1 00:00:01,775 --> 00:00:04,871 China Airlines Flight 676 falls out of the 2 00:00:04,903 --> 00:00:06,839 sky, crashing into a residential 3 00:00:06,887 --> 00:00:10,487 neighborhood in Taipei. It was a horrific 4 00:00:10,551 --> 00:00:13,383 crash. Pieces of the wreckage fell onto 5 00:00:13,439 --> 00:00:16,983 houses. More than 200 people are dead, 6 00:00:17,119 --> 00:00:19,679 making this the worst aviation disaster in 7 00:00:19,727 --> 00:00:21,115 Taiwan's history. 8 00:00:23,135 --> 00:00:25,567 Okay, let's get to work. I want photos of 9 00:00:25,591 --> 00:00:29,099 everything amid the devastation. Recovery 10 00:00:29,147 --> 00:00:31,775 teams locate the crucial flight recorders. 11 00:00:32,475 --> 00:00:34,451 Hopefully the NTSP can extract some useful 12 00:00:34,483 --> 00:00:37,275 data from them. Crash experts from around 13 00:00:37,315 --> 00:00:39,575 the world assist in the investigation. 14 00:00:40,675 --> 00:00:42,787 I've got something for you. We're on it. 15 00:00:42,891 --> 00:00:45,563 But 11 seconds of the flight are almost 16 00:00:45,619 --> 00:00:47,811 impossible to explain. What the heck 17 00:00:47,843 --> 00:00:51,443 happened here in 11 seconds is 18 00:00:51,459 --> 00:00:53,687 like an eternity. That verges is 19 00:00:53,711 --> 00:00:54,635 unbelievable. 20 00:01:22,315 --> 00:01:26,211 At least the weather has improved for 21 00:01:26,243 --> 00:01:30,667 now. China Airlines Flight 676 22 00:01:30,771 --> 00:01:33,451 is cruising over the South China Sea as it 23 00:01:33,483 --> 00:01:35,735 nears Taipei City, Taiwan. 24 00:01:37,235 --> 00:01:40,211 The flight is getting the land at Taipei's 25 00:01:40,243 --> 00:01:42,235 Chiang Kai Sheikh airport, the primary 26 00:01:42,315 --> 00:01:45,379 gateway to Thai1 and is the busiest 27 00:01:45,427 --> 00:01:48,915 airport on the island. This evening, 28 00:01:48,995 --> 00:01:51,695 Captain Longlinin Kang is at the controls. 29 00:01:51,995 --> 00:01:54,435 An air force veteran, he's clocked over 30 00:01:54,475 --> 00:01:57,375 2000 hours on the AirB Bar 300. 31 00:01:57,955 --> 00:01:59,615 This flight is special. 32 00:02:01,715 --> 00:02:03,459 I hope you're enjoying the flight tonight. 33 00:02:03,587 --> 00:02:06,571 Can I give you some water? His wife and 34 00:02:06,643 --> 00:02:08,683 children are on board returning from a 35 00:02:08,699 --> 00:02:11,523 holiday in Bali. It's a special moment for 36 00:02:11,619 --> 00:02:14,141 the family. There is no one that you will 37 00:02:14,173 --> 00:02:16,105 have rather fly them than yourself 38 00:02:16,485 --> 00:02:20,037 ensuring of their safety. Tonight's flight 39 00:02:20,101 --> 00:02:23,221 left Bali at 3:27pm and is 40 00:02:23,253 --> 00:02:25,477 scheduled to touch down in Taipei just 41 00:02:25,541 --> 00:02:26,385 after eight. 42 00:02:30,205 --> 00:02:31,917 I'll talk to the cabin. We're seven 43 00:02:31,941 --> 00:02:34,109 minutes from descent Tonight Kang is 44 00:02:34,157 --> 00:02:37,509 paired with a rookie. Okay, thanks. He has 45 00:02:37,557 --> 00:02:40,355 far more experience on the A300 than his 46 00:02:40,395 --> 00:02:42,827 first officer D. Sheng Jiang. First 47 00:02:42,891 --> 00:02:45,899 officer's job is help pilot flying to 48 00:02:45,947 --> 00:02:49,331 handle all the non essential job all non 49 00:02:49,363 --> 00:02:52,595 essential dututy so pilot of flying can 50 00:02:52,635 --> 00:02:54,855 concentrate on flying the aircraft. 51 00:02:56,875 --> 00:02:59,515 Yes. Hello. Hi. We're about seven minutes 52 00:02:59,555 --> 00:03:02,267 until weart descent. The descent is the 53 00:03:02,291 --> 00:03:04,475 busiest, most critical phase of any 54 00:03:04,515 --> 00:03:07,845 flight. Weather is 16 degrees and cloudy. 55 00:03:08,185 --> 00:03:09,681 We won't need any wheelchairs tonight, 56 00:03:09,713 --> 00:03:12,005 right? No, not tonight. 57 00:03:17,985 --> 00:03:21,361 Look at the EC cam, the trimm tank. The 58 00:03:21,393 --> 00:03:23,201 captain notices an error message 59 00:03:23,313 --> 00:03:27,089 concerning the trim tank. The trim tank 60 00:03:27,177 --> 00:03:29,337 system moves fuel between tanks in the 61 00:03:29,361 --> 00:03:31,769 wings and tail keeping the plane balanced 62 00:03:31,817 --> 00:03:33,125 and fuel efficient. 63 00:03:35,695 --> 00:03:37,195 Let's see what we can do. 64 00:03:38,215 --> 00:03:47,535 The auto trim Seems 65 00:03:47,575 --> 00:03:49,551 like we don't need to do anything. Yeah, 66 00:03:49,583 --> 00:03:52,795 we do. See the trim tank system fault. 67 00:03:53,095 --> 00:03:55,263 The captain guides the first officer 68 00:03:55,319 --> 00:03:57,759 through the issue the trim tank system 69 00:03:57,847 --> 00:03:58,595 fault. 70 00:04:00,635 --> 00:04:03,603 Trim tank system in flight if no forward 71 00:04:03,659 --> 00:04:06,335 transfer not our case. 72 00:04:06,955 --> 00:04:10,775 It's not a critical problem. Okay, clear. 73 00:04:11,195 --> 00:04:13,563 So you see, that's it. That's a trim tank 74 00:04:13,619 --> 00:04:17,459 system for procedure. It's just a 75 00:04:17,547 --> 00:04:19,923 minor hiccup. It's easy to clear the 76 00:04:19,939 --> 00:04:20,615 message. 77 00:04:27,505 --> 00:04:29,897 The plane is now 25 minutes from the 78 00:04:29,921 --> 00:04:41,377 airport descent. 79 00:04:41,521 --> 00:04:43,577 The captain prompts the first officer to 80 00:04:43,601 --> 00:04:45,165 check in with controllers. 81 00:04:48,785 --> 00:04:51,441 Control Dynasty 676. 82 00:04:51,633 --> 00:04:55,717 Requesting Dynasty 676. 83 00:04:55,781 --> 00:04:57,517 Descend and maintain flight level two 84 00:04:57,581 --> 00:05:00,645 fifty. The tower issues a new altitude of 85 00:05:00,685 --> 00:05:03,973 25,000ft. Descending to flight level 86 00:05:04,029 --> 00:05:07,825 250. Heading 360 on our way. 87 00:05:17,965 --> 00:05:21,183 Checke get the crew also needs to 88 00:05:21,239 --> 00:05:22,927 confirm their gate information with 89 00:05:22,951 --> 00:05:24,879 airline operations on the ground. 90 00:05:25,047 --> 00:05:28,035 Operations Dynasty 676. 91 00:05:29,295 --> 00:05:31,475 Disregard Dynasty 676. 92 00:05:32,095 --> 00:05:34,831 The first officer makes a careless mistake 93 00:05:35,023 --> 00:05:35,955 this side. 94 00:05:39,015 --> 00:05:42,555 Unfortunately, E4 made a request on the 95 00:05:42,855 --> 00:05:46,141 ATC China. So not the 96 00:05:46,173 --> 00:05:48,717 correct frequency of the company reduce. 97 00:05:48,821 --> 00:05:51,333 It's quite embarrassing because basically 98 00:05:51,429 --> 00:05:53,989 you callus the operress with your callsine 99 00:05:54,117 --> 00:05:56,869 so every pilot in that area. 100 00:05:57,037 --> 00:05:58,985 No humid mistake. 101 00:06:01,445 --> 00:06:03,893 The plane is now 45 miles from the 102 00:06:03,909 --> 00:06:04,625 airport. 103 00:06:07,325 --> 00:06:09,505 And the weather is getting worse. 