Every Single Plane Crash – Air Disasters Seasons 15 & 16

A look at every single plane crash featured on Air Disasters Seasons 15 and 16.

#SmithsonianChannel #AirDisasters

Paramount+ is here! Stream all your favorite shows now on Paramount+. Try it FREE at https://bit.ly/3qyOeOf

Subscribe to The Smithsonian Aviation Channel: https://bit.ly/2UNavJO
Twitter: https://bit.ly/33lH712
Instagram: https://bit.ly/3iw9Iay
Facebook: https://bit.ly/3kkVOZp

[MUSIC PLAYING] Confirmed descent from 11,500. Let me know when you are positioned over Oksibil. Oksibil, we won’t be over the airport. We’re going to fly direct to a left base leg to Runway 11. Copy that. Left base leg? That’s not what it says in the charts. They took a shortcut? Oh, god.oh, god! [SCREAM] Pakistan International Airlines Flight 268 is en route to Kathmandu, the capital of Nepal. On the flight deck, Captain Iftikhar Janjua is in command. He’s a former officer in the Pakistani Air Force, who’s now a senior pilot instructor with the airline. The first officer is Hassan Akhtar, a junior pilot with the airline with almost 1,500 hours flying the Airbus A300. Flight 268 leaves its cruising altitude as it begins the approach into Kathmandu. In 1992, Kathmandu Airport doesn’t use radar to track its planes. Every movement, planned or otherwise, must be accurately reported by the flight crew. The mountains surrounding Kathmandu require pilots to make what’s known as a stepped approach. Pilots must descend to a series of lower altitudes at certain distances from the airport. Here it is. Watch my levels. The details of the stepped approach are laid out on a chart for the pilots. The crew works its way through the steps towards Kathmandu, approaching from the South. They should be touching down in about six minutes. Alpha Charlie Hotel, continue to taxiway number two. Pakistan 268 is at 10 miles. Report your level. We crossed out of 8,500– 8,200 now. Roger. Clear for final. Report 4 miles. Runway 0-2. We’ll call you at 4 miles, Pakistan 268. Clear for final. Runway 0-2. Runway 0-2. Flight 268 is now less than 10 miles from the runway. Air Traffic Control waits for the flight crew’s next position report. Three minutes go by. Flight 268 should have reported that they’ve reached the next step by now. Pakistan 268, Tower. Pakistan 268, this is Tower. As the minutes tick by, air traffic controllers come to a horrifying realization. Flight 268 is missing somewhere in the Himalayas. It doesn’t take much time for Kathmandu air traffic controllers to get word that PIA Flight 268 has crashed somewhere in the Himalayas. Within hours, the plane’s wreckage is found on the side of a mountain. The flight was carrying more than 20 tons of fuel, and the plane is completely consumed by fire. All 167 people on board have died. [MUSIC PLAYING] Nepalese soldiers have retrieved the aircraft’s black boxes from the wreckage of PIA Flight 268. Investigators look at the plane’s flight path as the crew prepares for landing. They started their descent at 16 miles out. They’re at 10,500 feet, then down to 9,500 at 13 miles, 8,200 at 10 miles, then quickly down to 7,500, until impact at 7,280. The data reveals an astonishing fact. So 1,000 feet too low right from the get go. Wait a minute. There you go. Perfect match. They were flying the approach, but one step ahead. What if the problem is the actual approach chart itself? To avoid the mountains surrounding Kathmandu, pilots follow an approach chart about the size of a large postcard. It provides the specific altitudes pilots must take at set distances from the airport. They study the same chart used by the flight crew. So is, let’s say, 8,200 the right altitude for 10 miles or 8 miles? It’s hard to say. Even harder when it’s actual size. Investigators need to know how the crew would have prepared themselves for the complicated approach. They examine the cockpit layout of an identical Pakistan International Airlines Airbus A300. There’s no place to clip an approach chart here. There is just here. Now, only the captain can see it. And he’d have to keep turning his head. Unlikely, they used that. More likely, he put it here. The other clip is a full 3 feet away from either pilot. That’s not much better. It’s almost impossible to read. And that handle is blocking the descent profile. They learned that Pakistan International Airlines has only two places for pilots to clip their charts. Both positions make it very difficult to read a 4 inch wide piece of paper. Could it be? Then, quite by accident, the investigators discover a key clue regarding the pilot’s misreading of the approach chart. This definitely makes it look like 10,500 is the correct altitude at 16 miles. Robinson finally has a theory he can put to the test. Kathmandu Tower, good afternoon. Pakistan 268, 25 miles at 11,500. Just a few minutes before impact, the first officer reports being 25 miles from the airport at the correct altitude of 11,500. Pakistan 268, report 16 miles. Wind 180. Expect Sierra approach. Roger. Call you at 16 miles. Expecting Sierra approach. Sierra approach, commander. But when the captain looks at the approach chart, his thumb may have obscured the correct next altitude. Here it is. Watch my levels. Now, instead of maintaining 11,500, Captain Janjua dials in the next altitude on the chart– Descending to 10,500. –oblivious that his plane’s approach is now one step ahead. The crew has now made a fatal mistake, unaware the airliner is now descending towards a violent collision with the side of a mountain. [MUSIC PLAYING] 33,000 above the border between Norway and Sweden, a CRJ200 jet cruises on autopilot during a late night mail run for one of Sweden’s largest cargo companies. ILS approach to Runway 01. Inbound heading 009. Minimum– It’s been a routine flight. And the plane is expected to land in about 30 minutes. We need to climb and turn right. Acknowledged. And according to last [INAUDIBLE],, we can expect light wind and zero– what the hell? Suddenly, the captain sees his plane is in a steep climb. [BEEPING] He pushes the nose down. [BEEPING] What? What? The pilots can no longer make sense of what the plane is doing. Mayday! Mayday! Mayday! Air [INAUDIBLE] 294. Mayday! Mayday! Mayday! 294, understood, mayday. What is the nature of your emergency, please? The plane loses almost 10,000 of altitude and speeds towards the ground at more than 450 miles an hour. Oh! Bank angle. Bank angle. Bank angle. Bank angle. West Air Sweden Flight 294– [CRASHING] –hits the ground at almost 600 miles an hour. The plane is obliterated. It takes three hours before rescue helicopters arrive. Even from the air, it’s obvious the pilots could not have survived. The next day, a team of investigators from the Swedish Accident Investigation Authority, known as the SHK, arrives at the crash site. Investigators waste no time analyzing the wreckage to pinpoint the northbound plane’s orientation as it crashed. That’s the front of the plane. So they were traveling east. This is the left wing, all right. But it’s in the wrong place for a plane flying east. This is [? Arvidsen’s ?] first clue about the accident. The left wing is found in the south end of the crater, and the right wing is in the north end. For a plane traveling east, it should be the other way around. If they were coming in this way and the left wing is over there, then they must have been upside down. The position of the wreckage and the size of the crater paint a picture of a plane hitting the ground off course, upside down, and at a blazingly high speed. [CRASHING] Incredibly, both of the plane’s black boxes survived the crash. They’re sent to France, where the data can be downloaded. Investigators hope the data will shed light on why the plane was so badly out of control. [MUSIC PLAYING] Investigators in Stockholm try to understand discrepancies in Flight 294’s flight data. Right here, the data shows a steep pitch up, but the plane flies level, and then pitches down, not up. Investigators soon discover that it’s not just the pitch data that’s off. Right here, heading and roll are wonky, too. All four parameters come from what’s called an Inertial Reference Unit, or IRU. It’s made up of gyroscopes that provide information to the cockpit displays and the flight data recorder. There are two IRUs, one for each pilot’s display. The FDR gets its data from the captain’s side IRU. The captain’s display and the flight data recorder both get their pitch data from the same source. Now, this is what the captain was seeing. Investigators are coming to a troubling conclusion. The captain’s instrument was telling him the plane was pitching up when it was still flying level. And that it was rolling to the right when it was actually rolling left. What the hell? It’s now clear, the pilot was receiving bad information from a faulty IRU. [BEEPING] Is it possible, the first officer was, too? The FO’s instrument has its own gyro. Investigators learned that IRU one was only feeding the captain’s instrument. A second IRU feeds the first officer’s display and is not recorded by the flight data recorder. So is it possible that both sides could have failed? If the first officer’s instrument was correct, he should have seen that the plane was flying straight and level. So why did he allow the captain to push the plane into a high speed dive? To find out, the team recreates the flight in a simulation based on the data from IRU one. OK. Start the animation. It paints an almost incomprehensible picture. Three seconds after his instrument shows a pitch up, he pushes the nose down. When the captain pushes the nose down, his ADI continues to show a pitch up, so he keeps pushing the plane into a steeper and steeper dive. And they begin to roll to the left. The plane continues to roll, until it’s on its back. Eventually, they do reach a speed of 508 knots? Flight 294 hits the ground inverted at a speed of almost 600 miles an hour. Investigators can now see what happened, but they still don’t understand, why the failure of a single instrument led to such a sudden and catastrophic crash. How do you go from level flight at 33,000 to 1,000 kilometers per hour impact in, what, one minute, 20 seconds? [MUSIC PLAYING] UNI Air Flight 873 is nearing Hualien Airport in Taiwan. Flaps 11. Flaps 11. The crew prepares to land right on schedule. But as they close in, the first officer struggles with the landing. Too high. Too high. [SCREECHING] You are sinking too fast. I have control. The captain takes over the controls. You have control. The captain engages the thrust reversers and hits the brakes. The plane touches down with plenty of room on the runway to come to a stop. But as the plane slows– [THUD] [FIRE] [SCREAMING] What the hell? [SCREAMING] A fire erupts immediately and spreads