British 6711 : La Balise Trompeuse

Foreign hello, my friends. One of the problems that pilots sometimes encounter is when they take off from countries where aviation is a highly structured industry. People do their job professionally, there are checks at every level and they have to land in countries where aviation is not taken very seriously. This can cause quite a few disasters. So on March 31, 2003, a British Airways plane took off from Hatro, not far from here, bound for Disabeba with a stopover in Alexandria. There were 75 people on board. It was British Airways Flight 671. After takeoff, it headed for Alexandria in Egypt. It landed some fuel and took off again for the final leg of its flight. It crossed Egypt to the south, flying over the Nile Valley, it crossed Sudan, then arrived in Ethiopia and entered the terminal area of Addis Ababa airport. Clouds covered the airport. Storm cells with cumulonbus. Clouds had been observed in the mountains around the airport. Addis Ababa airport is at an altitude of 7,500ft. It is a very high airport, which poses challenges for arriving and departing aircraft. Because aircraft performance is not very good at these altitudes. A go around is difficult to do. A takeoff is poor performance. Also during the approach, the aircraft go fast. They will fly faster to compensate for being at altitude. The true speed of the aircraft is faster. The rate of descent that it must display to capture and maintain a glide path to a Runway also requires high rates of descent. And all this in the mountains. So when the aircraft arrived at the terminal at Addisaba airport, it was after 11pm so it was pitch black and cloudy and the airport was equipped with a bit of a lack of equipment. So the aircraft wanted to do a vordme approach. The pilots wanted to do a vordme approach on Runway 25 left. We've talked about the NDB many times. In the last videos we've said that the NDB is a very old beacon, one of the first beacons that were invented that tells the pilot, I'm here. And the pilot has a needle that shows him the direction of the beacon. That means that if I take this plane. Let's imagine that this is the NDB beacon. The aircraft is there while the beacon is straight ahead. If the aircraft is there, the beacon is always straight ahead. And here it is always straight ahead. So in fact, one of the difficulties with the NDB beacon is that it goes against the fundamental principle of navigation. Because when you're navigating, what you want to know is where the plane is, not where the beacon is. The NDB indicates where the beacon is. But it doesn't really tell you where the plane is. You have to do calculations because the difference between all these positions is the heading. So you have to do a calculation, including the heading and the bearing relative to the beacon to know where you are. And even then, with random precision. The VOR is a much more precise instrument. The vor, if you like, without going into too much detail, is a beacon that divides the area around the VOR, which divides the space around the VOR into 360 radials. In fact, okay, it's 360 routes. And pilots know exactly which route they're on using the instruments they have on board. Plus, when it's coupled with distance measuring equipment called a dme, it's possible to know the distance relative to the vor. So it's an instrument that allows navigation, but also, as in this case, allows landing. So one of the problems with VOR type approaches is that it doesn't give vertical guidance, it gives lateral guidance. This means that the pilot can capture the axis of the Runway with great precision and know that he is really on the axis of the Runway. And then in terms of descent, he doesn't have extremely precise guidance on the descent. In fact, there is no descent guidance at all, whether it's precise or not. And so he has to follow a certain number of markers, a certain number of milestones, if you like, marked by a distance. And at a certain distance, he has to be at a certain height and so on. Okay? He can't see the Runway, he can't see the city, he can't see the mountains, he can't see anything at all. He really feels like he's in instrument flight conditions. Okay, now there's a detail I'd like to tell you about. This approach is based on the ADZ Alpha Delta Sera vor, which is the Addis Ababa flight, located at the entrance to the Runway. This VOR is a VOR manufactured by Tail. It is a Type C and VOR. There are two types of VORs, C, C and D. The Dvor and the Cvor. The Devore is more modern. It is a Doppler VOR. But we don't really care about Infini, because in an airplane the receiver doesn't make the difference and the pilots don't make the difference and don't see the difference. But I simply want to tell you, because it is important here, that this is a fairly conventional DES4 type of equipment manufactured by Tail. A few days before this fateful flight, the local technicians opened the COVID of the VOR antenna to be able to make an intercom. They did not close the COVID of the vor. In the meantime, there was very heavy rain falling. And what happens B the electronic circuits of the Addis Abeba VOR are flooded. So what do they do? And so he decides to repair it. So he repairs it. He sees that some capacitors, it's vor, they have capacitors as big as tin cans. And in fact, these circuits with the capacitors are replaced. He replaces the circuits and they restart the vor. When you start a vor, what you have to do is calibrate it. That means you have to check that it is emitting the radials, the 360 radials, that it is emitting them in the right direction. This requires there are devices that allow the ground to go to known locations whose route is actually known in relation to the vortex, and then to check that it is correct, or even by an aircraft that circles the area that goes at different distances. And this aircraft has measuring equipment that allows it to check the accuracy of this flight. The tolerance at the aci, the oci, the Civil Aviation Organization, says that if the wind shifts by more than one degree in any direction, it must be stopped. It must be stopped and recalibrated. Okay? This work is not done by local technicians. The technician starts the vortex. He takes a portable receiver, he sees that it works more or less, and then it's good, he validates, and that's it. So this was the day before this flight. The pilots at the start commit to the approach as a reference. So they have this vortex and they also have a