104 00:06:12,115 --> 00:06:13,575 The lightning is pretty strong. 105 00:06:18,155 --> 00:06:21,003 The power of that lightning. Lightning 106 00:06:21,099 --> 00:06:23,163 does not pose a direct physical threat to 107 00:06:23,179 --> 00:06:25,131 the aircraft. But it usually comes with 108 00:06:25,163 --> 00:06:27,163 severiet thunderstorms, strong winds. 109 00:06:27,259 --> 00:06:29,267 That's not the kind of conditions that you 110 00:06:29,291 --> 00:06:31,699 want to land at the airport. The captain 111 00:06:31,747 --> 00:06:33,507 wants to get his plane on the ground 112 00:06:33,611 --> 00:06:37,727 before the storm worsens. Flaps 23 113 00:06:37,751 --> 00:06:39,719 miles from the airport. The pilots extend 114 00:06:39,767 --> 00:06:41,835 the flaps as they reduce speed. 115 00:06:42,775 --> 00:06:47,591 Flaps 20 below 116 00:06:47,623 --> 00:06:49,719 them. The busy streets of Taipei's 117 00:06:49,767 --> 00:06:52,295 sprawling suburbs are shrouded in fog and 118 00:06:52,335 --> 00:06:52,915 mist. 119 00:06:55,175 --> 00:06:57,127 Feasibility is always a concern for 120 00:06:57,191 --> 00:06:59,287 pilots, especially when landing. If you 121 00:06:59,311 --> 00:07:02,015 can't see the Runway, you poste a danger 122 00:07:02,055 --> 00:07:04,713 to the safety of the aircraft. You could 123 00:07:04,769 --> 00:07:07,641 either miss the Runway altogether or even 124 00:07:07,673 --> 00:07:09,485 worse, a control flight into terrain. 125 00:07:11,985 --> 00:07:14,737 God, that tail was really strong. The 126 00:07:14,761 --> 00:07:16,937 captain pitches the nose down for the 127 00:07:16,961 --> 00:07:17,605 descent. 128 00:07:20,305 --> 00:07:24,249 Landing lights on. They're just 129 00:07:24,297 --> 00:07:27,337 moments from landing. Landing checklist 130 00:07:27,401 --> 00:07:30,089 complete. But the captain doesn't like the 131 00:07:30,137 --> 00:07:33,735 approach. Gor go around. 132 00:07:34,595 --> 00:07:37,371 Just 13300ft from the ground. The captain 133 00:07:37,403 --> 00:07:39,135 aborts the landing attempt. 134 00:07:42,835 --> 00:07:45,243 Pilots can go around for any reasons. 135 00:07:45,379 --> 00:07:49,251 Like a non stable approach or runong 136 00:07:49,283 --> 00:07:53,347 way is blocked or by atc'or 137 00:07:53,491 --> 00:07:56,791 pilot not satisfied for the approach. And 138 00:07:56,823 --> 00:07:59,271 we have go around in the airport. Maybe 139 00:07:59,383 --> 00:08:02,679 one go around per day. It's quite common 140 00:08:02,727 --> 00:08:05,855 operations. But this go around is going 141 00:08:05,935 --> 00:08:07,155 tragically wrong. 142 00:08:20,575 --> 00:08:23,575 Dynasty 6:76 confirm go around. 143 00:08:25,035 --> 00:08:27,475 An alarm warns the crew that the plane is 144 00:08:27,515 --> 00:08:28,535 losing lift. 145 00:08:33,035 --> 00:08:36,655 With no lift, airplane knowse 146 00:08:37,595 --> 00:08:40,963 will drop and then airplane will de 147 00:08:41,059 --> 00:08:43,135 and then you get in trouble fast. 148 00:08:53,075 --> 00:08:55,497 The aircraft is up front, is clearng left 149 00:08:55,521 --> 00:08:57,785 of the Runway anding towards a densely 150 00:08:57,825 --> 00:09:01,513 populated residential area. Dynasty 676, 151 00:09:01,569 --> 00:09:03,245 Taipei what is your status? 152 00:09:18,865 --> 00:09:21,413 All vehicles and A300 has just crashed on 153 00:09:21,429 --> 00:09:23,305 the roadway north of 05 left. 154 00:09:27,005 --> 00:09:29,437 The burning remains of China Airlines 155 00:09:29,501 --> 00:09:32,445 Flight 676 are scattered along a busy road 156 00:09:32,525 --> 00:09:34,625 near Chiang Kai Shek Airport. 157 00:09:37,885 --> 00:09:40,213 It was a horrific crash. There was a 158 00:09:40,229 --> 00:09:42,933 severiet fire from the jet fuel and it 159 00:09:42,949 --> 00:09:45,125 took more than 60 firefighters from across 160 00:09:45,165 --> 00:09:46,705 the city to put out the fire. 161 00:09:54,095 --> 00:09:56,343 How could a modern jetliner performing a 162 00:09:56,359 --> 00:09:59,199 go around fall out of the sky and crash 163 00:09:59,287 --> 00:10:01,115 into a residential neighborhood? 164 00:10:17,875 --> 00:10:18,935 Any survivors? 165 00:10:20,915 --> 00:10:24,403 All 182 passengers and 14 crew 166 00:10:24,459 --> 00:10:27,531 on board China Airlines Flight 676 are 167 00:10:27,563 --> 00:10:30,535 dead including the captain's family. 168 00:10:33,395 --> 00:10:35,451 Six more people have been killed on the 169 00:10:35,483 --> 00:10:36,055 ground. 170 00:10:38,415 --> 00:10:40,919 It fell into a residential neighborhood. 171 00:10:41,047 --> 00:10:44,183 Pieces of the wreckage fell onto houses 172 00:10:44,319 --> 00:10:47,687 into cars. It'taiwan's worst 173 00:10:47,751 --> 00:10:48,955 air disaster. 174 00:10:53,695 --> 00:10:56,063 It was a very devastating incident know 175 00:10:56,079 --> 00:10:59,555 with many innocent lives lost at instant. 176 00:11:00,495 --> 00:11:03,511 It's also China Airline'second crash in 177 00:11:03,543 --> 00:11:04,995 less than five years. 178 00:11:06,455 --> 00:11:09,671 In 1994, Flight 140 crashed in 179 00:11:09,703 --> 00:11:13,991 Nagoya, Japan killing 264 of the 271 180 00:11:14,023 --> 00:11:18,155 on board that aircraft was also an A300. 181 00:11:21,295 --> 00:11:23,743 The Taiwanese people are very concerned 182 00:11:23,799 --> 00:11:25,391 and frightened because they want to make 183 00:11:25,423 --> 00:11:27,831 sure that the government will ensure that 184 00:11:27,863 --> 00:11:30,575 the airline is safe to travel. Desperate 185 00:11:30,615 --> 00:11:32,693 to allay the public's fears, the 186 00:11:32,709 --> 00:11:34,797 government grounds the airline's entire 187 00:11:34,861 --> 00:11:37,637 fleet of A300s until the planes can be 188 00:11:37,661 --> 00:11:40,517 checked for faults. The decision throws 189 00:11:40,581 --> 00:11:42,785 Taiwan's air travel into chaos. 190 00:11:45,645 --> 00:11:48,261 Taiwan is an island, so air transport is 191 00:11:48,333 --> 00:11:50,125 incredibly important and with the aircraft 192 00:11:50,165 --> 00:11:52,397 grounded, a lot of flights were cancelled 193 00:11:52,501 --> 00:11:54,237 and many people's travel plans were out 194 00:11:54,261 --> 00:11:55,065 the window. 195 00:11:57,985 --> 00:12:00,097 The Taiwanese Civil Aeronautics 196 00:12:00,161 --> 00:12:02,577 Administration will lead the high profile 197 00:12:02,681 --> 00:12:05,105 investigation. The pressure to find 198 00:12:05,145 --> 00:12:06,725 answers is intense. 