quickly. Passengers instantly panic. Moments after the fire starts, the power goes out. Mayday! Mayday! Mayday! Black smoke spews from the still rolling airplane. [FLAMES CRACKLING] [SCREAMING] Panic is as dangerous as the fire. You have to stay calm! The terrified passengers aren’t waiting for the plane to stop. They rush for the exits. Seconds after the blast, the plane comes to a complete stop. The captain issues the order– Evacuate! Evacuate! Evacuate! [SCREAMING] The 90 passengers and six crew members only have seconds to escape. The captain is the last crew member to leave the aircraft. Like some ship captains do, he stayed with his aircraft and made every effort to make sure everybody was off. [COUGHING] Anybody there? It takes more than an hour to contain the massive fire. [SIRENS] By the time it’s out, only 59 of the 90 passengers are accounted for. When firefighters finally extinguish the massive inferno, they board the ruined plane looking for the missing passengers. To everyone’s relief, there are no bodies aboard the burned wreckage of UNI Air Flight 873. The missing passengers are located, but it could have been much, much worse. [MUSIC PLAYING] Taiwanese and American investigators are struggling to determine what triggered the massive explosion aboard UNI Air Flight 873. They turned to the black boxes for clues. Hualien Tower, good afternoon. And as investigators listen, they soon discover, it’s uneventful. Except for a somewhat bumpy landing. After touchdown, the captain engages the thrust reversers and hits the brakes. What the hell? It came right out of the blue. There was absolutely no hint of trouble before the explosion. The sound itself could be the biggest clue on the audiotape. It wasn’t the very rapid, high frequency spike of a bomb. Send it in for analysis, please. The sound of the explosion on board Flight 873 is compared to sounds of various types of explosions. [DIFFERENT EXPLOSION SOUNDS] We determine it was a gaseous explosion. But the flight data recorder shows no irregularities with the fuel systems on UNI Air Flight 873. Whatever it was, exploded here. Investigators now focus on the overhead bin above seat 8B. They wonder if someone brought explosive materials on board. Investigator Tony James decides to examine the burnt out plane for clues that may have been missed. James focuses on the area around the overhead bin at row eight. On the far side, underneath seat 7C, he spots something unusual. I’ve found something. It’s not the look of the fragment that grabs his attention, but the smell. It smells like gasoline. The bottle fragment was found in row seven, just one row forward of the exploded bin. Could this be the source of the gas explosion? Investigators order up a chemical analysis of the bottle fragment. The fragment is soon identified as coming from a bottle of a common brand of laundry bleach. But lab results showed that it wasn’t bleach in the bottle. No question about it. It’s gasoline. It’s a remarkable discovery. An innocent-looking 750 milliliter bottle is now the prime suspect in the accident investigation. Investigators believe someone swapped the bleach for gasoline. The gasoline could have leaked out during the flight. Maybe. The team gets their first big break, but it raises new questions. Why would anyone bring gasoline onto an airplane? [MUSIC PLAYING] Parking brake. Set. Detroit Metropolitan Airport. The crew of Northwest Airlines Flight 299 prepares to depart from Memphis. Man, I don’t think I’ve ever seen fog this bad. In the tower, controllers are dealing with fog that’s blanketed the entire airport. The ground controller clears Flight 299 to taxi to the active runway three center. It will follow a series of taxiways known as Oscar, Foxtrot, and X-ray that will take you to the runway’s threshold. Northwest 299, what’s your position now? OK. We just turned down onto X-ray, 299. Roger. Switch to tower control 118.4. Roger. As Flight 299 nears the runway threshold, control of the plane passes from the ground controller to the tower controller. The 727 is now at the runway threshold, preparing for takeoff. Northwest 299, Metro Tower. Runway three center clear for takeoff. Roger. Flight 299 is cleared for takeoff. [PLANE SOUNDS] The 727 is seconds away from leaving the ground, but danger lies ahead. A DC-9 is stopped halfway down the runway. In the cockpit, Captain Bill Lovelace and First Officer Jim Schifferns are lost in the fog. This is a runway. Yeah. Oh, no. Oh, damn! There’s very little time to avoid a collision. Oh, god. oh, god! [CRASHING] [SCREAM] [SCREAMING] Inside the DC-9, it’s chaos. [SIRENS] Firefighters and emergency response teams raced to the scene. Eight passengers are killed, making it the deadliest incident of its kind on US soil in almost two decades. [MUSIC PLAYING] OK. Let’s see where they go. And let’s confirm any directional changes with the FDR data. Investigators combine the cockpit voice recording with the heading changes from the flight data recorder to better understand Northwest Flight 1482’s every movement on the ground. Three center, exit the ramp at Oscar Six. Did you get all that? Yeah, but I’m going to need you to help. Kind of wind around here and Oscar Six is going to be right around the corner here. OK. Investigators discover that as they began to taxi, Captain Lovelace asked his first officer to navigate. Left turn or right turn? Yeah. Well, this is the inner taxiway here. We’re still going for Oscar Six. So left turn? Yeah. So when they should go straight through Oscar Six, they turn off it and go east. We’re headed eastbound on Oscar Six here. Just as the controller said, you can’t go east on Oscar Six, Oscar Six runs north-south. Northwest 1482, you were on the outer taxiway. Yeah, that’s right. The investigators next hear the course correction that was supposed to get Flight 1482 back on track. Northwest 1482, at Oscar Four, take the right turn on X-ray. Report crossing 927. Roger. At an Oscar Four, make the right turn onto X-ray. The DC-9 crew was definitely told to continue eastbound and make a hard right turn onto X-ray to get back on track. So what do we do here? You make the right turn and report crossing 27. The captain is now relying completely on his first officer. OK. So the last instruction was at Oscar Four, turn right onto X-ray. This is– Should be 927. Are you sure? That’s what he said. Yeah. Yeah, this is 927. They are totally disoriented. Tell them we’re out here. We’re stuck. Somehow, they’ve made it here. And we’ve got two planes facing each other on the same runway. But there’s signs and surface markings all along here. I don’t get it. Why didn’t the airport signage and runway marks prevent the pilots from losing their way? This is a runway. [MUSIC PLAYING] Just after midnight, Aires Flight 8250 is carrying a planeload of tourists and locals from Bogota to the small Colombian island of San Andrés, an ocean playground in the Caribbean Sea. There are signs of a thunderstorm in the distance. Now past the halfway point of the trip, the captain is increasingly concerned with the possibility of heavy storms gathering over their destination. It doesn’t look like we’re going to get any relief from this storm. Yeah. Lots of activity tonight. Get a report from San Andrés for me. OK. I’ll ask. The San Andrés Air Traffic Controller is also keeping a close eye on changes to the weather. San Andrés, good morning. Aires 8250. Go ahead. We want to know how the weather is at the airport, please. Aires 8250, there is a thunderstorm over the airport right now. OK. Roger. Thank you. Gear down. The weather in San Andrés is changing fast. Aires 8250, there is now heavy rain at the airfield, and visibility reduced to 5 kilometers. In deteriorating conditions and just 7 miles from landing, the captain confirms the procedure in the event they have to abort the landing at the last minute. In case we miss the approach, we go right, correct? Yes. Correct, sir. 500 feet from the ground, the crew spots the runway. The captain lines up his plane. Aires 8250, winds at 60 degrees at 50 knots. Then, the wind picks up. 50 knots head on. In case of wind shear, we use maximum thrust for a go around. Don’t touch flaps or gear. Agreed. Yes, sir. Caught in a sudden deluge, visibility decreases again. 50. Look out, captain. The first officer now realizes they are too low to make a safe landing. Can you make it? 20. Climb, captain! The captain pulls back on his control column, but it’s too late. [CRASHING] [SCREAMING] At 1:47 AM, Aires 8250 slams into the ground with 131 passengers and crew on board. [SIRENS] In the aftermath of the recovery, one passenger is dead after being hit in the chest by a tray table. A second passenger, who was severely injured on impact, dies in the hospital two weeks later. The Colombian Aeronautical Authority needs to find out what brought down one of the most widely-used passenger planes on Earth. [MUSIC PLAYING] Aires 8250 has crashed just short of touchdown in San Andrés and broken into three pieces. While wreckage crews transport the debris to a nearby facility for analysis, investigators wonder what role the severe weather played in the accident. Can I show you this? The team already knows the weather was changing quickly just before the crash. But they need to determine if the air traffic controller kept them up to date. I updated them regularly once they where my airspace. San Andrés, good morning. Aires 8250. Go ahead. We want to know how the weather is at the airport, please. OK. Aires 8250, wind, calm. Visibility more than 10 kilometers. Scattered clouds at 1,500 feet. There’s some bad weather near the airfield, but right now, there is no rain. Every time the weather changed, even a bit, I told them. Gear down. Aires 8250, there’s a little drizzle over the runway. A little what? Drizzle. Ah. Minutes later, it turned into a full blown storm. Aires 8250, there is now heavy rain at the airfield and visibility reduced to 5 kilometers. Did they seem concerned about the weather? No. They didn’t sound overly concerned. I was very careful to tell them everything. Investigators need to know if there was lightning near the 737 as it approached San Andrés. Sure looks like lightning. The timing fits. The team studies the static wicks from the accident plane. Static wicks are located on the trailing edges of airplane wings and help discharge any buildup of static energy. If the plane was struck by lightning, they should be melted or burned. But they are not. Investigators can find no signs of a direct lightning hit on the aircraft surfaces. They’re clean. The team still has a lot of unanswered questions about the weather at the time of the accident. Wind shear is a phenomenon in which the wind’s direction or speed changes extremely abruptly. The team learns that just two minutes before the crash, as the storm built, winds near the runway were gusting. Digging further into the flight data. So lots of headwind, but almost no downdraft. Investigator Julian [? Etcheverry ?] discovers the powerful winds were hitting the plane head on not from above. Not enough to slam them to the ground. [MUSIC PLAYING] May 18, 2011. Sol Flight 5428, you are clear for takeoff. Roger. Sol Flight 5428 clear for takeoff. B1. Rotate. At 8:05 PM, Sol Flight 5428 lifts off from the airport in Neuquén. Positive rate. The plane will fly directly South from Neuquén to Comodoro Rivadavia on the Patagonian Coast. As per their flight plan, the Saab 340 will climb to a cruising altitude of 19,000 feet. Looks like we’re picking up some ice. Nothing to worry about. We’re expecting some light icing. It should be better when we get to 19,000. At 19,000, water droplets freeze solid, which keeps them from clinging to the aircraft. [TICKING] Building up some ice on my wing. OK. Let’s get it down to where it’s warmer to melt it off. Ask to descend to a lower altitude. 