Where is the RUNWAY?! Unraveling the Mystery of Turkish Cargo flight 6491

(suspense music) - This Boeing 747 is on final approach to Manas International Airport in Kyrgyzstan. The plane is working perfectly and the pilots are following the ILS approach preparing to execute a normal autoland due to very thick fog, but are they really? (suspense music) A very little-known feature within the 747's autopilot logic is about to test these two pilots very soon. Stay tuned. (jet engines whirring) On January 15, 2017, a

flight

crew of four pilots working for ACT Airlines had been planned to operate two

flight

s on a wet lease on behalf of Turkish Cargo. The first flight with call sign Turkish Cargo

6491

would start in Hong Kong and then head to Manas Airport in Bishkek, Kyrgyzstan,

where

there would be a crew change and then the last flight would continue to Atatürk Airport in Istanbul, Turkey.

Before the flight, the pilots had enjoyed a fairly long rest period of 69 hours in Hong Kong which they had spent in a hotel in the city, so they were well rested when they found themselves on the bus heading to the airport. . in the afternoon of the scheduled flights. Although there were four pilots scheduled for this task, only two of them were to operate the first flight. The remaining two traveled passively during the first leg and would then take over on the second flight to Istanbul. The captain who was to operate the first flight was a 59-year-old Turkish citizen with just over 10,800 hours of total time.

He had only been operating the Boeing 747-400 for approximately 820 hours at the time of this flight and all of those hours he had flown as pilot in command. His prior experience flying civilian aircraft was in the Turkish Air Force, as was the colleague he was about to share the cockpit with, a 59-year-old first officer. The first officer was also quite experienced, with about 5,900 hours of total time and 1,758 of them flown on the 747. The pilots had flown together before, so they knew each other well enough to jump right into pre-flight planning. as soon as he arrived at the airport.

The plane they were going to fly was a 14-year-old Boeing 747-412 freighter, with no technical problems reported. The only items mentioned in the technical log were a few dents here and there and they were all within safe limits for operation. Once the captain verified this, the pilots turned their attention to the weather, which looked like it was going to be a little more complicated for the first flight. The weather en route was good, but a cold front had passed over the Bishkek area, bringing with it very cold air and freezing frontal fog. This fog was forecast to reduce visibility to 200 meters with vertical visibility of only 100 feet at its destination overnight and at approximately the time of its expected arrival.

But it was forecast to start improving around 6am the next day. Manas Airport was equipped with a Category 2 approach system. I will explain what that is soon, but even with that, visibility was so poor that the airport could not be used as a planned destination. Fortunately, the two alternatives, Astana and Karaganda, were considerably better and had mainly fog and light snowfall, so the flight was still legal, although both pilots knew that the approach to their destination could become quite challenging. Anyway, that didn't really matter. Pilots often have to start flights to a destination in bad weather, as long as we have at least two substitutes who are working.

In the worst case scenario, this will mean that we will have to divert, but in the vast majority of cases, a Category 2 ILS will get us so close to the

runway

that we will be able to land safely even in very dense areas. fog. So the pilots continued checking the NOTAMs, deciding on their departure fuel of 96,640 kilos for their six-hour trip to Manas and once they were done, they walked to the plane and began preparing it for departure. Now that the arrival weather was expected to be so marginal, the captain decided that he would be the pilot of the flight, since that is what the ACT Airlines manual dictated.

Now this is a little different than the way we, at my airline, operate in a climate like this. We always choose the first officer to be the flying pilot if we see that there is a possibility of a Cat 2 or Cat 3 approach. And the reason for this is that we fly these approaches as monitored approaches, which means that the The first officer will operate the aircraft to minimums and also through a possible missed approach, but the captain will take over controls if sufficient visual references can be seen at the minimums. This is a very good way to use the crew.

It also means that the captain can look out for the approach lights and therefore will not have to switch from scanning the instruments to external references if a landing is to be made. But for some reason, this is not the way ACT Airlines did it. Instead, the captain would fly on approach waiting for the first officer to tell him if he saw something outside, and if he did, the captain would then switch from his instruments, start looking outside, and decide whether or not they could land before reaching the minima. This technique will end up having serious consequences in the story, but we'll get to that.