199 00:12:08,465 --> 00:12:10,125 Okay, let's get to work. 200 00:12:13,185 --> 00:12:16,033 I want photos of everything. We all want 201 00:12:16,049 --> 00:12:17,945 to know what happened quickly and get this 202 00:12:17,985 --> 00:12:19,889 solved. But the same token, you don't want 203 00:12:19,897 --> 00:12:22,121 to leave something unfound because you 204 00:12:22,153 --> 00:12:23,993 rush through it. And so even though there 205 00:12:24,009 --> 00:12:26,689 may be pressure from the government, you 206 00:12:26,777 --> 00:12:28,569 have to push back, say we'll do this as 207 00:12:28,617 --> 00:12:32,685 fast as we can. Amid the devastation, 208 00:12:33,065 --> 00:12:35,585 investigators focus first on recovering as 209 00:12:35,625 --> 00:12:37,985 much of the plane as they can. Some pieces 210 00:12:38,025 --> 00:12:39,617 may not be important, but the first thing 211 00:12:39,641 --> 00:12:41,929 is to do you have all the parts of the 212 00:12:41,937 --> 00:12:43,377 airplane? Do you have the wing tips, the 213 00:12:43,401 --> 00:12:46,225 tail tips, the nose? But when you have 214 00:12:46,385 --> 00:12:48,601 buildings and cars and everything else, 215 00:12:48,713 --> 00:12:52,053 it becomes a lot more difficult they also 216 00:12:52,109 --> 00:12:54,101 plot the position of every piece of 217 00:12:54,133 --> 00:12:57,461 debris. Soon a pattern emerges that paints 218 00:12:57,493 --> 00:12:59,869 a vivid picture of the way the plane came 219 00:12:59,917 --> 00:13:02,213 down. Whatever went wrong, it caused them 220 00:13:02,229 --> 00:13:04,145 to come in steep and flat. 221 00:13:07,645 --> 00:13:11,461 They pancake looking just at the impact. 222 00:13:11,573 --> 00:13:13,301 It gives you initial indication that this 223 00:13:13,333 --> 00:13:15,693 airplane, although it was coming down 224 00:13:15,749 --> 00:13:17,983 vertically very quickly, it didn't have a 225 00:13:17,999 --> 00:13:20,823 lot of forward velocity to it. Also, the 226 00:13:20,879 --> 00:13:22,519 pitch att added toe. The airplane wasn't 227 00:13:22,567 --> 00:13:24,255 severely pitched down or pitched up. It 228 00:13:24,755 --> 00:13:26,895 was a very kind of more of a flat impact, 229 00:13:27,015 --> 00:13:30,135 which is a little bit unusual. The high 230 00:13:30,175 --> 00:13:32,647 velocity impact destroyed most of the 231 00:13:32,671 --> 00:13:36,303 plane, but both engines survived. I see a 232 00:13:36,319 --> 00:13:39,463 distinct damage pattern. If the 233 00:13:39,479 --> 00:13:42,095 engines are running at impact, basically 234 00:13:42,175 --> 00:13:44,643 you'll bend or shear off the fan blades 235 00:13:44,699 --> 00:13:46,415 opposite to the direction of rotation. 236 00:13:47,275 --> 00:13:49,747 Also you'll see where they ingest a fair 237 00:13:49,771 --> 00:13:52,547 amount of debris, usually vegetation. At 238 00:13:52,571 --> 00:13:55,811 the same time engines 239 00:13:55,843 --> 00:13:59,443 are running when they hit, hard to know 240 00:13:59,459 --> 00:14:03,091 for sure if they're working properly. So 241 00:14:03,203 --> 00:14:06,051 there are numerous mechanical hints that 242 00:14:06,123 --> 00:14:08,027 investigators can get to determine if 243 00:14:08,051 --> 00:14:10,479 there's power on the engine. How much 244 00:14:10,527 --> 00:14:12,055 power? Well, that's a little bit 245 00:14:12,095 --> 00:14:14,167 different. Then you need to turn to flight 246 00:14:14,231 --> 00:14:16,967 recorders or some sort of recorded data to 247 00:14:16,991 --> 00:14:20,639 get at that. Investigators also recover 248 00:14:20,687 --> 00:14:23,175 the jack screw mechanism used to operate 249 00:14:23,215 --> 00:14:27,111 the flaps. If a pilot chooses 250 00:14:27,143 --> 00:14:29,351 the wrong flap settings, there can be two 251 00:14:29,383 --> 00:14:33,167 dangers. One is you select flaps 252 00:14:33,231 --> 00:14:35,231 at too high of a speed and that could 253 00:14:35,263 --> 00:14:37,267 structurally damage the flaps. The other 254 00:14:37,291 --> 00:14:39,875 danger is that you could be at a low speed 255 00:14:39,915 --> 00:14:41,819 and retract flaps. Now you don't have 256 00:14:41,867 --> 00:14:44,267 enough lift and the airplane is going to 257 00:14:44,291 --> 00:14:46,915 start coming out of the air. The nuts 258 00:14:46,955 --> 00:14:50,011 almost at the stop. So the flaps were 259 00:14:50,043 --> 00:14:53,335 partially extended at 20 flaps 20. 260 00:14:53,915 --> 00:14:56,507 There's no sign of structural failure and 261 00:14:56,531 --> 00:14:58,615 the flaps were in the correct position 262 00:14:59,155 --> 00:15:02,035 early on the investigation. The process is 263 00:15:02,075 --> 00:15:03,751 fairly simple. A shotgun approach. We're 264 00:15:03,783 --> 00:15:06,847 looking at everything. There's one theory 265 00:15:06,871 --> 00:15:09,039 in particular that investigators want to 266 00:15:09,087 --> 00:15:09,755 explore. 267 00:15:12,215 --> 00:15:13,743 Maybe the weather had something to do with 268 00:15:13,759 --> 00:15:17,015 this. We know from the weather forecast 269 00:15:17,095 --> 00:15:18,755 that the weather was going to get worse. 270 00:15:19,215 --> 00:15:21,415 There was fog moving in, there was 271 00:15:21,455 --> 00:15:23,751 lightning. And so you start looking at all 272 00:15:23,783 --> 00:15:27,311 those factors, the power of that 273 00:15:27,343 --> 00:15:30,523 lightning. Investigators turned to 274 00:15:30,539 --> 00:15:33,163 the tower controller for more insight into 275 00:15:33,179 --> 00:15:35,419 the weather around Chiang Kai Shek Airport 276 00:15:35,467 --> 00:15:38,815 at the time China Airline 676 crashed. 277 00:15:39,315 --> 00:15:40,979 So what kind of winds are we looking at? 278 00:15:41,107 --> 00:15:43,283 Did a dangerous condition called wind 279 00:15:43,339 --> 00:15:45,215 shear bring the plane down? 280 00:15:47,715 --> 00:15:51,003 God, that tail'really strong. Wind shear 281 00:15:51,059 --> 00:15:53,467 is caused by powerful winds that can formm 282 00:15:53,531 --> 00:15:55,829 close to the Runway during thunderstorms, 283 00:15:55,987 --> 00:15:59,001 creating forces so powerful they can slam 284 00:15:59,033 --> 00:16:06,441 a plane into the Ground visibility 285 00:16:06,513 --> 00:16:09,085 was above minimum. No wind shear. 286 00:16:10,025 --> 00:16:13,033 The weather was fine. The weather theory 287 00:16:13,089 --> 00:16:16,249 is quickly ruled out. We knew there was 288 00:16:16,297 --> 00:16:19,297 lightning approaching the airport. The 289 00:16:19,321 --> 00:16:21,281 winds, however, were fairly light. No 290 00:16:21,313 --> 00:16:23,739 severe turbulence, no severe downdraft. 291 00:16:23,857 --> 00:16:26,991 We didn't see any weather out there. That 292 00:16:27,023 --> 00:16:29,447 by itself should cause this airplane to 293 00:16:29,471 --> 00:16:32,023 be, you know, pushed out of the sky. Take 294 00:16:32,039 --> 00:16:35,335 a look at this. But controllers did notice 295 00:16:35,415 --> 00:16:36,951 something which is troubling about the 296 00:16:36,983 --> 00:16:37,555 flight. 297 00:16:39,775 --> 00:16:43,423 Interesting. Airlines usually 298 00:16:43,479 --> 00:16:45,671 descend to the Runway at a gradual 3 299 00:16:45,703 --> 00:16:49,223 degree angle. But Flight 676 had a 300 00:16:49,239 --> 00:16:51,535 much faster, steeper descent. 