14,000. Roger. Then– Feel that? –the situation suddenly changes. Propeller’s vibrating. The pilots feel some vibrations. It could be a sign of ice building up on the propellers. Put them on max. Roger. [RATTLING] Oh, hell! [BEEPING] What’s wrong? I don’t know! The turboprop is suddenly pitching down and banking. [SCREAMING] Pull back! Pull back! Pull back harder! I’m trying! The pilots struggled to get the plane under control. Come on! Come on! Come on! Come on! Come on! [BEEPING] Pull up. [CRASHING] The turboprop crashes and explodes in a remote region of Patagonia. [FIRE CRACKLING] [SIRENS] Argentina’s Civil Defense Agency sends emergency crews to the crash site. But because of its remote location on a desert plateau in Rio Negro province, it takes hours for them to arrive. [HELICOPTER WHIRRING] When the crews do arrive, they find there are no survivors. [MUSIC PLAYING] OK. Let’s add the flight path. Investigators now turn their attention to the weather conditions on the night Sol 5428 crashed. This is the flight path from Neuquén to Comodoro. Look what’s waiting for them here. They flew right into this cold front. Investigators wonder if the pilots were aware of the icing conditions and took the proper steps to protect the plane. It’s something only one person can answer. The airline’s flight dispatcher provided the crew with crucial information. Investigators discover the report the flight dispatcher provided did not indicate the potential for severe icing. The report you gave them suggested icing would be mild, but it was actually severe. Where did you get your report? The team learns that because the airport’s meteorological office closed at 4:00 PM, the weather report the dispatcher provided was more than five hours old and no longer accurate. But establishing that the pilots encountered more severe weather than they expected is not enough to prove that ice brought down their plane. In icing conditions, pilots use the plane’s anti-ice protection. On the Saab 340A, that consists mostly of pneumatic boots made of rubber. The boots inflate and break up the ice on the wing’s leading edge. But did a malfunction of the ice protection bring down Sol Flight 5428? [BEEPING] Sol Flight 5428, mayday! Mayday! Mayday! [MUSIC PLAYING] Investigators of Sol Flight 5428 scour the wreckage for confirmation that a failure in the ice protection system brought the plane down. All that survived are the valves. OK. Let’s get them checked out. Yeah. The primary function of the valves is to control the flow of pressurized air that inflates the ice protection boots. Was there a problem with the valves? When they test the valve components, it’s discovered that enough pressurized air was passing through the valves to inflate the boots. Take a look. No failures. If the pilots knew they were flying in ice and the ice protection system was working, then ice alone doesn’t explain the accident. [MUSIC PLAYING] Cougar Flight 91 prepares for liftoff. This is a commuter run from Saint John’s, Newfoundland to offshore platforms in the Hibernia oil fields. The Hibernia oil fields are 170 miles from Saint John’s. By helicopter, it’s only a 90-minute flight. So for workers on oil rigs, it’s the primary mode of transportation. 28 minutes into the flight, Cougar 91 is leveled off at the standard cruising altitude. 9,000 feet. 9,000 feet, copy. Three weeks on, three weeks off. Life on a rig. I don’t know what’s worse, three weeks off, I’d get bored. [BEEPING] Gearbox pressure. Gearbox pressure. MGB oil pressure. Emergency checklist. While First Officer Lanouette looks for the checklist– Gander Center, Cougar 91, mayday. Cougar 91, go ahead. Sir, we have a main gearbox oil pressure problem. Request immediate clearance back to takeoff. Cougar 91, roger. You can make a right turn heading 300. The controller contacts the search and rescue headquarters in Halifax. Halifax, Gander Center. We have a mayday call from a Sikorsky S-92 currently on return to St. John’s, main gearbox problem. The pressure gauge indicates that the main gearbox oil pressure is well below the normal range of 45 to 70 PSI. Confirmed. MGB oil temperature greater than 130 degrees. But the reading on the temperature gauge is not what it would be if there were a loss of oil pressure. Cougar 91, dispatch. Go for 91. Rescue is asking if ditching is imminent, probable, or possible? Dispatch 91, ditching is possible. If possible, can you update our senior pilot on the situation? Dispatch, I think, we have an oil pump problem or an oil pressure sensor problem. Matt, can you describe the symptoms? The gearbox oil temperature is still normal. I don’t think we’ve lost all the oil in the system. I’m on a heading for closest landfall if this goes south on us. Cougar 91 is still 35 miles from the nearest land. But now– [RATTLING] Oh, hell. Dispatch, we’re ditching. [BEEPING] Ditching. Airspeed 122, Matt. Brace! Brace! Brace! Brace! [CRASHING] The passengers and crew of Cougar 91 are now trapped inside a sinking helicopter. [WAVES CASHING] [MUSIC PLAYING] 17 people are dead. There is only one survivor. The close-knit offshore oil community is shattered and looking for answers. The answers lie 554 feet beneath the surface of the Atlantic Ocean. [MUSIC PLAYING] Investigator Alan Chaulk’s recovery of Cougar 91 provides a big break in the investigation. Within days of the wreckage recovery, the helicopter’s manufacturer, Sikorsky, has questions about the crash. That’s right. The whole main gearbox. Sikorsky. All right. Yeah, good. Yeah. We’ll be expecting you. Great. They’re coming? Yeah. And the FAA. They want to take a look at the whole main gearbox. We are. It turns out, Sikorsky and the US Federal Aviation Administration, or FAA, are extremely interested in the same piece of wreckage as Cunningham. Investigators learned that eight months earlier, a Sikorsky S-92 in Australia got the same gearbox warning. [BEEPING] Gearbox pressure. Gearbox pressure. But that S-92 reached land in less than seven minutes without incident. The circumstances are identical. Is there some flaw in the design of the main gearbox? So what went wrong? Right here, the oil filter bowl. The oil filter bowl filters the 11 gallons of oil circulating through the main gearbox. It’s fastened to the gearbox with three studs. The investigators now focus on this key component. Here we go. Wow. Two of the studs are gone. | discovery confirms what investigators saw in the FDR data, an extremely sudden loss of oil. They now have concrete evidence that oil leaked because of the broken studs. So what made the studs fail? What is it made out of? These ones? Titanium. This looks like galling. Galling is a form of wear caused when two surfaces are moving against each other, particularly two different metals under pressure. Titanium is prone to galling, especially when in contact with steel. Sikorsky studied the Australian helicopter’s broken studs and concluded the galling was caused by the steel nuts used to fasten them. Here. Take a look at this. Alert two operators of the S-92. Replace all titanium studs within a year or 1,200 flying hours. Six weeks before the Newfoundland crash, Sikorsky issued an advisory to swap the titanium studs for steel studs. They knew the studs could fail. Sikorsky said that the titanium studs would last at least a year. Cougar got the replacement parts, but maintenance didn’t think that it was urgent enough to replace the studs immediately. The question now is, why did Cougar delay replacing the studs? [RATTLING] Oh, hell. Dispatch, we’re ditching. [MUSIC PLAYING] Sentani Airport in Papua Province, Indonesia. Trigana Air Service pilots Captain Hasanuddin and First Officer Aryadin Falani are on a short layover before their fifth flight of the day. The pilots have already flown four short flights today. This leg will take them back to a remote airport in the Oksibil area. The plane reaches its cruising altitude of 11,500 feet. The flight is so short, the captain is already preparing for the descent into Oksibil. Without the aid of advanced landing systems seen at bigger airports, pilots flying into Oksibil must keep the runway in sight during the approach. Wings there, 1250. Confirm descent from 12,000. Report when you are positioned over Oksibil. Five minutes later, the air traffic controller at Oksibil realizes Flight 267 hasn’t reported in as expected. Hey, this is Oksibil. We’ve lost contact with Flight 267. Have you heard from the pilots? You haven’t. We might have a problem here. Word of the missing plane travels fast. As the news spreads, investigators with Indonesia’s National Transportation Safety Committee, the KNKT, fear the worst. Flight 267 was only 15 miles from Oksibil Airport when it last checked in with Air Traffic Control, but search efforts come up empty. Until a pilot flying out of Oksibil Airport spotted smoke billowing from nearby Tangok Mountain. [PHONE RINGING] Yeah? 140 degrees. 29 minutes. 