As the pilots prepared the plane for departure, there was some delay in loading its

cargo

into the main

cargo

hold. This meant that when the pilots were finally able to call for thrust and start at time 1912, they were delayed by more than two hours. This, in turn, would mean that they would now have to operate until at least around 02:00 in the morning, approaching the window of the circadian minimum. Taxiing, takeoff and initial climb from Hong Kong were uneventful. The captain engaged autopilot at about 400 feet and then continued his climb to his cruising altitude of flight level 340, which is about 34,000 feet.

The flight then proceeded normally until approximately two hours before landing, when the captain handed the controls to the co-pilot, began preparing for the approach, and began preparing for his approach briefing. And this focus briefing and what the captain decided to include but also leave out will prove very important to what's about to happen and I'll tell you all about that after this. While working on the script for today's video, I was trying to find some news articles and accident reports, but blocked websites were slowing me down, but then I remembered that I could use today's sponsor, NordVPN, to unblock those articles. since Nord services are specifically designed to break those types of digital borders just like that. (finger snap) And Nord does more than just help digital access.

They also function as a kind of shield in the digital world. Its cutting-edge cybersecurity features—which, by the way, go far beyond a regular VPN—are designed to help you avoid hackers, malware, and those pesky phishing attempts. So whether you simply want help accessing more international research material, watching your favorite shows abroad, or simply keeping your data safer, NordVPN has your back. Now, if you click the link below, which is nordvpn.com/pilot, you will support my work and unlock four months free every time you sign up for a two-year plan. Also remember that you can try NordVPN risk-free thanks to its 30-day money-back guarantee.

Thanks, Nord. Now let's continue the story. So, unfortunately, the cockpit voice recorder started recording a little after the captain had already started his briefing, but from the content of it, investigators were still able to deduce that he most likely did not discuss the different distances versus altitude controls indicated in the ILS. approach chart. The most important of these would have been the final approach point

where

the last aircraft needed to intercept the ILS glide slope, as long as they maintained the correct altitude. That point was located at a distance of 3.2 nautical miles from a VOR called MANAS and at an altitude of 3,400 feet, but since the two navigation receivers were to be tuned to the ILS frequency that was different from the VOR, it is not certain if this VOR station was tuned during the actual approach.

In any case, those same distances were clearly marked in the navigation database that the pilot had programmed into his flight management computer so that they could be seen on his navigation screen, but as I said, this was never mentioned in the report. Now, when the captain began his report, they were still too far from the airport to receive the actual weather, but the captain based his report on the weather forecast that indicated that

runway

26 would likely be in use. His briefing included the expected weather that was fog, the NOTAMs that might affect them, and the type of approach they were going to fly, in this case, a fully automated Category 2 ILS approach followed by an automatic landing.

He also talked about the minimum sector altitudes for terrain clearance and the vertical profile for the approach, here he did not talk about the control heights or what minimums they were going to use, which in this case was 100 feet of radial altitude. . He went on to explain the missed approach point and what to do in case they had to fly the missed approach procedure and in addition to these points, which are standard in any approach briefing, the captain also explained the differences in procedures between flying a fully automatic focus plane like they were going to do and a normal focus.

This included how it would handle the autopilot system and what to do if certain faults arose. Now the final report contains the cockpit conversation translated from Turkish to English, which means it is not perfect, but when it came time to discuss the different faults that could arise, the captain said: "If I see the runway lights , I will call: 'Landing'". ' and if I don't, I will call 'Turn around' and also if any emergency occurs above and below 1000 feet, let's say we have a no landing autopilot three or no landing above 1000 feet or if we have autopilot or no autoland, we will execute a go-around." He then continued, "If the autopilot disengages, we will attempt to reengage it.

Um, if it happens below 1000 feet above the landing zone, we'll turn around again. And if we get an ILS deviation below 1000 feet, we will do a loop due to the ILS deviation." That might have sounded a little complicated and, like I said, probably had to do with the translation, but it clearly indicates that the captain was planning to go around for almost any problem once the plane had passed below 1000 feet on approach. Remember that at this point, it's probably worth explaining a little about the approach these pilots were on as well. about to perform. An ILS approach is by far the most common type of approach we use for instrument arrivals.