301 00:16:52,835 --> 00:16:54,915 You're right, they're way too high on 302 00:16:54,955 --> 00:16:58,435 approach. All the way in at a point 303 00:16:58,475 --> 00:17:00,459 it's about five miles from the Runway. 304 00:17:00,547 --> 00:17:03,147 The airplane is 1,000ft higher than it 305 00:17:03,171 --> 00:17:06,215 should be for that position. And so 306 00:17:06,795 --> 00:17:08,015 they've got a problem. 307 00:17:10,315 --> 00:17:13,963 Go over, go around. The finding explains 308 00:17:14,059 --> 00:17:16,485 why the captain called for a go around. 309 00:17:16,675 --> 00:17:18,833 But it still doesn't explain what caused 310 00:17:18,889 --> 00:17:22,321 the crash. A go around inter self is 311 00:17:22,353 --> 00:17:25,785 not a cause for concern at the crash site. 312 00:17:25,905 --> 00:17:28,225 The painstaking search of the debris pays 313 00:17:28,265 --> 00:17:31,289 off. Investigators find both of the black 314 00:17:31,337 --> 00:17:34,405 box flight recorders. Great job. 315 00:17:35,745 --> 00:17:37,425 The flight data recorders are essential 316 00:17:37,465 --> 00:17:39,289 for finding those little bits of 317 00:17:39,337 --> 00:17:41,857 information that you can't see in the 318 00:17:41,881 --> 00:17:45,107 wreckage itself. It gives us a lot more 319 00:17:45,171 --> 00:17:49,019 data to now put into our little basket of 320 00:17:49,067 --> 00:17:52,291 information. Investigators get more 321 00:17:52,323 --> 00:17:54,443 critical information with the discovery of 322 00:17:54,459 --> 00:17:57,355 the plane'ph FedEx or full authority 323 00:17:57,475 --> 00:18:00,891 digital engine controllers. The PHADEC is 324 00:18:00,923 --> 00:18:03,211 just like a flight data recorder, except 325 00:18:03,243 --> 00:18:06,495 it's specific only to the engine operation 326 00:18:06,795 --> 00:18:08,747 or if there were any engine anomalies that 327 00:18:08,771 --> 00:18:10,365 were occurring during the flight. 328 00:18:11,425 --> 00:18:13,401 Hopefully the NTSP can extract some useful 329 00:18:13,433 --> 00:18:15,817 data from them. Recovering the data from 330 00:18:15,841 --> 00:18:18,281 the US made controllers requires special 331 00:18:18,353 --> 00:18:20,601 expertise. So investigators ask the 332 00:18:20,633 --> 00:18:22,769 National Transportation Safety Board to 333 00:18:22,817 --> 00:18:25,025 oversee the download. They were slightly 334 00:18:25,065 --> 00:18:28,105 damaged, but they weren't burned. And that 335 00:18:28,145 --> 00:18:30,545 gave investigators hope. While the PH 336 00:18:31,045 --> 00:18:32,697 FedEx are transported to Connecticut in 337 00:18:32,721 --> 00:18:34,045 the United States. 338 00:18:37,365 --> 00:18:39,885 Welcome to Canberra. The flight recorders 339 00:18:39,925 --> 00:18:41,829 are sent to the Bureau of Air Safety 340 00:18:41,917 --> 00:18:45,085 Investigation or BASI in Canberra, 341 00:18:45,205 --> 00:18:48,101 Australia. Well, we've had expertise and 342 00:18:48,133 --> 00:18:51,141 facilities for the replay and analysis of 343 00:18:51,173 --> 00:18:54,385 flight recorders going back to the 1970s. 344 00:18:56,045 --> 00:18:59,037 Let's find you some answers. We've had 345 00:18:59,101 --> 00:19:02,165 close links in the Asia Pacific region for 346 00:19:02,245 --> 00:19:05,265 many decades. So it's natural that our 347 00:19:05,805 --> 00:19:07,637 expertise would be made available to the 348 00:19:07,661 --> 00:19:11,125 Taiwanese in these international acts 349 00:19:11,205 --> 00:19:13,277 investigations. Each country brings a 350 00:19:13,301 --> 00:19:15,469 different area of expertise and the 351 00:19:15,517 --> 00:19:17,845 purpose behind this is global aviation 352 00:19:17,885 --> 00:19:20,101 safety. Because if an accident happened in 353 00:19:20,133 --> 00:19:21,597 one place, well, it could have happened 354 00:19:21,661 --> 00:19:25,145 easily in your own country in Connecticut. 355 00:19:26,945 --> 00:19:30,409 Hold it there. NTSB systems Analyst 356 00:19:30,457 --> 00:19:33,001 Jeff Gazeti begins recovering the Data 357 00:19:33,073 --> 00:19:36,313 from the Fedadeex of Flight 676 with helpd 358 00:19:36,369 --> 00:19:37,725 from the manufacturer. 359 00:19:39,385 --> 00:19:42,177 The manufacturer is very much needed. 360 00:19:42,201 --> 00:19:44,041 It's their laboratory. However, the 361 00:19:44,073 --> 00:19:46,001 procedure is dictated by the 362 00:19:46,073 --> 00:19:48,401 investigators. That box was not opened 363 00:19:48,473 --> 00:19:51,225 until investigators like myself showed up. 364 00:19:51,345 --> 00:19:54,459 And in front of everybody, we unlocked, 365 00:19:54,547 --> 00:19:57,475 unsealed the boxes and began to extract 366 00:19:57,595 --> 00:19:59,215 the components that we needed. 367 00:20:03,315 --> 00:20:04,975 Well, here goes nothing. 368 00:20:08,075 --> 00:20:10,171 There's no guarantee that the data 369 00:20:10,243 --> 00:20:13,843 survived. These chips were not designed to 370 00:20:13,859 --> 00:20:16,339 be crash worhy. They were designed so that 371 00:20:16,387 --> 00:20:18,587 an airline can download information from 372 00:20:18,611 --> 00:20:21,263 the PHC and conduct maintenance on the 373 00:20:21,279 --> 00:20:22,555 engines as needed. 374 00:20:24,895 --> 00:20:28,167 Today they get lucky. 375 00:20:28,351 --> 00:20:30,871 So I was very relieved to see that the 376 00:20:30,903 --> 00:20:33,111 memory did survive on the chips of the PHA 377 00:20:33,143 --> 00:20:34,935 deck. So now we had something to work 378 00:20:34,975 --> 00:20:35,555 with. 379 00:20:40,535 --> 00:20:43,535 Okay, let's see what we've got. Were the 380 00:20:43,575 --> 00:20:46,679 engines functioning properly? The phadec 381 00:20:46,727 --> 00:20:49,367 memory will not only point out the anomaly 382 00:20:49,431 --> 00:20:51,863 of an engine, but exactly when it happened 383 00:20:51,959 --> 00:20:54,199 and at what altitude. And that data is 384 00:20:54,247 --> 00:20:55,875 crucial for investigators. 385 00:20:57,295 --> 00:20:59,487 The data confirms the investigator's 386 00:20:59,511 --> 00:21:01,831 observation that the engines were working 387 00:21:01,903 --> 00:21:04,615 when the plane hit the ground. But 388 00:21:04,695 --> 00:21:07,031 Something inexplicable happened 20 seconds 389 00:21:07,143 --> 00:21:08,075 before that. 390 00:21:14,315 --> 00:21:15,855 What the heck happened here? 391 00:21:17,635 --> 00:21:19,843 Engines were operating normally during the 392 00:21:19,899 --> 00:21:22,811 entire accident flight, except during a 393 00:21:22,843 --> 00:21:24,755 certain portion of the go around in which 394 00:21:24,795 --> 00:21:27,379 both engines surged at the same time, 395 00:21:27,507 --> 00:21:29,615 which is anomalous. Shouldn't happen. 