51.18 seconds east. OK, got it. Thank you. The wreckage is at approximately 8,300 feet on a ridge of Tangok Mountain, 10 nautical miles from Oksibil Airport. A search and rescue team takes a KNKT investigator to the crash site. At over 4,000 feet above ground level, there are no roads, runways, or clearings. The only way to get to the crash site is on foot. The team finally reaches the crash site. A field of debris stretches through Tangok’s thick forest. [PHONE RINGING] [BEEP] Go ahead, Oksibil. Jakarta. We reached the crash site. No survivors. [MUSIC PLAYING] KNKT investigators are facing intense pressure to determine why Flight 267 crashed. Without recorded flight data, the plane’s cockpit voice recorder is investigator’s last hope to learn what was happening in the air at the time of the crash. Fortunately, the CVR data is successfully recovered. All right. We’re ready? OK, play it. The CVR contains two hours of audio, including from the two flights prior to the accident flight. I’ll call down and let them know we’re coming. Investigators first listened to the pilot’s flight into Oksibil from earlier in the day, in the hopes of finding clues. Oksibil Trigana 267. Trigana 267, Oksibil, copy. We’re beginning our approach. Would like to descend from 11,500 feet. Confirm descent from 11,500. Let me know when you are positioned over Oksibil. Oksibil. We won’t be over the airport. We’re going to fly direct to a left base leg to runway 11. Copy that. Left base leg? That’s not what it says in the charts. They took a shortcut? The CVR reveals that the pilots made an intentional deviation in their approach during their prior flight into Oksibil. The official approach directs aircraft to fly over the airport, and then circle back and land. But on the earlier flight, the pilots flew directly to runway 11 without looping around. So they take the same shortcut for the accident flight. The last flight proceeds much like the crew’s earlier flight into Oksibil. Big plans for the family? Investigators listen closely for any signs of danger. Yeah. Looking forward to some time off. It’s been very busy these past few weeks. It’s what they don’t hear that raises questions. They’re not doing their approach briefing and checklist. The team now knows that the crew missed a crucial step on their approach. While listening to the final moments of the crash, investigators make a baffling discovery. That’s it. [MUSIC PLAYING] It happened so fast. Sounds like they didn’t even see the mountain. [CRASHING] [MUSIC PLAYING] There is still no evidence why LAM 470 flew into the ground in Namibia. The investigation moves from the field to DAAI offices in Windhoek, Namibia. Is this the last inspection? Looks like it. They begin looking into the way Mozambique Airlines was maintaining the plane. November 28. Day before the accident. Everything seems to be done by the book. The Embraer 190 purchased brand new just a year ago has a spotless record. Yeah. There was nothing wrong with the aircraft. With no evidence of a mechanical problem, the team turns his attention to outside factors. They review meteorological reports from the day of the crash. All right, Dennis. The data from [? Mon ?] weather station is in. Did bad weather bring down Flight 470? Great. Why don’t we start from the takeoff in Maputo? The team scrolls through a series of satellite images spanning the duration of the flight. Just some scattered clouds 3,000 feet throughout the flight. Nowhere near the cruising altitude. What was the wind like? Nothing stronger than 11 knots. So no problem there. With both the machine and the environment ruled out, investigators have only one more avenue to pursue. OK. Let’s look into the pilot records. Were the pilots responsible for the crash of Flight 470? Investigators consider if the pilots of LAM 470 played a role in causing the crash. Captain Herminio dos Santos Fernandes. Born in Mozambique, 49 years old. Here we go. How familiar was Captain Fernandes with the Embraer 190? Positive rate. Gear up. Gear up. The team scrutinizes his pilot records. Captain Fernandes was not only well trained– 2,500 on the Embraer 190. 9,000 flight hours total. –he was a very experienced airman. Investigators have exhausted all available avenues of inquiry and come up empty. Any news on the recorders? Not yet. But I’ll check in. OK. Search-and-rescue aircraft deploy to try and find Mozambique Airlines Flight 470. But it’s park rangers in Namibia’s remote Bwabwata National Park who first come across a tangle of wreckage. It’s clearly Mozambique Airlines Flight 470. All 27 passengers and the flight crew are dead. The site is in a roughly 280-mile strip of Namibia, straddling Botswana and Angola, putting the investigators from Namibia’s DAAI in charge. OK. We got a lot of ground to cover. Let’s find the point of impact. The team combs through the debris for any clues that could help explain the crash. The last radar plot of the flight showed the plane was in an unusually steep descent moments before the crash. But the crash site adds important details. OK, so the plane came down there and was headed in this direction, towards Luanda. Flight 470 was headed in the direction of its destination. It was not off course. Investigators look for other wreckage patterns to see what else they can learn. At the point of initial impact, investigators find two nearly identical pits. 32 feet. That’s the distance between the two engines. It’s an important lead that shows the investigators the way in which the plane hit the ground. Investigators study the actual shape of the crash site. OK, we’re 487 meters from the first point of impact. The wreckage pattern is long and narrow. The extensive trail of debris leads to one possibility. Maybe they were trying to land. To answer that question, the team examines the landing gear. Tires aren’t blown. No damage to the treads. No puncture marks. They were retracted. This wasn’t an emergency landing. It’s the first tangible clue about what the pilots may have been doing. If they weren’t trying to land, why were they coming in so fast? In Douala, Cameroon, Kenya Airways Flight 507 is almost an hour behind schedule. Ladies and gentlemen, we are going to wait for the weather conditions to improve before we take off. A severe storm must pass before the pilots can depart. Tower. Kenya 507. Looks like there’s a break in the weather. Requesting startup. OK, 507. Startup approved. Startup checklist. Quickly, please. Uh, generator’s on. A few minutes before midnight, Flight 507 is finally ready for takeoff. OK. Takeoff thrust is set. Speed building on both. Check. 80 knots. Check. V1. Rotate. Just after midnight, Kenya Airways Flight 507 finally departs for Nairobi. [INAUDIBLE] [BEEPING] Bank angle. Bank angle. Suddenly, the 737 rolls dangerously to the right. The captain fights to level the plane. Bank angle. Bank angle. But it continues banking further and further right. [BEEPING] Bank angle. Bank angle. Despite the pilots’ efforts, they keep rolling. [BEEPING] Bank angle. We’re crashing! Yeah, we are crashing! Left! Left! Bank angle. Left correction. Bank angle. Bank angle. Five hours later, the controller in Douala is ending his shift. [PHONE RINGS] Douala tower. No reason I can think of. They should be there by now. Kenya Airways Flight 507 should have landed in Nairobi 45 minutes ago, but the plane has still not arrived. [INAUDIBLE] And no one even knows where to start looking for it. Two days later, search teams finally locate the wreckage of Kenya 507. Its 3 and 1/2 miles southeast of Douala Airport. The 737 has crashed into a mangrove swamp and is submerged in mud and water. There are no survivors. The crash of Kenya Airways Flight 507– a key piece of evidence is quickly found– the 737’s flight data recorder. Well, looks pretty good, considering. Technicians in Canada have been able to extract the FDR data. Any malfunction warnings during the flight would provide an important clue. No warnings. They find no signs of any system failure. As the team continues looking through the data– But look at this. –they discovered that just over a minute into the flight, a different kind of alert did go off. Bank-angle alert at an altitude of 2,800 feet. For investigators, this is their first real clue about what happened to Flight 507. Any rule exceeding 35 degrees is considered extreme. Once reaching that angle, a warning alerts pilots of danger. Rotate. When investigators review the pilots’ inputs after takeoff, they make a critical discovery. Gear up. OK. He was turning left from the moment they left the ground. And then, he keeps making corrections to the left. It looks like he’s trying to keep the plane level. But what caused the slow roll to the right in the first place? Bingo. It’s the way the flaps are rigged. It was a slight right roll. The flaps increase a plane’s lift at slower speeds. On this particular airplane, the left flap provided slightly more lift than the right one. This discovery explains why the captain kept his control column turned to the left immediately after takeoff. Well, that’s weird. But it doesn’t explain an even more puzzling element of the flight data. Look at this. The captain’s control wheel inputs seem to stop. It’s like he just let go of it. Six degrees right, he does nothing. 11 degrees, 15 degrees, 20 degrees, 30 degrees, and still nothing. As the plane’s bank angle approached 35 degrees, neither pilot took action to stop the increasingly dangerous roll– [BEEPING] Bank angle, bank angle. –until the bank angle warning sounded. Bank angle. And then, he does this. Right, left, right again, left again. Bank angle. Bank angle. Bank angle. The data paints a baffling picture of the 1.5-minute flight. Who flies like that? That was puzzling to us. So it really highlighted that we needed more information and more data. And that data is the CVR. The CVR is what tells us the conversation. What might they have been looking at? But after weeks of searching the crash site, there’s still no sign of the cockpit voice recorder. It’s 6:45 PM. Loganair Flight 6780 cruises over the North Sea near Scotland. Take a look at that. I don’t like it. There’s some bad weather developing off the end of the runway. Approach, Logan 6780. There’s a big storm cell on radar just off the runway. We might need to discontinue. 6780, roger. The pilots consider their options. How much fuel do we have left? 2,500 kilos. You want to head back to Aberdeen? Well, let’s circle. Make another attempt. As the pilots circle for another landing attempt– Oh, crap! Circuit breakers look fine. I have control. But something’s wrong. The controls feel really heavy. Mayday, mayday, this is Loganair 6780. Please clear the airspace. 6780, copy. Without knowing what’s wrong with the plane– It’s really fighting me. –it’s too risky to attempt landing in a storm at Sumburgh Airport. We should divert back to Aberdeen. Better conditions. Agreed. Aberdeen Airport is 190 miles away. Let’s try climbing to 4,000. 6780, we’re going to divert back to Aberdeen. Something’s wrong. I can’t get the plane to climb. How’s your side? It’s really heavy. And then, the unthinkable happens. Oh, no! Come on! We’re dropping! [GROANS] Flight 6780 plunges uncontrollably towards the North Sea. The pilots wrestle with their controls as the plane speeds toward the water at 350 miles an hour. Speed. Speed! In a last-ditch effort to save the plane, the captain increases engine power. We’re climbing. Requesting flight level 240. 