It works by using two sets of antennas on the ground, the localizer and the glide slope, sending out lobes of radio signals. of two different frequencies that are then detected by the instruments on board the plane. Now, depending on how those lobes interfere with each other, the instruments will be able to detect whether the center line of the localizer is to the right or left of the aircraft or if it is. glide slope is above or below. This type of approach is extremely precise and, in the case of a standard Category 1 ILS, can guide an aircraft to a height of about 200 feet above the ground.

From that height, the pilot flying must see enough of the approach environment and runway to be able to safely maneuver the aircraft to a manual landing, but there are also Category 2 and Category 3 ILS approaches. These work essentially the same way. , but they are only used in low visibility conditions, such as fog, and have quite a few additional requirements. These include greater separation between aircraft taking off and landing, a larger area around antennas that must be protected, and also increased backup electrical power requirements at the airport. Now, when pilots fly Category 2 or 3 ILS approaches, they almost always do so using the airplane's autoland feature.

This means that the aircraft will handle the actual landing and the pilots will only monitor its performance and the pilot in charge will take over control at some point after landing. That, in turn, means that both the aircraft pilots and the airport must be certified to operate these types of approaches and we, the pilots, are verified for our ability to do so normally every six months. Now, to make it even more complicated, but adding a super important detail to what is about to happen, theAircraft that are certified to fly Category 2 and 3 approaches may do so as either an operational failure or a passive failure.

If the aircraft's autopilot system is failure passive as it is on the Boeing 737 fleet I fly, the two autopilots will independently engage with the ILS and begin flying the approach to the runway, but if one of those autopilots suddenly started telling the aircraft that it should, for example, initiate a turn but the other autopilot wouldn't indicate the same, well then the system wouldn't know which of the autopilots was correct and all the system would simply disconnect. And with the autopilots disabled, an immediate missed approach will have to be carried out. But if the aircraft fails, the internal components will allow the aircraft to continue flying on approach even if a failure is detected.

In the case of the Boeing 747-400, it was equipped with three autopilots that allowed it to be operational in the event of a failure and therefore continue flying in an automatic approach even if certain failures occurred. The captain then finished his approach briefing and shortly thereafter, the airplane began to approach its planned maximum descent point. The pilots completed the descent checklist and at that moment, 51 minutes after midnight, the first officer called Bishkek Control and requested his first descent. They were initially cleared to descend to flight level 220, which is about 22,000 feet and at that point, the flight was about 131 miles east of MANAS VOR, over an area with very high mountains.

The captain initiated the descent using LNAV and VNAV autopilot modes, meaning the aircraft was now following its pre-programmed trajectory toward the airport. But like I said, there were many high mountains below them, so they had been assigned the 22,000 foot restriction to keep them away from that terrain. As they descended, the air traffic controller called the crew to update them with the latest weather information for their destination. He told them that runway 26 was in use as planned and that the runway's visual range, which is a way of measuring visibility, was 400 meters at the start of the runway and 325 meters at the center and end. .

On top of that, vertical visibility was reported to be 130 feet. So this wasn't great, but it was largely in line with what the pilots already expected and also exceeded the minimum visibility of 350 meters in the landing zone required to initiate the approach. About seven minutes into the descent, the aircraft stabilized at flight level 220 and the autopilot entered altitude hold. The flight management computer almost always calculates with a continuous idle descent from the start of the descent, so this leveling out meant that they were now becoming increasingly higher in profile, which is not uncommon in areas of high terrain.

The pilots were flying in clear air at this point, so they could clearly see the mountains peeking through the clouds beneath which they also commented. After a while, the controller called them again and repeated the same visibility and weather information they had received before. The first officer confirmed that he had received the information the first time and, following the captain's instructions, requested to continue descending, but the controller simply told them to keep the flight at 220 until further notice because there were still mountains below them. At this point, the cockpit voice recorder began to pick up slight agitation on the captain's part.