396 00:21:30,355 --> 00:21:33,403 Aircraft engines ingest huge quantities of 397 00:21:33,419 --> 00:21:36,491 air and use it to spin a turbine to create 398 00:21:36,563 --> 00:21:39,863 thr ###ust if that air is disrupted, the 399 00:21:39,879 --> 00:21:42,167 engine can't produce enough thrust and 400 00:21:42,191 --> 00:21:45,955 will backfire or surge. It's kind of like 401 00:21:46,735 --> 00:21:49,415 walking into a room that suddenly has no 402 00:21:49,455 --> 00:21:52,559 air and you're gasping and you're 403 00:21:52,607 --> 00:21:53,395 surging. 404 00:22:00,255 --> 00:22:03,075 Campbell, I've got something for you. 405 00:22:03,465 --> 00:22:06,337 The NTSB updates the Australians about 406 00:22:06,361 --> 00:22:09,245 their findings. Engine search. 407 00:22:10,345 --> 00:22:12,605 Okay, we're on it. 408 00:22:14,825 --> 00:22:19,305 But what caused the surge could 409 00:22:19,345 --> 00:22:22,361 be ingesting birds into the engine. 410 00:22:22,473 --> 00:22:25,161 It could be a mechanical malfunction 411 00:22:25,193 --> 00:22:27,297 that's allowing too much fuel or too 412 00:22:27,321 --> 00:22:30,389 little air into the eng and we just didn't 413 00:22:30,437 --> 00:22:31,025 know. 414 00:22:34,525 --> 00:22:37,965 Let's take a look at the Australian Bureau 415 00:22:38,005 --> 00:22:40,909 of Air Safety investigation. Neil Campbell 416 00:22:40,997 --> 00:22:43,133 reviews a simulation of the final seconds 417 00:22:43,189 --> 00:22:46,261 of Flight 676 based on the information 418 00:22:46,373 --> 00:22:48,705 captured by the flight data recorder. 419 00:22:50,205 --> 00:22:52,389 In this case, a flight Data recorder 420 00:22:52,477 --> 00:22:56,021 recorded 115 separate parameters. It tells 421 00:22:56,053 --> 00:22:59,025 us things about the aircraft performance, 422 00:22:59,325 --> 00:23:01,597 its configuration, whether the flaps are 423 00:23:01,621 --> 00:23:04,093 up or down, the landing gears up or down. 424 00:23:04,269 --> 00:23:06,825 It tells us what the crew inputs were, 425 00:23:07,205 --> 00:23:09,733 the aircraft responses to those inputs, 426 00:23:09,909 --> 00:23:12,925 and also the status of various aircraft 427 00:23:13,005 --> 00:23:15,345 systems. Nearly at the Airport. 428 00:23:16,165 --> 00:23:19,093 Campbell scrutinizes the data carefully, 429 00:23:19,269 --> 00:23:21,293 looking for anything that might explain 430 00:23:21,389 --> 00:23:24,261 why the engines of a modern A300 surged 431 00:23:24,333 --> 00:23:25,735 and lost trust. 432 00:23:30,555 --> 00:23:32,135 He comes to go around. 433 00:23:33,955 --> 00:23:35,375 Go over. Go around. 434 00:23:41,435 --> 00:23:44,175 Hold on. What have we here? 435 00:23:45,035 --> 00:23:47,975 15 seconds after the go around begins, 436 00:23:48,275 --> 00:23:49,675 investigators noticed something 437 00:23:49,715 --> 00:23:51,715 surprising. Stop. 438 00:23:53,335 --> 00:23:56,551 We saw the aircraft pitched up to a pitch 439 00:23:56,583 --> 00:23:59,591 attitude of over 40 degrees. This is a 440 00:23:59,623 --> 00:24:02,679 very large value, very unusual. Normally 441 00:24:02,727 --> 00:24:05,007 during a go around, we'd expect to see 15 442 00:24:05,071 --> 00:24:07,759 degrees pitch attitude. No wonder the 443 00:24:07,767 --> 00:24:11,167 engine surged with the nose of the plane. 444 00:24:11,231 --> 00:24:14,143 Pitched up so steeply. Air can't flow 445 00:24:14,199 --> 00:24:17,403 through the engine to create thrust. The 446 00:24:17,419 --> 00:24:19,403 airplane was at a very high attitude. 447 00:24:19,459 --> 00:24:21,491 This prevented the engine from taking in 448 00:24:21,563 --> 00:24:25,515 smooth, continuous airflow. Ah, now the 449 00:24:25,555 --> 00:24:27,643 steep angle has another disastrous 450 00:24:27,699 --> 00:24:31,035 consequence. If an aircraft pitches up too 451 00:24:31,075 --> 00:24:33,587 steeply, it disrupts air FLW over the 452 00:24:33,611 --> 00:24:35,459 wings, preventing them from providing 453 00:24:35,507 --> 00:24:39,027 lift. Pitched up so high, they stalled 454 00:24:39,051 --> 00:24:39,775 the plane. 455 00:24:43,055 --> 00:24:46,031 And when an airliner turns straight up at 456 00:24:46,063 --> 00:24:48,951 such a high angle, the airspeed bleeds off 457 00:24:49,063 --> 00:24:51,303 significantly and gets you closer to an 458 00:24:51,319 --> 00:24:54,111 aerodynamic stall. With the loss of 459 00:24:54,143 --> 00:24:56,527 airflow into the engines and lift from the 460 00:24:56,551 --> 00:24:59,951 wings, Flight 676 begins to drop from 461 00:24:59,983 --> 00:25:03,471 the sky. You're not flying. You're almost 462 00:25:03,503 --> 00:25:05,367 becoming a falling object at that point. 463 00:25:05,511 --> 00:25:06,635 See that again. 464 00:25:13,995 --> 00:25:17,051 He's trying to get the nose down. The data 465 00:25:17,123 --> 00:25:19,067 shows the captain did what he could to 466 00:25:19,091 --> 00:25:22,531 regain control. Pilots are 467 00:25:22,563 --> 00:25:24,275 trying to get the nose down and get the 468 00:25:24,315 --> 00:25:26,571 speed back up. Airflow over the wings 469 00:25:26,643 --> 00:25:29,827 again. This take time and the situation 470 00:25:29,931 --> 00:25:33,155 where seconds are a big deal, but it's too 471 00:25:33,195 --> 00:25:33,775 late. 472 00:25:42,285 --> 00:25:43,665 They ran out of time. 473 00:25:48,765 --> 00:25:51,397 Unfortunately, the altitude was too low 474 00:25:51,541 --> 00:25:53,093 and they were not able to recover the 475 00:25:53,109 --> 00:25:55,597 aircraft in time. If they'd had another 476 00:25:55,621 --> 00:25:58,309 few thousand feet, there's a good 477 00:25:58,437 --> 00:26:00,077 possibility that the crew would have been 478 00:26:00,101 --> 00:26:02,305 able to recover the aircraft successfully. 479 00:26:03,375 --> 00:26:06,735 It's a tragic discovery, but why did 480 00:26:06,775 --> 00:26:09,471 Flight 676 pitch up so steeply in the 481 00:26:09,503 --> 00:26:12,391 first place? Campbell combs through data 482 00:26:12,463 --> 00:26:15,311 from earlier in the flight. Throughout the 483 00:26:15,423 --> 00:26:17,415 investigative process, everything gets 484 00:26:17,455 --> 00:26:21,447 looked at and relooked at. Look at 485 00:26:21,511 --> 00:26:24,555 this. That's weird. 486 00:26:25,575 --> 00:26:27,679 Something strange took place on board 487 00:26:27,727 --> 00:26:31,155 flight 676 just seconds before the stall. 488 00:26:35,035 --> 00:26:36,935 Bring up pitch and elevator, please. 