6780, roger, climb and maintain, 240. The controls are working better now. The pilots of Loganair Flight 6780 are on final approach to Aberdeen Airport. Aberdeen, 6780, established on the localizer, runway 16. Not knowing which instruments they can trust, the pilots carefully configure the plane for landing. Flaps 35. Flaps 35. Control feels normal. Having avoided disaster twice already, their sole focus is to get the plane safely on the ground. 130. Speed looks good. Decision height. Runway in sight. Continue. All right. 50 knots, coming out of reverse. Check. I’m happy to be on solid ground. Puzzled by the crew’s account of the incident on board Flight 6780, investigators turned to the cockpit voice recorder for answers. It will answer many questions as to why things were happening the way they were and how the crew were working together and interacting together. Aberdeen ground, 6780, taxiing on whiskey for parking stand 7. But something’s not right. Aberdeen? This is from when they landed back at Aberdeen. Right, can you stop it and go back to the top to play it again, please? Aberdeen ground, 6780, taxiing on whiskey for parking stand 7. That’s all there is. Well, that’s not going to help us very much now, is it? The CVR has recorded over the critical moments of the flight. It’s a major setback for the investigation. Right, this is what we know so far. The lightning struck. The autopilot disconnected. They had control problems. Then, the plane did a nosedive. And let’s review the data. Will the data recorder provide the answers they need to solve this case? Flight data recorders often add a level of detail that simply can’t be gained from the witnesses themselves. And crucially, quite often, the information that’s gathered from a recording device offers a slightly different perspective to what we might get from personal recollections. Stop. This is where the lightning struck, at 2000 feet. Right. And then for the next 2.5 minutes, there’s a slow, uneven climb to 4,000 feet. And then, they are in a very steep nosedive, 20 seconds towards the North Sea. What were the pilots doing to recover from the dive? Look at the control column data. Right after the lightning struck, the crew pulled back on the control column to pitch the nose up. The FDR data confirms the pilots were tackling a control problem. But why? The captain said they were also applying pitch trim. Pitch trim moves the tail elevators up and down to maintain the pitch of the aircraft. Could we take a look at the pitch trim data, please? All right. Investigators make a puzzling discovery. Look at that. The elevators are trying to get the nose to pitch down instead of up. After the lightning strike, some unknown force was fighting the pilots’ inputs to both the control column and the pitch trim. The crew said the lightning struck, the autopilot disconnected, and then they had control problems. Right. Could we take a look at the autopilot data, please? Finally, the team zeros in on the answer. The autopilot was on almost the entire time. Investigators discover that after the lightning strike, the pilots were in a tug of war with the plane. So we then had to look at, why did the crew misunderstand the status of the aircraft? Execuflight 1526 is flying a short 35-minute flight northeast from Dayton to Akron, Ohio. Zipline 1526, descend to 13,000. "Zipline" is Execuflight’s call sign. Descending to 13,000 feet. Thank you, Zipline 1526. Let me check the weather. Automated weather observation. Wind, 290 at 07. Overcast, 1,800. Temperature, 09 Celsius. All right. We have overcast weather. The crew prepares for possible bad weather in Akron. OK, let’s see. Akron. Right. Heading? 249. Flight 1526 continues its descent. Akron, visibility 1-and-1/2 mist. Sky condition, overcast, 600, broken. The crew learns the weather in Akron is getting worse. Temperature, 11 Celsius. They need to know if it’s still possible to land there. 1-and-1/2-mile visibility. What visibility does this approach want? 1-and-1/4 miles. All right, so we have visibility. Six minutes from Akron, the crew has a new distraction. A flight instructor is teaching a student pilot how to land in bad weather on the same runway assigned to Flight 1526. Flight 1526 is now four miles from the runway as it starts its final approach. Can you check if I’ve got everything? Ignition. Everything is all set. Stand by. 2 and 1/2 miles from the air, the pilots are still searching for the runway. Finally, the plane breaks through the clouds. Ground. Keep going. OK, OK, level out now. I got it. Pull up. No, no, no, no, no, no, no, no! Execuflight 1526 has crashed into a two-story residential building. Everyone on board is dead. Incredibly, none of the residents were home during the crash. The NTSB must now determine what caused this fatal accident. Execuflight 1526 plunged into a residential neighborhood in Akron, Ohio. At the crash site, investigators combed through the wreckage for clues. We found the angle of attack indicator in the cockpit wreckage. That’s important, because at a certain angle of attack, the wing will stall. Oh, what angle were you at? It was damaged extensively. But they could see that the needle was in the red band. Investigators determined that the plane stalled as it approached the airport. Now, they must understand why. Pull up. Pull up. That’s great. Get that to Washington. The cockpit voice recorder from Execuflight 1526 is recovered and sent to NTSB headquarters for analysis. The CVR is doubly important in this case, since the Hawker 700 wasn’t equipped with a flight data recorder. While investigators wait for the voice recording, they work with what they have– the documents found in the cockpit wreckage. Weight and balance. Will you look at that? We found that the weight and balance didn’t account for the auxiliary power unit. It’s a little jet engine in the back that helps power the aircraft when it’s on the ground. So they had no APU. The team wonders how this compares to what they discovered at the crash site. Hello, APU. Looks like they were carrying more weight than they thought. Investigators believe they found an error in the plane’s documented weight and balance. Were they too heavy? Wonder how much this plane truly weighed. The pilots made their calculations without accounting for any APU. But there clearly was one on board. The NTSB calculates the actual weight of the plane during its final flight. The APU weighs 300 pounds. And according to the aircraft refueller, they were loaded with 8,160 pounds of fuel, but they only wrote down 7,700. So how much in total were they over by? Oh, there were only 286 pounds overweight. The plane’s actual weight was slightly more than what the pilots recorded. It wouldn’t really have made a performance difference on the aircraft. The weight of the APU and the additional fuel was not enough to affect the balance of the airplane. But it did tell us that this crew and this company wasn’t following their procedures appropriately. Somebody wasn’t watching what they were supposed to be doing. Investigators need to look elsewhere to explain why Flight 1526 stalled. It’s almost 6:00 in the morning in Lexington, Kentucky. The pilots of Comair Flight 5191 prepare for their pre-flight briefing. Ladies and gentlemen, we’d like to take this time to welcome you on board Comair Flight 5191, direct to Atlanta. We’ll try to keep it as quiet as possible. Hopefully, you can catch a nap on the way there. It’s our pleasure having you on board today. There are 47 passengers and one flight attendant on today’s flight. The pilots begin their briefing. Right flex. Takeoff procedures off of– he said what runway? 24? It’s 22. Today, they are bound for Atlanta, Georgia, just a 67-minute flight straight south from Lexington’s Bluegrass Airport. Lexington’s air traffic controller clears Flight 5191 to the runway. At your leisure, Comair 121. Ready to go. Comair 191. Lexington tower. Fly runway heading, clear for takeoff. Oh. All yours, Jim. Captain Clay hands control of the aircraft to First Officer Polehinke for takeoff. My brakes, my controls. Set thrust, please. Thurst set. That is weird with no lights. Yeah. 100 knots, checks. Something’s not right– V1, rotate. –as 5191 prepares to lift off. Whoa. Damn it! Comair Flight 5191 hurtles into a field less than 1/2 a mile from the runway. This is Lexington. Alert 3. West side of the runway with a Comair regional jet taking off. 49 people are dead after the tragic takeoff of Comair Flight 5191. First Officer Polehinke is the only survivor. The question NTSB investigators now need to answer is why Comair Flight 5191 couldn’t get off the ground. NTSB investigators begin looking for clues to discover why Comair Flight 5191 crashed less than 1/2 a mile from Lexington’s Bluegrass Airport. The wreckage path tells investigators that the plane was struggling to get off the ground. Was it engine trouble or something else? It looks like they took off from Runway 26. But hang on. 26 wasn’t in use last night. Lexington Airport has two runways– Runway 26 and Runway 22. According to the flight plan, they were supposed to take off from Runway 22. How the heck did they end up over here? Investigators are surprised to discover that Flight 5191 took off from the wrong runway. But being on the wrong runway doesn’t explain why the plane barely made it off the ground. Let’s take a closer look at those engines. Yeah, they were spinning all right. Evidence of deformed blades suggests the engines were running on impact. Something else must have kept ’em from getting airborne. Maybe they were too heavy? With engine failure ruled out, the team wonders if the plane was loaded with too much weight to take off. Hmm. Oh, here it is. CRJ’s max takeoff weight is 50,178 pounds. And the load manifest said the plane weighed 49,087 pounds on that day. It’s close, but it’s within their limit. The aircraft wasn’t too heavy to take off. In that aircraft on that day, with that weight, how much runway would be needed to take off safely? Using the actual weight of the aircraft and the CRJ 100’s specifications, investigators calculate how much runway the plane needed to lift off. 3,744 feet. 3,744 feet. And Runway 26 is– 3,501 feet. Investigators reach an astonishing conclusion. Runway 26 is 243 feet too short. They ran out of runway. They now understand why the plane couldn’t get airborne. But they’re still mystified. Why didn’t Flight 5191 take off from the longer runway? It’s just after 5:00 AM in Jakarta, Indonesia, as 181 passengers settle in for a short domestic flight. Lion Air Flight 610 is a 90-minute journey from Jakarta over the Java Sea to Bangka Island. At 6:20– Rotate. –the MAX 8 lifts off the runway at Jakarta. [BEEPING] But immediately, there’s trouble. The captain’s control yoke starts shaking– a warning that the plane is about to stall. Take off config. OK, but what? The pilots can’t identify the source of the problem. They have no choice but to continue climbing. 