He was probably looking down at his browsing screen and saw that he was moving higher and higher in the profile and he didn't like that. He told the first officer that at 66 miles from the airport, they would absolutely need a further descent. That corresponds to three times your altitude, which is a common way we use to manually calculate how many miles of runway we need for descent. Now one trick you can use if you find yourself in a situation like this is to start slowing down and that's because the total energy state of the plane is a combination of the potential energy which is the altitude and the kinetic energy which is the altitude.

It's your speed. So if you're reaching a high altitude, then you can start slowing down to compensate for that a little bit and when later on you get more descent clearance, you can just put the nose down and let the airplane pick up that speed and hopefully you'll come back. to be close to your profile, but the captain didn't do this and to be completely honest, at this point he still had many miles of runway left, so there would be plenty of other ways to fix it. . He now told the first officer to request descent again and this time they were authorized to descend to flight level 180 and maintain that level until at least an RNAV point called RAXAT.

Now, according to his letter, the minimum altitude to maintain above that point, to clear the terrain, was flight level 170, so this irritated the captain even more, who thought that the controller was keeping them high on purpose, but In reality, the reason the controller had given them an extra 1,000 feet here was because the temperature in the area was very cold. Cold air will take up less volume than warm air, so when the temperature is very low, an airplane will be closer to the ground than in warmer temperatures. This means that the minimum terrain clearance altitudes must be corrected for this temperature, so that is exactly what the controller was actually doing.

When the aircraft stabilized again, the area controller told the crew to switch to the Bishkek approach controller and the crew quickly did so. But as they changed the frequency, both of them also muttered some underhanded comments about that previous controller. So why is the fact that they did that so important? Well, because it shows that the pilots, especially the captain, were probably starting to feel some stress here, perhaps from the flight delay or from staying above profile and possibly also due to some initial fatigue. The problem is that this type of mentality can lead to something we know as "getting there," which basically means that pilots will be more lucid than they should and therefore potentially miss important clues that should lead them to make decisions. . other decisions.

Anyway, once the first officer checked in with the approach controller, they had already flown over RAXAT, so they were now cleared to descend to flight level 60 and proceed via TOKPA 1 arrival for the ILS approach, track 26. This was exactly what they had. planned and the approach chart showed the minimum altitude of Tokpa flight level 60. But again, this authorization seemed to irritate the captain, who muttered, "They left us up high again!" At this point, the plane was flying significantly higher than it should have, but it was now cleared to descend 12,000 feet without restrictions, so here the captain had a great opportunity to return to that profile.

However, that would require him to react quickly and increase the resistance by immediately selecting the speed brakes and possibly also increasing the speed to achieve that descent speed, but he did none of this. Instead, he initially simply selected the gear shift and reduced thrust to idle, keeping speed at around 260 knots. He also told the first officer that he planned to be flight level 60 by TOKPA. Now TOKPA was only 27 nautical miles away at this point and as I told you, they had about 12,000 feet to descend. So with the formula we talked about before, three times the descent required will give you the runway miles, you can see that 12 times three is actually 36.

This meant that 27 miles was too close to achieve what I planned to do at unless you took some immediate steps to correct it. This is exactly the type of calculations we pilots have to constantly make to ensure we have our descent planning under control every time we fly. The captain must have understood this little by little, because he soon began to increase speed to 280 knots, but it was still too little. The plane now entered a cloud layer that extended to the ground. The outside temperature was well below freezing so the pilots now had to turn on the engine and wing de-icing and this also had a negative effect on their descent performance as the engines had to rev up a bit. little to provide the extra air needed for the anti-icing systems and that meant that the rate of descent now slowed further.

The captain continued to gradually increase speed and at 01:08:20, the first officer reported that the ILS frequencies were already tuned and identified, meaning that the navigation system was already configured and ready for the next approach. The captain continued trying to get back on profile, but he still wasn't using the speed brakes on him. NOTE Instead, he switched the autopilot to vertical speed mode and increased it to 2,400 feet per minute. At the same time, he decreased the selected speed to 280 knots, but since the plane, in vertical speed mode, will prioritize vertical speed as the name suggests, the speed continued to increase to 317 knots.