489 00:26:41,595 --> 00:26:43,371 You can see the pilots making control 490 00:26:43,403 --> 00:26:46,215 inputs here and here. 491 00:26:48,875 --> 00:26:53,051 But during these 11 seconds, the plane 492 00:26:53,123 --> 00:26:55,955 is pitching up. Investigators determined 493 00:26:55,995 --> 00:26:58,139 that once the pilots initiated the go 494 00:26:58,187 --> 00:27:00,841 around, no inputs on the control column 495 00:27:00,873 --> 00:27:03,081 were made as the plane continued to pitch 496 00:27:03,113 --> 00:27:05,633 up dangerously. The aircraft will 497 00:27:05,729 --> 00:27:08,049 naturally want to pitch up as the engine 498 00:27:08,097 --> 00:27:10,841 thrust increases. So I would expect the 499 00:27:10,873 --> 00:27:14,325 crew to counteract that pitching up by 500 00:27:14,665 --> 00:27:17,377 pushing down on the Controls. Looking at 501 00:27:17,401 --> 00:27:19,177 the flight recorder data and seeing that 502 00:27:19,201 --> 00:27:22,113 there was no pitch controls, no changes in 503 00:27:22,129 --> 00:27:25,277 those controls, it did make me wonder what 504 00:27:25,301 --> 00:27:28,333 the crew were thinking. So no 505 00:27:28,349 --> 00:27:31,949 one's flying the plane. It's very unusual 506 00:27:31,997 --> 00:27:33,917 that there would be no control inputs by 507 00:27:33,941 --> 00:27:36,133 the pilots during this period. Very 508 00:27:36,189 --> 00:27:38,893 surprising in 11 seconds in this situation 509 00:27:38,949 --> 00:27:42,581 is like an eternity in fact, to just sit 510 00:27:42,613 --> 00:27:44,925 there and do nothing while this airplane 511 00:27:45,005 --> 00:27:47,501 is pitching up, losing speed, having 512 00:27:47,573 --> 00:27:51,553 issues. That vertge is unbelievable. 513 00:27:51,649 --> 00:27:53,365 Do you think the controls are jammeded 514 00:27:54,025 --> 00:27:57,165 just for these 11 seconds? No way. 515 00:27:58,105 --> 00:28:00,305 There's three independent flight control 516 00:28:00,385 --> 00:28:04,605 systems for pitch, roll and 517 00:28:05,345 --> 00:28:07,945 yaw. So it's the elevators, ailerons and 518 00:28:07,985 --> 00:28:12,033 rudder. All those systems showed that 519 00:28:12,049 --> 00:28:14,153 there was no activity for that 11 second 520 00:28:14,249 --> 00:28:16,731 period. But prior to the go around, 521 00:28:16,803 --> 00:28:19,139 during the early part of the flight and 522 00:28:19,187 --> 00:28:22,027 during the recovery from the stall, there 523 00:28:22,051 --> 00:28:25,003 were inputs made on all those systems. 524 00:28:25,179 --> 00:28:27,815 And the aircraft response was as expected. 525 00:28:29,875 --> 00:28:33,187 If the controls were working, why didn't 526 00:28:33,211 --> 00:28:37,035 the crew do anything? This was a stunning 527 00:28:37,115 --> 00:28:39,675 finding to us. Whenever you're faced with 528 00:28:39,715 --> 00:28:42,659 evidence of improper actions by a flight 529 00:28:42,707 --> 00:28:45,381 crew, you have to ask why? Was it because 530 00:28:45,413 --> 00:28:47,581 they were perhaps intoxicated? Was it 531 00:28:47,613 --> 00:28:49,765 perhaps because they were incapacitated? 532 00:28:49,925 --> 00:28:51,581 You really have to get into the human 533 00:28:51,613 --> 00:28:53,065 factor element of this. 534 00:28:58,085 --> 00:28:59,461 We need to know everything about these 535 00:28:59,493 --> 00:29:01,621 guys. Background, training, favorite 536 00:29:01,653 --> 00:29:03,145 foods, the works. 537 00:29:16,225 --> 00:29:20,185 Thank you. Investigators searched 538 00:29:20,225 --> 00:29:22,993 to explain why two trained pilots allowed 539 00:29:23,049 --> 00:29:26,185 Flight 676 to pitch up so steeply that it 540 00:29:26,225 --> 00:29:28,085 lost lift and stalled. 541 00:29:30,025 --> 00:29:32,849 When you see a situation where the pilot 542 00:29:32,897 --> 00:29:35,851 actions frankly don't make any sense, you 543 00:29:35,883 --> 00:29:37,963 wonder is there something medically wrong 544 00:29:38,019 --> 00:29:40,763 with the pilots? Alcohol, drugs, you know, 545 00:29:40,819 --> 00:29:43,611 anything else going on? They review the 546 00:29:43,643 --> 00:29:46,215 pilot's post mortem toxicology tests. 547 00:29:50,875 --> 00:29:52,575 No san. They'incapacitated. 548 00:29:55,075 --> 00:29:57,275 You have to be very methodical. You can't 549 00:29:57,315 --> 00:30:01,163 allow any biases or opinions 550 00:30:01,299 --> 00:30:04,683 to get in the way of you collecting facts. 551 00:30:04,779 --> 00:30:06,571 Because at this point, that's all you're 552 00:30:06,603 --> 00:30:07,175 doing. 553 00:30:11,395 --> 00:30:13,131 No red flags in the Creu's flying records 554 00:30:13,163 --> 00:30:13,775 either. 555 00:30:18,675 --> 00:30:21,291 Nothing stood out that would explain what 556 00:30:21,323 --> 00:30:24,535 happened here. Thank you. 557 00:30:26,875 --> 00:30:29,075 Investigators pin their hopes of finding 558 00:30:29,115 --> 00:30:30,975 answers on the cockpit voice. 559 00:30:34,275 --> 00:30:36,095 Okay, let's hear it. 560 00:30:38,635 --> 00:30:40,619 The cockpit voice recorder gives you those 561 00:30:40,667 --> 00:30:42,475 little pieces of information, otherwise 562 00:30:42,515 --> 00:30:45,067 you wouldn't have. It's critical to 563 00:30:45,171 --> 00:30:47,859 understand timing of the crew. How are 564 00:30:47,867 --> 00:30:50,123 they working together? What's the tone of 565 00:30:50,139 --> 00:30:53,619 the cockhead? Whether's 16 degrees and 566 00:30:53,667 --> 00:30:55,907 cloudy. We won't need any wheelchairs 567 00:30:55,931 --> 00:30:58,363 tonight, right? The team starts the 568 00:30:58,419 --> 00:31:00,587 recording 30 minutes before the end of the 569 00:31:00,611 --> 00:31:01,175 flight. 570 00:31:04,555 --> 00:31:06,975 Look at the. The TR Trump tank. 571 00:31:08,955 --> 00:31:11,251 They listen intently to everything. The 572 00:31:11,283 --> 00:31:15,075 Pilots said trim 573 00:31:15,115 --> 00:31:17,683 tank system in flight. If no forward 574 00:31:17,739 --> 00:31:20,415 transfer. Not our case. 575 00:31:23,535 --> 00:31:26,391 Okay, clear. So you see, that's it. 576 00:31:26,543 --> 00:31:28,675 That's a trim tank system for procedure. 577 00:31:30,255 --> 00:31:31,743 This captain is spending an awful lot of 578 00:31:31,759 --> 00:31:33,195 time helping his first officer. 579 00:31:35,415 --> 00:31:37,135 Normally you have a division of labor. 580 00:31:37,255 --> 00:31:39,591 However, this captain is doing almost all 581 00:31:39,623 --> 00:31:41,927 the flying and monitoring himself. He's 582 00:31:41,951 --> 00:31:44,695 not engaging the first officer and really 583 00:31:44,775 --> 00:31:48,573 helping with things. Operations Dynasty 584 00:31:48,629 --> 00:31:49,865 676. 585 00:31:52,485 --> 00:31:55,941 Disregard Dynasty 676. When the first 586 00:31:56,013 --> 00:31:58,205 officer uses the wrong channel to call 587 00:31:58,245 --> 00:32:01,045 airline operations on the ground this 588 00:32:01,085 --> 00:32:03,533 side, the captain intervenes by making 589 00:32:03,589 --> 00:32:05,997 radio calls himself, which is contrary to 590 00:32:06,021 --> 00:32:08,573 procedure. I think in the captain's mind, 591 00:32:08,749 --> 00:32:11,085 this guy's incompetent and I can't trust 592 00:32:11,125 --> 00:32:13,453 him to do a simple task. And so the 593 00:32:13,469 --> 00:32:15,101 captain, what he do, he started doing more 594 00:32:15,133 --> 00:32:17,753 on his own. Don't write that down. We're 595 00:32:17,769 --> 00:32:20,969 descending. Okay. The captain begins 596 00:32:21,017 --> 00:32:23,081 criticizing the first officer's actions. 