5610, fly heading 248, follow standard instrument departure. Lion Air 610. The air traffic controller has no idea that there’s an issue in the cockpit. The pilots get a warning that their airspeed indicators do not agree. Airspeed disagree. What’s going on? Should we request a return to Jakarta? Landing gear up. Lion 610. Climb to flight level 270. Still unaware of any trouble, the controller instructs the crew to continue climbing to 27,000 feet. Altitude disagree. The first officer now notices that the altimeters also show conflicting readings. [? Knowledge ?] altitude disagree. The situation is deteriorating quickly. Now flying 5,000 feet above the sea, Captain Suneja struggles to keep the plane’s nose up. Flight path vector may be unreliable. Lion 610, turn right heading 070 to avoid traffic ahead. Set the pitch attitude. Roger, heading 070. Flight 610 is flying erratically over the Java Sea and becoming increasingly more difficult to control. The controller allows Flight 610 to fly at any altitude the pilots choose. Uh, please clear 3,000 above and below of traffic. OK, will do. What altitude would you like? 5– uh, it’s diving. It’s diving. It’s OK. It’s OK. Flight 610 is now speeding towards the sea. Airspeed. And the pilots are out of options. Fly up. Up. Up! 30 minutes after crashing into the Java Sea, the wreckage of Lion Air Flight 610 is located. There are no survivors. It takes three days for search-and-rescue teams to lock on to the signal coming from the MAX 8 Lion Air Flight 610’s flight data recorder. Divers recover it from a depth of 115 feet. The recorder has preserved data from the accident flight and 18 previous flights covering almost 1,800 different parameters. Master caution goes off as soon as they leave the ground, probably because airspeed and altitude don’t agree. Stick shaker activates here. The data shows a repeat of the problem on the previous flight. Faulty readings caused by a discrepancy between the left and right angle-of-attack sensors. [NON-ENGLISH SPEECH] From the FDR data we received, we learned that this plane had faulty angle-of-attack readings that affected both flights similarly. Left and right angle-of-attack values are off by 21 degrees for the entire flight. We suspected the new angle-of-attack sensor installed in Bali was either faulty, or the installation process was done incorrectly. On the MAX 8, the angle-of-attack sensor doesn’t just measure the airplane’s angle. Altitude disagree. It helps calculate precise airspeed and altitude. That explains why airspeed and altitude disagreed throughout the whole flight. The malfunctioning sensor on the captain’s side resulted in a difference between the left- and right-side speed and altitude displays. Then there’s this. The data clearly shows that for every nose-up trim input, there is a corresponding automatic nose-down trim input. [NON-ENGLISH SPEECH] The NTSB brought a representative from Boeing to help us in our investigation. We’re trying to understand these automatic inputs. It looks like the MCAS kicked in. What’s that? Boeing points to an obscure automated system known as MCAS, the Maneuvering Characteristics Augmentation System. It only kicks in when these three conditions are met. Boeing explains that MCAS only activates when it senses that the angle of attack is excessive, when the autopilot is off, and when the flaps are retracted– an extremely rare combination. Tragically, the data shows that because of the faulty maintenance and the angle-of-attack sensor, Flight 610 ended up meeting all three conditions. [NON-ENGLISH SPEECH] The series of problems occurred when the left angle-of-attack sensor was replaced in Bali. Investigators dig deeper into the data and discover the MCAS system had no failsafe. [NON-ENGLISH SPEECH] The MCAS installed in the plane relied on only one sensor. MCAS only took data from one angle-of-attack sensor, not both. [BEEPING] Most protection systems are designed with redundancies, so a single failure doesn’t result in catastrophe. Southwest Airlines Flight 1380 is boarding for a trip to Dallas, Texas. The crew flew in earlier today from Nashville. The four-hour trip to Dallas will be their second and final flight of the day. When this Flight 1380 climbs to cruising altitude, controllers at LaGuardia hand the flight over to New York area controllers. Southwest 1380, contact New York Center, 133.47. Copy that, 1380. 20 minutes after takeoff– Thank you. –everything changed. What the– oh! We had a very large bang. We had multiple warnings going off in the cockpit and a very severe vibration throughout the entire plane. First Officer Ellisor struggles to control the aircraft as it banked steeply to the left. I immediately grabbed the yoke to stop the roll. You still got it. I was not able to see any of the instruments, because the vibration was so severe. It was just a blur of colors. And so I can’t see anything. Still got it. Luckily, it was a clear day, a very clear horizon, and I was able to roll out of the bank and recover the airplane. First Officer Ellisor reduces engine power and begins a steep descent. Flight 1380 is 10 minutes away from Philadelphia International Airport and closing fast. At an altitude of just 1,000 feet, and 3 miles from Philadelphia International Airport, the pilots prepare for an emergency landing. Southwest 1380, runway 27 left, cleared to land. 27 left cleared to land, Southwest 1380. It’s seconds before touchdown. Flight 1380 is flying towards the runway at breakneck speed. Speed break. Armed with the green light. The pilots are unsure of the damage to the plane. They’re making a high-speed approach with a reverse thrust from only one engine to slow them down. They might not have enough runway to stop safely. 50 feet. 30 feet. Captain Scholz is a veteran Navy pilot. She’s landed F-18 Hornets in war zones. But this is a landing unlike any other. Speedbrakes up. The thrust reverser on their only engine deploys. If the reverser doesn’t work, the plane could overshoot the runway. We touched down. It was a great landing. Thank you, Lord. Thank you. Thank you. Thank you, Lord. I’m just going to pull her around here to the fire trucks. Flight 1380 rolls to a stop. Investigators study the inspection history for the fan blades in the left engine of Flight 1380. They did a full overhaul in 2012. Let me see. They learn that all the fan blades, including fan blade 13, were inspected during a major overhaul in 2012, six years before the incident on Southwest 1380. During the overhaul, the protective coating on each of the blades is stripped. Then, a fluorescent dye is used to help identify any fatigue cracking. Investigators review the work orders done on the fan blades during the 2012 overhaul. This checks out. It says they did a full inspection of blade 13. Investigators determined that at the time of the 2012 overhaul, all the fan blades were found to be in good condition. But how were the fan blades maintained by Southwest Airlines after the overhaul? They’re supposed to be lubricated and visually inspected between 1,500 and 3,000 flights. There are seven more routine checks here. They’re all comprehensive, all done on time. This was the appropriate and approved maintenance process that all technicians used at that time. If fan blade 13 was checked routinely for six years and passed all its inspections, investigators wonder when the crack began. So what do you got? Have a look. A microscopic examination of the fracture surface might tell them more about when the metal fatigue started. I see. The crack was growing. Using high magnifications, investigators can see tiny tracings called striations. There’s thousands of them. Investigators tally the striations on the fractured blade. That means counting tens of thousands of microscopic marks on a tiny piece of metal. There’s over 32,000 striations. What’s that tell you? By counting the striations in the base of fan blade number 13, investigators are able to date the beginnings of the crack. This crack could have started more than six years ago. It’s an important development. The fatigue crack on fan blade 13 likely began before the engine overhaul in 2012, when the blades were under close inspection. The pressure on investigators mounts. There are more than 4,000 Boeing 737s in service using the same type of fan blades. The same inspection regime that missed the growing crack is also used throughout the aviation industry. The catastrophe that struck Flight 1380 could happen again. US-Bangla Flight 211 cruises above the Himalayas. Their destination is Kathmandu Airport in Nepal. It’s a 90-minute flight from Dhaka to Kathmandu. The crew is flying a Bombardier Dash 8 Q400. BS 211 tower, wind 8 knots, Runway 02, continue approach. BS 211 is on final approach to Runway 02. Tribhuvan International Airport has one 10,000-foot airstrip. Planes approaching the airport from the south land on Runway 02, and from the north, on Runway 20. But BS 211 isn’t ready to land. Have you seen the runway? BS 211 has overflown the entire runway and is now headed north toward the mountains. I think we’re going to Runway 02. Concerned by what he’s seeing, the supervising controller steps in to assist Flight 211. OK, Bangla Star 211, runway 20, clear to land. Kathmandu air traffic control redirects BS 211 to turn left and land on Runway 20. But the plane isn’t lining up with the runway. Bangla Star 211, turn right. You have the runway to land. Confirm you have the runway in sight. Negative. Where is the runway? Sir, runway. Runway, 3:00. Affirmative. We have runway in sight. Request clear to land, sir. BS 211, clear to land. He’s lining up for the taxiway. BS 211, that is not the runway. Over, that’s not the runway. BS 211, that is not the runway, I say again. The pilots make a last-minute adjustment to try to line up their plane with the runway. But the airplane is headed directly for the control tower instead. The pilots of US-Bangla 211 struggle desperately to get their planes safely on the ground. Banking. [EXPLOSION] US-Bangla 211 bursts into flames 440 meter from the runway. 