Now below 10,000 feet the speed restriction almost everywhere is 250 knots and that was also the SOP at ACT Airlines but the captain now told the first officer that he would ignore this and maintain the subsequent speed for below 10,000 feet, at least at first. . Now, in controlled airspace, you can get ATC approval to do this and it's no big deal, but in this case, it was a sign that the plane was not in a good power state and this would become a problem. growing like now. He began to get closer and closer to the approach. So what could they have done instead?

Well, here are several ways to deal with this. As I mentioned before, using the brakes early at high speed can often help if you're far away, but as you start to get closer to the approach you often have to do the opposite, which is slow the plane down. . Now, if you do that, it will momentarily put you even higher above the profile, but it will also allow you to select flaps and even possibly extend the landing gear to increase overall drag, although using the landing gear is a bit extreme. But with the flaps and speed brakes selected at a lower speed, your vertical speed may not be as high but your descent per nautical mile will be greater, which is actually what you need as you get closer.

Now after you have tried this, if you still see that it is not enough, then the only right thing to do is to ask for a holding pattern or delay vectors. It can be irritating and set you back a few minutes, but it's always worth it, trust me. But these two pilots kept pushing and when they passed 12,200 feet, the captain finally began to extend his speed brakes. About a minute later, 250 knots was also selected on the MCP and the descent mode returned to the level change. This meant that the plane now began to slow down after all, but in doing so, it also reduced its vertical speed.

At that point, the approach controller called them with a weather update for the crew. The RVR had now dropped to 300 meters in the landing zone, which was less than the 350 meters they needed, but they were still allowed to start the approach and continue to 1000 feet, in the hope that the RVR would improve and, if No, obviously, then I would have to go around. Now this arrangement meant that when the approach controller asked them if they were still happy to continue, they said yes. If they had made the decision to stop here, this story would have had a very different ending.

At 01:11:18 the aircraft passed TOKPA point at an altitude of 9,200 feet. That's 3,200 feet higher than the captain had planned. A few seconds later, an autopilot caution announcement went off. Now this had to do with a time setting problem in one of the flight control computers and it went away quickly, but no crew member mentioned it. This glitch will have no bearing on what's about to happen, but the fact that the crew didn't notice or notice it could also be a sign that the pilots were starting to get a little distracted. The controller now cleared the crew for the ILS approach and asked them to report when the localizer was established.

He also told them that the transition level was flight level 60 and below that level they would need to fly with an altimeter of QNH 1023 hectopascal. This was read by the co-pilot and both pilots changed altimeters. They were still at 8,300 feet at this point with only 12.5 nautical miles to go—that's about half the miles of runway you'd actually need. So this situation was starting to getreally bad. The speed had now been reduced to 250 knots, so the captain asked that one of the flaps be selected. He then reduced the speed further to 240 knots and asked for five flaps and this was a really reasonable thing to do, as I explained before.

He told the first officer to request a further descent, and when he did, they were told to descend to 3,400 feet, which was the lowest altitude they could reach before capturing the glide slope. The captain selected that new altitude on the mode control panel and then asked the copilot to begin going through the approach checklist. At 01:13:28 the aircraft flew over a guide point located at radius 090° and eight nautical miles from the VOR MANAS. This point was the last point on arrival before turning toward the localizer and was designed to pass at a minimum of 4,400 feet. The plane passed that point at 6,500 feet, 2,100 feet above minimum and at a speed of 220 knots.

The captain asked for flaps 10 to be selected and when the first officer came up to select it, he also said, "Speed ​​controls, flaps 10, we might end up a little high here and we have speed too," clearly showing some discomfort with the situation. The captain did not respond to this indication from the first officer and the first officer did not mention his concerns again after this. The aircraft now continued toward the localizer and the captain kept the airbrakes extended although it was not recommended above flaps five to avoid shaking. This shows that the captain was probably well aware of the problematic height, but he still did not extend the landing gear, which could possibly have helped in this case.