597 00:32:23,193 --> 00:32:26,665 Okay. The captain 598 00:32:26,705 --> 00:32:28,401 felt that the first officer was not 599 00:32:28,473 --> 00:32:30,969 performing his task up to his standards, 600 00:32:31,137 --> 00:32:33,265 and it might be easier and faster to do 601 00:32:33,305 --> 00:32:34,845 everything himself. 602 00:32:36,465 --> 00:32:39,657 What are you doing? I'm keying in the 603 00:32:39,681 --> 00:32:41,685 data. Don't waste too much time on that. 604 00:32:41,785 --> 00:32:44,429 What's the Q and H? It's a risky 605 00:32:44,477 --> 00:32:47,685 situation. It sets up a problem where this 606 00:32:47,725 --> 00:32:50,821 first officer may be almost afraid to take 607 00:32:50,893 --> 00:32:52,837 action without the captain's permission. 608 00:32:53,021 --> 00:32:55,853 Statistics show that over 80% of aircraft 609 00:32:55,909 --> 00:32:58,613 accidents have human performance as a 610 00:32:58,669 --> 00:33:01,381 contributing factor. So it's very, very 611 00:33:01,453 --> 00:33:03,637 important to work together effectively as 612 00:33:03,661 --> 00:33:07,045 a team, to have good communication and 613 00:33:07,085 --> 00:33:10,191 to to make the best decisions they can. 614 00:33:10,383 --> 00:33:12,275 Even under stressful situations, 615 00:33:15,855 --> 00:33:17,991 you see that the lightning is pretty 616 00:33:18,023 --> 00:33:20,159 strong. The worsening weather puts even 617 00:33:20,207 --> 00:33:22,847 more pressure on the busy captain. The 618 00:33:22,871 --> 00:33:26,215 part of that lightning. Bad teamwork, 619 00:33:26,335 --> 00:33:29,075 task overload, and now weather. 620 00:33:31,975 --> 00:33:34,607 Cockpit was a stressful environment. The 621 00:33:34,631 --> 00:33:36,479 captain in particular was worried about 622 00:33:36,527 --> 00:33:40,679 the deteriorating weather at Taip AI and 623 00:33:40,727 --> 00:33:43,195 whether he might need to divert or not. 624 00:33:46,055 --> 00:33:48,511 It's what investigators hear next that 625 00:33:48,543 --> 00:33:50,275 gives them a vital clue. 626 00:33:53,455 --> 00:33:55,715 Hold on. You hear that again? 627 00:34:08,435 --> 00:34:11,415 Hu. That's the autopilot switching off. 628 00:34:13,155 --> 00:34:16,299 There's a distinct sound or an oral 629 00:34:16,347 --> 00:34:18,715 warning that occurs when the autopilot 630 00:34:18,755 --> 00:34:20,659 disengages. And it's called a cavalry 631 00:34:20,707 --> 00:34:21,335 charge. 632 00:34:25,575 --> 00:34:27,711 Switching the autopilot off is a routine 633 00:34:27,783 --> 00:34:32,115 part of the flight. Landing lights on. 634 00:34:34,015 --> 00:34:35,823 But at this stage of the flight, it was 635 00:34:35,919 --> 00:34:39,047 unusual to hear it and very unusual that 636 00:34:39,071 --> 00:34:42,471 there was no comment by the crew. The 637 00:34:42,503 --> 00:34:45,151 pilots don't discuss why the autopilot has 638 00:34:45,183 --> 00:34:48,671 disengaged. Come on, guys, talk to each 639 00:34:48,703 --> 00:34:52,071 other. You would expect the captain 640 00:34:52,103 --> 00:34:54,815 to say, I'm taking over. I have control. 641 00:34:54,935 --> 00:34:56,815 Or at least some comment between the two 642 00:34:56,855 --> 00:34:58,463 crew about the fact the autopilots 643 00:34:58,519 --> 00:35:01,103 disengaged, but there was no comment at 644 00:35:01,119 --> 00:35:03,635 all. Landing checkl was complete. 645 00:35:04,775 --> 00:35:07,431 In my opinion after the criticism by the 646 00:35:07,463 --> 00:35:11,199 captain to the first officer. He was just 647 00:35:11,247 --> 00:35:14,079 going to respond to any direct request 648 00:35:14,127 --> 00:35:15,115 from the captain. 649 00:35:17,815 --> 00:35:21,639 Goa go around. Six seconds after 650 00:35:21,727 --> 00:35:23,399 investigators hear the autopilot 651 00:35:23,447 --> 00:35:25,959 disengage, the captain initiates a go 652 00:35:26,007 --> 00:35:28,759 around without the autopilot. The captain 653 00:35:28,807 --> 00:35:31,119 needs to make inputs manually on the yoke 654 00:35:31,167 --> 00:35:34,435 to control pitch, but he doesn't. 655 00:35:35,095 --> 00:35:37,511 The autopilot is crucial because that 656 00:35:37,543 --> 00:35:39,711 autopilot prevents the airplane from 657 00:35:39,743 --> 00:35:41,863 pitching up. Whenever you're initiating a 658 00:35:41,879 --> 00:35:42,675 go around, 659 00:35:45,915 --> 00:35:47,419 I don't think the captain knows the 660 00:35:47,427 --> 00:35:50,947 autopilot is off. What the captain 661 00:35:51,011 --> 00:35:52,739 didn't seem to catch was that the 662 00:35:52,747 --> 00:35:54,563 autopilot was no longer flying the 663 00:35:54,579 --> 00:35:57,035 airplane. Even if the captain doesn't 664 00:35:57,075 --> 00:35:59,627 realize the autopilot is disengaged and 665 00:35:59,651 --> 00:36:01,323 the first officer doesn't bring it to his 666 00:36:01,379 --> 00:36:03,795 attention, it doesn't explain why neither 667 00:36:03,835 --> 00:36:06,163 pilot noticed the plane was pitching up to 668 00:36:06,179 --> 00:36:08,267 the point of stalling. Looking at your 669 00:36:08,291 --> 00:36:11,501 instmentation, you can tell that things 670 00:36:11,533 --> 00:36:13,021 are going wrong. You should make 671 00:36:13,053 --> 00:36:15,861 corrections. It becomes very difficult 672 00:36:15,933 --> 00:36:17,705 investigation to figure that part out. 673 00:36:19,565 --> 00:36:21,145 See up to the start of the go around, 674 00:36:23,285 --> 00:36:25,621 Investigators focus on the pilot's actions 675 00:36:25,733 --> 00:36:28,213 during the critical 11 seconds that led to 676 00:36:28,229 --> 00:36:30,637 a devastating stall of China Airlines 677 00:36:30,701 --> 00:36:32,265 Flight 676. 678 00:36:35,285 --> 00:36:36,705 Go lower. Go around. 679 00:36:40,395 --> 00:36:44,003 Gear up. Gear down. 680 00:36:44,179 --> 00:36:46,051 They discover that while the plane is 681 00:36:46,083 --> 00:36:48,491 pitching up, the pilots debate how to 682 00:36:48,523 --> 00:36:50,139 configure the plane for the go around 683 00:36:50,187 --> 00:36:52,295 maneuver. Cure up. 684 00:36:53,315 --> 00:36:55,491 Gear up. First. He should be calling for 685 00:36:55,523 --> 00:36:59,295 flaps before gear heading select plus 686 00:37:00,315 --> 00:37:03,355 plus 10 flaps. 