beacon, a db, which is also located at the entrance to the Runway. From the start, they had suspicions from the start. They were a little surprised that, for example, when they came in from the north and had to enter the approach pattern, the turn they made seemed to be a little bigger than it should have been. But it was little details like that that struck them. But they had to trust their instruments. They moved away from the Runway. They got to 13,000 nautical miles further than the Runway. So they moved away from the Runway to 13,000 nautical miles. And then they made a left turn, they came back towards the Runway and they began their descent. They descended and during the descent they couldn't see the Runway at a time when they were supposed to see it, and they began to worry. The ADF needle, which is an instrument that captures the DB beacon, tells them that they are not on the axis of the Runway, but they know that there are storm cells and sometimes storm cells tend to distort the adf. So the emission of the NDB beacon is very sensitive to storms. So as soon as there are storms, the degree, the rate of confidence that we have in an NDB decreases. And in any case, if we compare Vorind Blair, all the pilots in the world trust a VOR more than an ndb. At some point, the pilots do not understand anything and they decide to go around, say, it is not possible. We can't see the Runway. It should be there, but we can't see it. They go around, they go back into the holding pattern and they contact the air traffic controller. They ask him if the VOR ADS is working, is it okay with your vor? He answers it. He says, yes, the VOR is working okay. They have to trust him. He's an air traffic controller. He's there, he knows his equipment, he knows his airport. He tells you that the VOR is working. The VOR is working. In fact, not only was the VOR not working, it was completely out of calibration, it was not correct, it was not transmitting correctly. Means certain radials of the routes that were published in the approach charts. And these routes were supposed to guarantee the pilot a safe descent between the mountains. The minimum sector altitude at 10 Aba is 14,000ft. Just to let you know, there are mountains that rise very high compared to the airport. The airport is at 7,500 and there are mountains that rise very, very high around it. So the radial. Some radials issued by the Addis above board ensure the pilot that if you are on this route, you can descend and you are outside the mountains. They are on your left or your right, but you can descend on this radial. However, some of these radials considered safe and published on the maps ended up in the mountains. That's the problem. Uh, when the controller tells them, yes, it works, it means, yes, it is on. That's all. That doesn't mean that it is calibrated or that it guarantees anything. In addition, there was a cable. There was a cable that went from the VORD to the control tower and that went to equipment that regularly checked the health of the wind. This cable had been cut by a wind turbine for months and no one had repaired it. That's where the situation stood. Given the air traffic controller's assurances, Pilots put yourself in their shoes. They say to themselves, okay, we're going to try again. We're going to do another approach and this time we're going to do it with even more precision than before. So they return to the approach, move away from the Runway to the point marked by a distance of 13,000 nautical miles. @ this point, they already have the plane ready for landing. Flaps down, landing gear down. When a plane starts landing, there is a device called gpws. It is a proximity warning device or ground proximity alert device that warns pilots. It has many modes, it has many algorithms that warn pilots if they lose altitude after takeoff or if they approach the ground too quickly while the gear is retracted. If it approaches a terrain, it gives them information early enough to avoid this type of accident. This prevents cfit, which is controlled flight into terrain. However, when the aircraft is in landing configuration, the EGPWS switches to a mode called mode 2B. And this mode is less sensitive, of course, because we know that the aircraft wants to land, so the aircraft wants to go towards the ground. So some of the EGPWS alerts are deactivated. In fact, when the aircraft is in landing configuration, the pilots arrive at point D13, turn left dual loop and resume their descent. They start to descend. It's dark, they're in the clouds, they can't see anything at all. They trust their instruments, they trust the moving map that appears on their navigation display, on the navigation screens. And then suddenly, when they are about 8km from the Runway, there is a synthetic channel on the radio altimeter that announces 1,000ft. That's a bit low. It's a bit low, but hey, they know that there is terrain around, okay, they say, we're monitoring the terrain, it's coming. A few seconds later, the channel announces 400ft. So it's not that the plane has fallen 400ft, 600ft between the two notifications, it's simply that it is heading towards the terrain and the terrain is rising very quickly, because here we're not talking about altitude, we're talking about radio altimeter height. So when they hear 400ft, the 400 foot signal, the pilots are a little shaken. It's very, very low during an approach, especially since the airport is miles away and they can't see it, are completely in the clouds. The pilot, the captain, will hesitate for two or three seconds and then what he's going to do is take the stick that he has on the side, pull it all the way back and push the throttle all the way. He's going to put the throttle back on, the plane pitches up to 15 degrees and it starts to climb. So one of the advantages, let's say, of the Airbus over the Boeing in these situations is that the Airbus has envelope protection that prevents it from stalling. Because let's imagine. Remember the video about the American Airlines 965 that crashed in the mountains in Cali, Colombia? The pilots had two problems. On the one hand, they had to climb as quickly as possible, but at the same time, they had to keep the plane from stalling. Because if a pilot, in a panic or in the heat of the moment, pulls the stick too hard on a Boeing while the Boeing is on approach, he doesn't have much energy, he's relatively slow. He doesn't have much speed margin, so not much angle of attack margin. If he pulls the stick