49 people are dead. A multinational commission is formed to investigate the tragedy. It consists of delegates from Nepal, Bangladesh, and Canada, representing the Canadian manufacturer of the Dash 8 Bombardier. BS 211’s flight data recorder is ready for analysis. Will it reveal if the pilots were in control of their aircraft? These are the inputs the captain made to the control column. And these are the actual movements of the flight controls. Investigators compare the captain’s inputs to the actual movements of the plane’s flight controls. They make an important discovery. They’re identical. The plane was doing everything the captain commanded to do. Let’s see later on the flight. Take a look at these inputs. They’re so extreme. Bank angle. Bank angle. Sink rate. The data shows Captain Sultan was making some severe inputs to his control column near the end of the flight. Bank angle. Bank angle. If there wasn’t a flight-control problem, why was the captain flying the plane so erratically? This is the flight path the plane was supposed to take. To better understand exactly what the plane was doing as it approached the airport, investigators compile the heading data and chart the exact course of the plane. And this is the flight path they actually took. What the data shows is astonishing. It looks like they drifted way off course. And flew loops very close to the mountains. Looks like the trouble started right here, 17 miles out when they arrived at Guras waypoint. Pilots navigate by following a series of waypoints or GPS locations along their flight path. Guras is the last waypoint into Kathmandu Airport. Investigators dig deeper into the data to understand why BS 211 drifted off course at Guras. They programmed the flight management system for a holding pattern of the Guras. Here. The FDR data reveals that the crew pre-programmed a holding pattern about 34 miles before reaching Guras. And when they reached Guras, it looks like they start their holding pattern, right here. But the heading data shows the crew didn’t complete their holding pattern. The question is, why? It’s late afternoon at King Abdulaziz Airport in Jeddah, Saudi Arabia. Dozens of US Air Force KC 135 aircraft prepare for late-night missions during the Gulf War. Major Kevin Sweeney and the crew of US Air Force Flight Whale 05 review the final details of their mission. Dart switches. He’s the commander. Flight start. Sweeney has over 20 years of Air Force flying experience. It’s his job to know the mission and his plane. But the aircraft commander is just like the captain. If you have to make any tough decisions, it’s your responsibility, although it’s very important to take input from the rest of your crew members and listen to them. At 5:25 PM, Whale 05 lifts off from Jeddah. Tonight’s mission takes them along a tanker corridor, an aerial highway for tanker crews heading north from Jeddah. When they reach Waypoint Rita, they’ll turn east before making their final turn to the rendezvous point less than 180 miles from combat zones near the Kuwaiti border. Whale 05 reaches cruising altitude. As they get closer to enemy territory, the pilots reduce radio contact to avoid detection. Moments later– What the– –things go very wrong. [BEEPING] Oh! I got it. It takes just a second for the plane to roll 110 degrees to the left. The crew can’t tell if there’s some kind of malfunction or if they’re under attack. We’re going to lose her. We got to get her level. Just as it seems the plane is in an unrecoverable left bank, it snaps hard to the right. Just when all seems lost, Major Sweeney plays a hunch. Speed brake! Speed brakes are devices on airplane wings designed to increase drag during descent and landing. Sweeney deploys the speed brakes on both wings, hoping it will level the airplane. It’s a procedure Sweeney remembers from his training. It works. Incredibly, the pilots have managed to level the plane. But they’re not out of trouble yet. All right, I have lateral control, but we’re losing altitude. We’ve got fire lights on engines 1 and 2. The pilots discover a problem with the two left engines. Steve! How bad are the fires on engines 1 and 2? Checking. If there’s a fire in the engines, it could lead to disaster. Oh, my God. The only thing I could see was torn sheet metal on the wing where the engines were, and fuel being vented over the top of the wing. They’re not on fire. They’re gone. No fire? Affirmative. No fire. The engines are gone. Roger. US Air Force investigator Ike Stokes tries to understand how wake turbulence nearly destroyed a gigantic tanker aircraft. Separation between the two planes by the book. That’s it. Wind was 85 knots from the West. The wind was blowing enough at altitude to push the wingtip vortices from the preceding airplane into the flight path of the mishap aircraft. They accounted for everything. They were foiled by the direction of the wind. Perfect storm. But Stokes still doesn’t know how this perfect storm ever got a chance to form. When you’re taking off between 90 to 100 airplanes on a daily basis, the arrival and departure from the base is very critical. And in this particular case, the mishap aircraft was parked in such a way that it had to go first and be followed by the second airplane, the one that had further to go. The solution that they came up with was a perfectly responsible solution. And that was, we’re going to take off individually, and then you’re going to pass me. Had the number-two plane been parked to the left of the mishap airplane, the mishap never would have occurred. Investigators finally understand what happened to Whale 05. 05’s level. On a wartime mission over the Saudi Arabian Desert, two KC 135s switched position mid-flight. Altitude hold on. Altitude hold check. The wake turbulence generated by the passing plane is blown into the path of Whale 05. It creates a tremendous force that flips the plane so violently, G-forces rip both engines off the left wing. [GRUNTS] Often heard flying described as hours and hours of sheer boredom, followed by moments of stark, raving terror. In this case, the crew experienced that stark, raving terror. And there’s too much stress! The oscillations almost push the right engines and the aircraft beyond the point of recovery. Speed brake. But with the quick thinking of a seasoned wartime commander, the plane levels off. Coming back to level flight was a true feat of airmanship. Greg, are your NAV instruments working? Yes. INS is functional. Radar is still up. Well, give me a heading back to Jeddah. The coordination of a well trained crew– Nose gear down and locked. Runway in sight. –brings Whale 05 back home safely. Aircraft commander was an excellent airman. But the fact is he had every brain in that airplane working in tandem with him, side by side, to make sure they didn’t miss anything. Brakes. I was very fortunate to fly one of the best crews in the Air Force, in my humble opinion. They did their job in a critical situation. And it wasn’t just me getting back the airplane. It was the team got back the airplane. We did it. Whether it’s a refueling plane in wartime– We did it. –or a routine domestic flight in a 747, the importance of teamwork can never be overstated. It’s 9:00 in the morning as Ansett New Zealand Flight 703 cruises toward the city of Palmerston North, New Zealand. William McGrory is flying to his company’s head office for an early morning meeting. I was working for a plumbing company. I was based in Auckland, and they were based in Palmerston North. As we were on approach– I don’t think the landing gear’s down on the right side. Can you check your window? No, I can’t see the landing gear at all. In the cockpit, Captain Sotheran and his first officer are already troubleshooting the problem. As Flight 703 nears Palmerston North, the pilots follow the procedure for lowering landing gear manually. Airspeed below 140 knots. It’s 140. Karen was sitting right there in front of me, and the next minute– And then, I’ll have a few days off, so I’ll probably head down to– Another crash, and then we seem to be sliding. Eventually, we came to a stop. I guess it just clicked into survival mode and didn’t really matter what was wrong with me. Just, I was alive. Get out of this plane– that was probably the only single thing that was on my mind at the time. Saw a hole in front of on the right-hand side, and saw that as an opportunity to get out. In the airport’s tower, Controller Tony Chapman tries to contact Ansett Flight 703. I don’t know where they are. There’s no signals at all in. They’re off radar. Way out in front of the nose now, which was facing back the way we came, I saw my little briefcase. So I grabbed that and opened it up, knowing full well that my phone was in there. I rang 111, and they said, what emergency do you need? Ambulance, fire, or police? And I said, send the whole bloody lot. We got a plane crash. Send everything. And she immediately changed her tone and said, just hang up, and we’ll get back to you. Did he say where they were? OK, did you get a number? The operator has notified Palmerston North’s tower of McGrory’s call. Perfect. Thank you. [PHONE RINGS] Hello? William, whatever happens, do not hang up the phone. You have to stay on the phone with me. Can you see if you can find some kind of landmark, anything that can help them find where we are? The passenger who had come to help me said, I’ll have a look around. So he went off up the hill, and he had gone quite a distance and quite a while, really. And he came back some time later, and he said, there’s a big holding pen for sheep up the hill further. We’re next to a very large sheep holding pen. Looks like they’re near the Buckley stockyard up by Hall Block Road. With the location of the crash roughly pinpointed, rescuers make their way to the scene. Flight 703 has crashed into a hill 10 miles from the airport at Palmerston North. 