> 00:01:23,560 to sort of be near the critical point, and then normalize it, 17 00:01:23,560 --> 00:01:27,770 so that it has length 1. 18 00:01:27,770 --> 00:01:32,210 So the largest component is the first component. 19 00:01:32,210 --> 00:01:38,050 And the other two are small, but not too small. 20 00:01:38,050 --> 00:01:40,220 This is an easy problem numerically. 21 00:01:40,220 --> 00:01:42,450 There's no stiffness involved here. 22 00:01:42,450 --> 00:01:50,210 So I'm going to use ODE 23, integrate from 0 to 10, 23 00:01:50,210 --> 00:01:51,950 and here's the solution. 24 00:01:51,950 --> 00:01:56,980 The blue component is the first one, and it stays near 1. 25 00:01:56,980 --> 00:02:04,800 And the other two are periodic, rotating around the 0. 26 00:02:04,800 --> 00:02:08,449 Let's go back and take another starting condition. 27 00:02:08,449 --> 00:02:11,080 Here it is again. 28 00:02:11,080 --> 00:02:16,880 Now the other two components are really small. 29 00:02:16,880 --> 00:02:22,470 And when we integrate that, the blue component 30 00:02:22,470 --> 00:02:24,350 stays flat near 1. 31 00:02:24,350 --> 00:02:29,330 And the other two guys hardly move at all. 32 00:02:29,330 --> 00:02:33,550 Now I'm going to go to the third critical point, 0, 0, 1. 33 00:02:33,550 --> 00:02:34,570 Do the same thing. 34 00:02:34,570 --> 00:02:38,390 Take a random number near there. 35 00:02:38,390 --> 00:02:40,400 Use ODE 23. 36 00:02:40,400 --> 00:02:43,810 Now the yellow component stays near one. 37 00:02:43,810 --> 00:02:47,070 And the other two move periodically around 0. 38 00:02:51,160 --> 00:02:54,230 Run that again. 39 00:02:54,230 --> 00:02:57,170 The third component is near 1. 40 00:02:57,170 --> 00:03:01,530 The other two are not too big. 41 00:03:01,530 --> 00:03:04,650 And run the ODE 23. 42 00:03:04,650 --> 00:03:07,070 The other component stays near 1. 43 00:03:07,070 --> 00:03:10,480 And the other two rotate periodically around 0. 44 00:03:13,480 --> 00:03:17,850 Now we're going to go to the middle critical point. 45 00:03:17,850 --> 00:03:22,190 We're going to try and get the box to rotate around 46 00:03:22,190 --> 00:03:25,440 its middle axis. 47 00:03:25,440 --> 00:03:29,220 The second component is the one near 1. 48 00:03:29,220 --> 00:03:34,900 And now we see completely different behavior. 49 00:03:34,900 --> 00:03:42,802 This sienna component doesn't stay near 1. 50 00:03:42,802 --> 00:03:46,870 It goes down near -1, and comes back up. 51 00:03:46,870 --> 00:03:50,690 Let's integrate over a longer period, 52 00:03:50,690 --> 00:03:53,190 so we can see that behavior. 53 00:03:53,190 --> 00:03:55,570 So it's periodic. 54 00:03:55,570 --> 00:03:59,200 But it goes down to -1 and comes back to 1. 55 00:03:59,200 --> 00:04:08,260 And the other two move in large amplitudes around 0. 56 00:04:08,260 --> 00:04:16,365 So this is the instability of that middle critical value. 57 00:04:16,365 --> 00:04:17,115 Let's doing again. 58 00:04:20,320 --> 00:04:21,550 Same thing. 59 00:04:21,550 --> 00:04:23,810 1 down to -1, and back up. 60 00:04:23,810 --> 00:04:25,000 It's periodic. 61 00:04:25,000 --> 00:04:27,460 These solutions are all periodic. 62 00:04:27,460 --> 00:04:31,845 But that middle critical point is unstable. 