too hard, he pitches the plane up too much. What's going to happen? He's going to make it stall. Plus, we're on approach, we're not very high. When you stall an airliner at low altitude, you can't recover it. The only thing you can recover when you stall at low altitude is the black box. You can't recover the plane. So on Boeing, pilots in similar situations, in these situations have to be careful. Not only do you have to pull hard until you almost reach the stick shaker, but at the same time, you shouldn't be too hard on the plane because you can make it stall. On Arbus, they don't have that problem. They can pull hard, put the windows all the way down, and then let the plane sort itself out. That means the plane will go to the maximum angle of attack it can afford and won't go beyond that and won't stall. So that allows you to extract the maximum climb performance instead of hesitating. Because these hesitations, like saying, I'm afraid of making it stall, often lead to pilots following trajectories that aren't the best possible. There were even more possibilities, but the pilot didn't dare to be too timid. The plane at the start, the pilots didn't realize it, didn't see it. But later, the investigation revealed that they were 17 meters from hitting the mountain. They were at 56ft. The lowest they reached was 56ft, 17 meters from the edge of the mountain. So the plane took off again. The pilots returned to the holding pattern, and then they decided, they informed the controller. They said, listen, your ad system is completely rotten. Uh, it's transmitting incorrectly. We almost hit the mountain and we're not going to land here. And what did they do? They went to their alternate airport, which was in the Djibouti flight plan. So they went to Djibouti. They arrived around 1am they landed, everyone went to the hotel. You know, the night brings counseling. They slept, and the next day they decided to return to Sabah. It was daylight. The weather had improved, visual flight conditions. So they arrived in visual flight conditions. They return to Addisaba. They line up facing the Runway, on the Runway axis, and there they are in visual. So they see the Runway, they see the mountains, they see everything. They start to descend and then what do they do? They put the VOR just for the sake of awareness. They connect to the VOR of Addis Aba. And then what does they see there? They realize that depending on where they are, the VOR is offset by 22 to 30 degrees. It's huge, 22 to 30 degrees at the top. Because when a VOR is not well calibrated, the error it will make is not systematic. It is an error that can change depending on where you are. It is an error that is not the same. So this VOR was in fact broadcasting anything. And the controller had been informed the day before, and he did absolutely nothing. And the VOR continued to operate all night, night in imc. That means that if another plane had arrived after the British Airways and those pilots had perhaps a slightly less modern plane than the Airbus 320, or if those pilots had perhaps been less attentive, they could have taken a huge hit. Okay, so the pilots returned to the United Kingdom later, and it was the British authorities who had to write an urgent message to the controllers and the services that deal with these beacons in Ethiopia to tell them that this plane was doing something crazy and it must have been fixed or not, or no one knows anything. In fact. So also something else that was revealed following this affair, which was treated as an accident, even though there were no fatalities, there was no damage to equipment, it was treated as an accident because when a plane arrives at night in the middle of nowhere, 17 meters from a mountain, it's not a problem. It could have ended very, very badly. In fact, they found a problem on this Airbus. There was no gps. There were IRS or inertial measurement units. There were three of them, which calculate a position and then produce a sort of average. These IRS shift over time. That means they are gyroscopes. So these gyroscopes will initially stay aligned before taking off, but during the flight they will drift. They will drift very slowly. What does the FMS system do? The helm computer, if you like, when it picks up beacons, it will use the beacons to correct the information it has from the gyroscopes. That in fact, when it was navigating over Sudan. The system indicated that the degree of navigation accuracy was low because it had no beacon to confirm the position of the irs, the inertial systems. As soon as it captured the Addis Aba vor, it used the Atisaba VOR to correct the positions of the inertial navigation systems. In fact, it was to falsify them. That is to say, the map that was displayed on the pilot's navigation display had also slipped. Okay, so the map had slipped. It was wrong. That means the location where the Runway was displayed and the location where the plane was on the Airbus map was wrong. The Airbus was above the mountains while the mountains were displayed on the map, and they were still to the right. You see, that means that we have a beacon that has messed up. She messed up the approach. She messed up the aircraft's irs, the charts. She also messed up the egpws, because the ground proximity warning system also relies on the position calculated by the FMS to know where the aircraft is and issue alerts. So that system was also distorted. So what shocked investigators in this story is that in aviation, we always try to avoid. We always want redundancy. We always try to avoid situations where we have a single failure that can cascade and create others and which cannot be recovered otherwise. Whereas, in fact, we are in a case here of single point of failure. Here it means a single point of failure that will cascade, creating chain failures throughout the entire system. And that was particularly revealing. So from that point on, I think they must have stopped sending planes without GPS to places like this. Anyway, there you have it. They flew right past. Everyone got back safe and sound. And then, fortunately, the pilots had the reflex to go around, because, in fact, the pilot did go around, and it was only a few seconds later that the ground proximity warning went off. That means that the human pilot anticipated a reaction. His algorithm in his head acted several seconds before the plane's electronic system. That's it, my friends. I'll leave you with this story that ends well, and I'll see you next time.