15 passengers and both pilots have survived. Tragically, three passengers and flight attendant Karen Gallagher are killed. We were very, very lucky that 17 of us survived. I think so sadly for those that did lose their lives. The hostess and the others that died on that day– was so unnecessary. Investigators need to examine the wreckage of Ansett 703 to determine why the pilots were unable to get their landing gear down. But the muddy terrain is presenting a challenge. It was virtually impossible to get equipment onto that site. It would just slip, and they would get stuck and slide down hills and so on. Vance comes up with a solution. They had a huge helicopter owned by Russians. What we suggested to them that they do is get a big long cable and put the cable through the fuselage. The ribs were in good-enough condition, that they would basically hold the weight of that fuselage. With all the wreckage in a hangar, investigators are able to examine the right landing gear to understand why it didn’t come down. Vance is joined by Jim Donnelly, a maintenance engineer from de Havilland, the Dash 8’s manufacturer. When the landing gear is up, a latch holds a roller on the gear’s leg in the retracted position. When pilots lower the gear, an actuator moves the uplock latch to release the roller, allowing the landing gear to extend. This is probably where the problem was. Latch is definitely showing signs of wear. Over time, the roller wore a small groove into the right-side latch. It was enough to prevent it from sliding into the down position. Here’s another. Landing gear fails to extend. Yeah. It definitely was an issue. Investigators dig through the Dash 8’s history. Yeah. The inset fletcher had its share of problems. And both are Dash 8s by the looks of it. Ansett New Zealand’s Dash 8s had been experiencing landing-gear failures for years. Just the left side. Gotcha. Thanks for that. They replaced the mechanism on the left side. But we’re waiting on PATs for the right. Ansett only replaced the left uplock actuator, because that is where they experienced the majority of their issues. But all of these gear problems were easily dealt with. In every case, the pilots used the alternate method to lower the gear. And they all landed safely. If the gear didn’t lower normally, pilots could pull a handle in the cockpit that manually disengaged the latch, so that the gear could drop into position. The alternate system is 100% reliable. There has never been an issue with the alternate landing gear extension system failing to lower a landing gear. But evidence from the cockpit wreckage reveals the first officer didn’t pull the handle hard enough to release the landing gear. We saw that the handle that is normally pulled was partially pulled. Failing to lower the landing gear is unusual. But it doesn’t explain why the pilots of Ansett Flight 703 slammed into a hill just a few miles from the airport. It’s the first flight of the day for the crew of Propair Flight 420. Today’s flight is a 90-minute hop from Dorval to Peterborough, Ontario. The plane has been in the air for 12 minutes. [BEEPING] What? What is it? Looks like we lost hydraulics. Dorval approach. This is Popair 420. We’ve had dual hydraulic failure. Request clearance to return to Dorval. The metroliner has two hydraulic systems. One controls the flaps, the other the landing gear. Looks like we’re landing without flaps. With no flaps, the pilots can’t reduce their speed without stalling. Then just 30 seconds after losing hydraulics, before they’ve started back to the airport– What’s going on? It wants to roll left. Really? Holding a right. Something’s wrong with the controls. Need to trim half turn to the right. If the plane is rolling in one direction, applying trim avoids the need for continuous pilot inputs. Trimming it right brings the left wing up and levels the plane. As Flight 440 gets halfway back to Dorval– Fire. The left engine is on fire. –an even bigger problem emerges. Left engine shutdown procedure. The pilot’s attempt to extinguish the fire in the left engine. Left power lever. Confirmed left. The captain executes the engine shutdown procedure. Back to idle. Confirm left shutoff lever. Confirmed. Pulling left engine stop lever. Dorval approach Propair 420. Left engine is on fire. We’ve shut it down. Propair 420. I see you are returning to Dorval. I can give you direct to Mirabel. Affirmative, direct to Mirabel. While flight 420 is only 11 minutes from Dorval, they reroute to Montreal’s other airport, Mirabel, which is closer. I see flames now. Flames from the engine nozzle. The situation goes from bad to dire. Fire crews park alongside the runway at Mirabel Airport for the emergency landing of Flight 420. Captain Provencher is struggling to maintain control. Now, he has to lower the landing gear manually, with no guarantee it will work. Gear down now. Gear down. The nose and right wheels have dropped, but one light stays off. Prop air 420 is 20 seconds from touchdown. Rolling left. Not now! They’re just five seconds from being able to touch down when disaster strikes. The plane crashes into a watery ditch next to the runway. Despite the best efforts of rescue crews, no one makes it out of the plane alive. What is it? Looks like we lost hydraulics. Investigators now turn to the cockpit voice recorder of Propair 420 to determine why firefighters and the pilots both reported an engine fire. I’ve got the call I’m halfway to the [? right. ?] I can’t believe it’s taking this much time to hold it straight. Hang on. Control problems just 30 seconds after hydraulic failure. The Dorval is here. They’re barely out of the gate for the hydraulics fail here. Haven’t even begun their turn, and the controls start acting up here. Fire. The left engine’s on fire. Is that a passenger? Fire in the left engine? The passenger report of an engine fire confuses the crew. The engine overheat warning is off. Left engine shutdown procedure. The captain follows the checklist, but it doesn’t solve the problem. I see flames now, flames from the engine nozzle. I don’t have the fire lights. The cockpit voice recording provides investigators with their biggest lead yet. Maybe the fire started in the wheel well. That’s so close to the engine, the crew could have made that mistake. The team discovers that pieces of the left landing gear are burned almost beyond recognition. The team finds that several components of the left side breaks show significant heat damage. Piston housings are melted. Cylinders are blackened. Investigators can finally confirm that an in-flight fire aboard Propair 420 began in the left wheel well, but they still don’t know what started it. This is the left. Brake disk, correct. One component is key. The thing got pretty hot somehow. The grayish-blue color stands out. The landing gear would need a fuel source to ignite. But the wheel well is nowhere near the heavily reinforced tanks. Investigators focus on the tubing inside the left wheel well called the nacelle. Hydraulic line there. Dorval approach. This is Propair 420. We’ve had dual hydraulic failure. Request clearance to return to Dorval. The melted lines would cause the hydraulics to fail– the first problem reported by the crew. So the heat from the brakes melts the line. Hydraulic fluid pours out everywhere. There’s your fire. Right there. It’s a good theory, but they need evidence to back it up. Investigators design a test to determine if hydraulic fluid could ignite when exposed to overheated brakes. They heat the disk to degrees Fahrenheit, the temperature that would have been needed to give the left brake disks their grayish-blue color. Whoa. It’s an early summer morning when UPS Flight 1354 climbs over Kentucky. Tonight’s flight from Louisville, Kentucky to Birmingham, Alabama takes about an hour. The crew is flying the newest version of the Airbus A300, equipped with advanced computers and flight management systems to assist pilots. Notice to airmen– Runway 24 closed. Localizer Runway 18 in use. Landing and departing Runway 18. The runway they were expecting is closed for maintenance. They’ll have to use an alternate. The alternate runway is shorter. With the plane nearly at maximum weight, they’ll have to carefully manage their speed and altitude on approach. Landing on Runway 18 also involves a more demanding approach known as a non-precision approach. In a non-precision approach, pilots pre-program the flight computer to follow a virtual glide path or descent profile to the runway threshold. Five miles from the runway, the captain realizes something’s not right. Unbelievable. Too high. The autopilot hasn’t initiated its final descent to the airport. The captain tries to get the plane back on its programmed glide path. If the plane remains too high this close to the runway, the crew could overshoot it. Autopilot’s off. The captain prepares to fly the plane manually to touchdown. As they get closer to the airport– Did I hit something? Oh, no. Oh, God. The pilots can’t control the plane as it cuts through a small grove. No! Oh, my God! UPS Flight 1354 crashes just one mile short of the runway. Tower, did you see that? Yes. Yes. Airport 12, there’s been a crash. UPS 1354, heavy crash on the hill. Attention, aircraft crash, 3 miles final Runway 18. Rescue crews rush to extinguish the flames of UPS Flight 1354. Despite the plane coming down in a populated area and crossing a highway in Birmingham, Alabama, no one on the ground is injured. Tragically, both pilots are killed. Within hours, the National Transportation Safety Board begins the investigation. Verify the glide path agrees with the approach chart within 1 degree. Did a malfunction in the flight computer used to program the autopilot lead to the crash of Flight 1354? Verify approach. 0.1 degrees. The flight management system, to an airline pilot in an airline operation like this, is very critical. Because it is the automation. It is the typical way of flying a large aircraft. If you have bad data in, that data will cause bad things to happen. OK. Let’s hook this up. Investigators recover the flight computer’s memory card from the wreckage. They prepare to test it for signs of errors or malfunctions. Investigators went to great pains to figure out what exactly was loaded into the flight management computer. It was damaged, so they had to remove the motherboard and place it in a functioning unit and actually read it out. This should tell us if the computer was working. If the flight computer was operational, investigators should be able to download its memory. It’s working. It turned out that there was nothing wrong with the flight management computer. But had there been, that could have been a very important part of the accident sequence. The computer was working. Looks like they programmed it. Final approach is armed for a gradual descent of 3 degrees. Wait a minute. They’ve got two separate destinations programmed. They forgot to clear the conflict. They discovered the crew missed a step in planning their route to Birmingham Airport. They failed to clear a previously programmed destination. It’s a troubling find. You can load a flight plan into it, and then if you deviate from that particular flight plan, the flight management computer doesn’t really know what’s going on and can put out false data. And that’s called a discontinuity. The crew programmed the plane to fly directly to Birmingham Airport. 20 miles out, they needed to clear their flight path and program a specific approach to Runway 18. But the crew didn’t clear the initial plan, which created the discontinuity, a confusion in the system. There was a conflict between where the pilots told the airplane to start the approach and where the computer knew the approach had to start, and that was a flight plan discontinuity. That’s why the autopilot wouldn’t initiate the descent path. The captain basically was chasing the incorrect guidance that the display was telling him by trying to descend as quickly as he could, when there was no reason for it. The team knows the crew didn’t clear the conflict. The question is why. [MUSIC PLAYING]