Instead, he armed the localizer mode, preparing the autopilot to capture it and at time 1:14:05, the localizer beam was indeed captured and the aircraft began to turn towards the runway. They were now only 6.1 nautical miles from landing, with an altitude of around 5,700 feet and a speed of 200 knots, still desperately high. The first officer shouted, "Locator capture," to which the captain responded, "Approach mode selected." This meant that the autopilot would now be ready to capture the glideslope as well, but the problem was that the three-degree glideslope they were actually aiming for was now located well below them. It is now possible to capture a glide slope from above, but you must be very careful if you are going to do so.

Due to the nature of how radio signals are transmitted, they will often bounce off the ground as well as being transmitted in the intended direction and this will create false glide tracks at regular intervals. The most common are set at 6° and 9° from the transmitter, extending above the correct glideslope and the autopilot will not be able to distinguish between the correct glideslope and a false one. Because of that, we typically never engage approach mode unless we are capturing the glideslope from below or are within a point of the glideslope from above and the altitude makes sense. Anyway, the plane was now moving towards the runway still well above glide.

The autopilot was operating with the captured localizer, giving it lateral orientation and level change, descending it to 3,400 feet as authorized. When they passed 5.4 miles from the VOR, the captain finally asked them to slow down. He now only had 2.2 miles to lose almost 2,000 feet to reach the glide slope intercept point at 3.2 miles at the correct altitude. Speed ​​was reduced to 190 knots, flaps 20 were selected and the airbrakes were finally retracted. The GPWS has now announced... - 2,500. - It showed that the radio altimeter had begun to detect the ground beneath them and when he heard that, the first officer responded: "Check." But the captain did not respond even though low visibility procedures required it.

This was yet another sign that he was possibly now becoming overloaded with tasks. Now remember we talked earlier about getting there? Well, this was an excellent example of that. The captain was now focusing all of his attention on getting the airplane onto the glide slope when the much easier decision would have been to simply tell the controller that they were too high, level the airplane at the missed approach altitude, and then just fly away. the failed approach. That would have taken maybe a few more minutes and would have been a little tricky, no doubt, but the plane would have been in a much better position and they could have done a perfectly normal, relaxed ILS after that.

They had plenty of fuel so that wasn't a problem and they were the only plane in the area. It is exactly in situations like this that pilot monitoring plays such an important role in being that voice of calm and sanity that advocates for decisions like this to be made, but unfortunately they did not happen. Instead, an argument ensued between the two pilots that will have very dire consequences. The captain said, "Look, you can actually see the ground down there," probably seeing an opening in the fog below them. The first officer responded, "Well, I'm monitoring the instruments," as he should be, but to that the captain responded, "Instruments?

I monitor the instruments. Look out." And the first officer responded in agreement. Now, I can't stress enough how important that discussion was. The captain probably said this because he wanted to get an early indication of any runway or approach lights in sight so he could make the decision to land, but what he was effectively doing was removing the pilot's monitoring role from his monitoring role and leaving him looking ahead. outside. in the dark fog. Anyway, the plane had reached a distance of 3.2 miles from the VOR, which is where the glide slope needed to be captured in order to continue the approach, but they were still at 4,000 feet, that's a good 600 feet down. above the glideslope and that meant the glideslope indicator was showing full downward deviation and probably wasn't even moving.

As I mentioned before, mandatory checkpoints such as final approach correction were supposed to be announced verbally and verified towards altitude, but that had not been covered in the previous approach briefing, so it was now passed through completely high. Instead, the plane simply continued descending toward 3,400 feet, now with little to no chance of being able to capture that glide slope, at least not the correct one. Aside from not being able to fly the approach accurately, this still wasn't too bad or dangerous at this point. Had they recognized their situation, they could have abandoned the approach, made one turn, and then proceeded to another.

But since they didn't, things would begin to take a very sinister turn. The captain was aware that he was flying above the profile, but he put the blame almost entirely on the controllers and now began shouting over them and even insulting them. The first officer tried to calm him down by pointing out that nothing had really happened here. The captain simply muttered in response and then called for flaps 25 to be selected as the plane continued its gradual descent, getting closer and closer to the runway. At this point, the pilots had forgotten to report that they had captured the localizer, so the approach controller called them and asked them to confirm that they were set.