687 00:37:04,095 --> 00:37:06,303 It distracts them from their instruments 688 00:37:06,359 --> 00:37:08,295 which show the plane is in danger of 689 00:37:08,335 --> 00:37:13,911 stalling. It's for a mess during 690 00:37:13,943 --> 00:37:16,367 the go around. That's when you really have 691 00:37:16,391 --> 00:37:18,183 to have your act together. You're close to 692 00:37:18,199 --> 00:37:20,167 the ground, a lot of things are happening 693 00:37:20,271 --> 00:37:22,199 and you've got to be on the same page. 694 00:37:22,327 --> 00:37:25,575 That didn't happen here. Investigators now 695 00:37:25,615 --> 00:37:27,543 have a better understanding of the pilot's 696 00:37:27,559 --> 00:37:30,387 behavior prior to the stall. But it 697 00:37:30,411 --> 00:37:33,307 doesn't answer a key question. Who turned 698 00:37:33,331 --> 00:37:35,095 the autopilot off and why? 699 00:37:36,395 --> 00:37:38,571 The captain was under a lot of stress and 700 00:37:38,603 --> 00:37:40,739 high workload. Under those conditions, 701 00:37:40,867 --> 00:37:43,539 people can make mistakes. He's near the 702 00:37:43,547 --> 00:37:45,415 end of his approach and he's too high. 703 00:37:45,795 --> 00:37:47,835 The team returns to the flight data 704 00:37:47,915 --> 00:37:50,651 recorder for clues. Somehow the autopilot 705 00:37:50,683 --> 00:37:54,155 gets switched off. God, the tail was 706 00:37:54,195 --> 00:37:57,025 really strong. Data reveals the captain 707 00:37:57,065 --> 00:37:59,233 did something unusual right before the 708 00:37:59,249 --> 00:38:03,665 autopilot is disengaged here. 709 00:38:03,785 --> 00:38:06,049 He's really at a crunch point. He's still 710 00:38:06,097 --> 00:38:07,841 too high. So he gives the control column 711 00:38:07,873 --> 00:38:08,925 an extra push. 712 00:38:11,185 --> 00:38:13,329 Investigators suspect the captain pushes 713 00:38:13,377 --> 00:38:16,169 the nose down hard to try to salvage his 714 00:38:16,217 --> 00:38:17,485 too high approach. 715 00:38:20,025 --> 00:38:22,337 But when they dig into the Airbus 300 716 00:38:22,401 --> 00:38:24,951 manual, they find the captain's actions 717 00:38:25,023 --> 00:38:27,195 had another tragic consequence. 718 00:38:29,735 --> 00:38:30,875 Take a look at this. 719 00:38:33,215 --> 00:38:35,879 The autopilot is designed to be disengaged 720 00:38:35,927 --> 00:38:37,767 in a number of ways. There's the routine 721 00:38:37,831 --> 00:38:39,835 way of just switching it off normally, 722 00:38:40,175 --> 00:38:43,775 but on this aircraft, if, if you apply 33 723 00:38:43,815 --> 00:38:45,927 pounds of force, it will automatically 724 00:38:45,991 --> 00:38:49,135 disengage. He disengaged the autopilot by 725 00:38:49,175 --> 00:38:52,311 accident and I 726 00:38:52,343 --> 00:38:55,095 think that's due to the task overload he's 727 00:38:55,135 --> 00:38:57,223 going through that. There's so much going 728 00:38:57,239 --> 00:38:58,999 on in his mind, there's so much happening 729 00:38:59,047 --> 00:39:01,799 here that he doesn't catch the fact that 730 00:39:01,847 --> 00:39:04,895 he's been pushing like mad and whoops, 731 00:39:04,935 --> 00:39:06,675 the autopilot just disengaged. 732 00:39:07,895 --> 00:39:10,471 Investigators now understand why China 733 00:39:10,543 --> 00:39:13,875 Airlines Flight 676 fell from the sky. 734 00:39:14,255 --> 00:39:15,727 It starts with bad teamwork in the 735 00:39:15,751 --> 00:39:19,063 cockpit. Seems like we 736 00:39:19,079 --> 00:39:20,195 don't have to do anything. 737 00:39:22,135 --> 00:39:24,791 Yeah, we do. See the trum tank system 738 00:39:24,823 --> 00:39:28,191 fault trim tank system in flight. 739 00:39:28,303 --> 00:39:30,151 In aviation, it's definitely not a good 740 00:39:30,183 --> 00:39:33,295 idea to do everything yourself. In a two 741 00:39:33,335 --> 00:39:35,519 person cockpit, there should be a proper 742 00:39:35,567 --> 00:39:38,551 division of labor. If there isn't that one 743 00:39:38,663 --> 00:39:40,567 pilot that's doing everything becomes 744 00:39:40,631 --> 00:39:43,179 saturated with tasks and doesn't have 745 00:39:43,307 --> 00:39:45,619 adequate capacity to handle potential 746 00:39:45,667 --> 00:39:48,099 challenges. God, that till'really strong. 747 00:39:48,187 --> 00:39:50,539 The overworked captain is too high on 748 00:39:50,587 --> 00:39:53,131 approach and when he attempts to descend 749 00:39:53,203 --> 00:39:56,123 faster, he unintentionally disengages the 750 00:39:56,139 --> 00:40:00,335 autopilot. Landing lights on. 751 00:40:00,755 --> 00:40:03,339 Neither pilot comments on the autopilot 752 00:40:03,507 --> 00:40:06,683 not in checklist complete when they 753 00:40:06,699 --> 00:40:10,191 are still too high on approach. Gor go 754 00:40:10,223 --> 00:40:13,535 around. Captain Kang calls 755 00:40:13,575 --> 00:40:14,715 for a go around. 756 00:40:17,415 --> 00:40:19,959 Not realizing the autopilot has shut off, 757 00:40:20,047 --> 00:40:23,231 he assumes it will control the plane. It 758 00:40:23,263 --> 00:40:23,955 doesn't. 759 00:40:27,975 --> 00:40:30,639 The Airbus pitches so high it begins to 760 00:40:30,687 --> 00:40:34,731 stall. By the 761 00:40:34,763 --> 00:40:36,255 time the captain notices, 762 00:40:38,235 --> 00:40:39,575 it's too late. 763 00:40:45,955 --> 00:40:48,975 What a shame. It was totally preventable. 764 00:40:50,715 --> 00:40:53,747 The aircraft was serviceable. And if the 765 00:40:53,771 --> 00:40:57,883 crew had acted differently, then the end 766 00:40:57,939 --> 00:40:59,655 result would have been different as well. 767 00:41:01,255 --> 00:41:03,719 In the final report, investigators 768 00:41:03,767 --> 00:41:06,183 recommend more training, including better 769 00:41:06,239 --> 00:41:08,527 cockpit communication for all China 770 00:41:08,591 --> 00:41:11,319 Airlines pilots. In response, 771 00:41:11,447 --> 00:41:13,863 China Airlines begins recruiting pilots 772 00:41:13,919 --> 00:41:16,591 with proven track records while improving 773 00:41:16,623 --> 00:41:19,743 their pilot training program. After the 774 00:41:19,759 --> 00:41:22,095 loss of this aircraft, crew resource 775 00:41:22,175 --> 00:41:24,495 management training was required to 776 00:41:24,655 --> 00:41:26,487 improve the communication and teamwork 777 00:41:26,551 --> 00:41:30,175 skills between the pilots. Taiwan also 778 00:41:30,255 --> 00:41:32,985 overhauls its Transportation Safety Board. 779 00:41:34,125 --> 00:41:36,165 It's now got an international reputation 780 00:41:36,245 --> 00:41:38,773 for their expertise in the technical 781 00:41:38,829 --> 00:41:40,669 aspects of investigation, including 782 00:41:40,757 --> 00:41:43,325 recorders, and also in publishing high 783 00:41:43,365 --> 00:41:45,613 quality reports. So that's been a real 784 00:41:45,669 --> 00:41:47,905 positive after the accident as well. 785 00:41:49,845 --> 00:41:53,237 In all accents, you want to gain the most 786 00:41:53,301 --> 00:41:55,505 out of it so you can prevent the next one. 787 00:41:56,045 --> 00:41:59,079 So from this tragedy, we learned and we 788 00:41:59,087 --> 00:42:02,207 adjusted the system and what we like to 789 00:42:02,231 --> 00:42:04,727 have is that we don't ever see an accent 790 00:42:04,751 --> 00:42:05,255 like this again.