63 00:04:34,780 --> 00:04:38,220 Now I want to view these in a different way, graphically. 64 00:04:41,930 --> 00:04:46,570 The differential equations have these three critical points. 65 00:04:46,570 --> 00:04:50,700 Any solutions to start exactly in these initial conditions 66 00:04:50,700 --> 00:04:53,470 stay there. 67 00:04:53,470 --> 00:04:57,830 But what happens if you start near these initial conditions? 68 00:04:57,830 --> 00:05:04,850 Well it turns out, that x and z are stable critical points. 69 00:05:04,850 --> 00:05:08,490 But y is an unstable critical point. 70 00:05:08,490 --> 00:05:13,130 If the angular momentum is near x or near z, 71 00:05:13,130 --> 00:05:15,430 it stays near there. 72 00:05:15,430 --> 00:05:21,030 But if it starts near y, it moves away quickly. 73 00:05:21,030 --> 00:05:28,630 You can think of x as the short axis, and z is the long axis. 74 00:05:28,630 --> 00:05:32,570 Rotation near the short axis is stable. 75 00:05:32,570 --> 00:05:36,060 And rotation near the long axis a stable. 76 00:05:36,060 --> 00:05:40,990 But rotation near the middle axis is unstable. 77 00:05:40,990 --> 00:05:46,480 We can see that in the following graphic. 78 00:05:46,480 --> 00:05:48,730 It turns out that if a solution starts 79 00:05:48,730 --> 00:05:52,630 with an initial condition that has norm 1, 80 00:05:52,630 --> 00:05:55,220 it stays with norm 1. 81 00:05:55,220 --> 00:05:59,330 So the solution lives on the unit sphere. 82 00:05:59,330 --> 00:06:03,610 Here's our unit share with our three critical points, x, y, 83 00:06:03,610 --> 00:06:04,690 and z. 84 00:06:04,690 --> 00:06:08,800 If this were the Earth, z will be the North Pole. 85 00:06:08,800 --> 00:06:14,690 Axis where the 0-th meridian crosses the equator. 86 00:06:14,690 --> 00:06:19,160 That's in the east Atlantic, a little off of West Africa. 87 00:06:19,160 --> 00:06:25,070 y would be where the 90th meridian crosses the equator. 88 00:06:25,070 --> 00:06:30,260 That's in the Indian Ocean, west of Sumatra. 89 00:06:30,260 --> 00:06:35,210 If we start with an initial condition near x, 90 00:06:35,210 --> 00:06:38,040 the solution orbits around x. 91 00:06:38,040 --> 00:06:44,100 That's stable rotation around the short axis. 92 00:06:44,100 --> 00:06:48,510 If we start with an initial condition near z, 93 00:06:48,510 --> 00:06:51,330 the solution orbits around z. 94 00:06:51,330 --> 00:06:56,150 That's stable rotation around the long axis. 95 00:06:56,150 --> 00:07:00,150 But if we start near y, the solution 96 00:07:00,150 --> 00:07:09,990 takes off, goes over to near -y, turns around, 97 00:07:09,990 --> 00:07:11,860 and comes back to y. 98 00:07:11,860 --> 00:07:19,200 Periodic, but goes clear around the globe. 99 00:07:19,200 --> 00:07:26,310 That turns out, that's a circle actually, an orbit around x. 100 00:07:26,310 --> 00:07:37,560 If we come up a little above y, we get an orbit around z. 101 00:07:37,560 --> 00:07:45,740 Go down a little below y, we get an orbit around -z. 102 00:07:45,740 --> 00:07:52,230 Go to the right of y, we get an orbit around -x. 103 00:07:52,230 --> 00:07:54,615 Let's zoom in a little bit. 104 00:08:02,600 --> 00:08:12,571 And we can see that y is a classic unstable critical 105 00:08:12,571 --> 00:08:13,070 point. 106 00:08:27,100 --> 00:08:31,010 Let's conclude by drawing a few orbits.