The co-pilot responded and confirmed that they were because the localizer had been captured but did not mention the glide slope. This confirmation caused the approach controller to ask the crew to change to the tower controller's frequency and never asked them to confirm that they were fully established. Now you might be wondering why the controller, at least at some point during arrival, didn't ask the crew about their high altitude or try to warn them. Well, it turns out that although the controller could see the aircraft's altitude on their radar screens, it was his responsibility to make sure they didn't descend too low or cross any minimum altitude.

The descent profile that the aircraft maintained above those minimum altitudes was left entirely in the hands of the pilots. The first officer now checked in at the tower, who cleared Turkish Cargo 6941 to land on runway 26. The surface wind was calm and the RVR was now at 400 meters, which was above its minimum visibility, meaning the crew could now legally continue the approach. But the problem was, of course, that without having captured the glide slope, the plane now stabilized at 3,400 feet still with the glide slope armed and ready to capture. They passed the outer beacon which was another possible point to verify the correct altitude without anyone saying anything.

Instead, at 01:15:50, the aircraft was still maintaining level flight at 3,400 feet with only 0.4 nautical miles to reach the VOR. Here, while the co-pilot responded to the landing clearance, the captain requested flaps 30, which were the intended landing flaps. Since the first officer didn't respond immediately, the captain probably reached out and selected it himself, and as he did so, the glideslope indicator suddenly started moving. Now remember how I told you about false glide runways? Yes, the airplane had flown close enough to the glide slope antenna to be within range of the false 9° glide slope. The autopilot quickly caught on and began to descend the plane at a vertical speed of about 1,400 feet per minute.

If the pilots had monitored their distance here, they would have realized the absurdity of this capture, but remember, both navigation frequencies were set to the ILS, so the distance from the DME probably wouldn't even be displayed, but in any case case, if they had done it. If you've looked at your navigation screens, you should have seen that you were practically already on the runway, and this is where the design of the Boeing 747's failed autopilot operating system begins to play a role in this story. Approximately six seconds after glide slope capture, the LAND 3 announcement was displayed, meaning that all three autopilots were now ready to fly the planned autoland.

The pilots saw this and the first officer shouted, "Glide slope captured," to which the captain responded, "Check, 4,400 feet set when you set the missed approach altitude on the MCP. At this stage, the pilot should have completed the landing checklist, but for some reason this was never done. The plane was now flying over the runway below and the glide slope antenna next to it, but the ILS signal was still being received. everywhere with deviations. well above one point, which is the maximum deviation allowed in an ILS under normal circumstances. But, interestingly, it seems that this glide slope indication was not even being monitored by anyone at this time.

The first officer had been told to look outside, so that was probably what he was doing and the captain could have concentrated on the aircraft's descent performance, which interestingly was almost perfect. Normally, an aircraft attempting to follow a glide slope that moves all around would be pitching. and down like a wild horse, but this plane was descending nice and smooth, along something that looked and felt more or less like a three-degree glide slope. So how was that possible? Well, remember I told you about the operating system failure on the 747, the system that allowed the approach to continue even if a failure was detected?

Well, it turns out that this system also had a built-in function for unreliable glide slope signals. Then, should the glide slope signal momentarily disappear, the system would return to flight with a calculated inertial trajectory of three degrees, based on the aircraft's inertial reference system. This route would then be followed until the glide slope signal was restored or until an automatic landing was reached. And ads for LAND 3 or LAND 2 would continue to appear everywhere. But of course, this wouldn't happen without any precautions. Then, approximately 15 seconds after glide slope capture, as the airplane descended to about 1,000 feet above the ground, an EICAS autopilot warning message appeared along with two master caution lights.

The FMA glide slope warning also passed through an amber line, the flight director bars were eliminated, and all of this was accompanied by an audible warning, consisting of four consecutive beeps that could be heard several times subsequently. Neither pilot noticed or warned any of this, and as the autopilots continued flying on the now completely unguided approach, the plane continued to descend. This is a bit difficult to explain since the captain had advised that any deviation from the glide slope below 1,000 feet, as well as any warning, should result in an immediate maneuver, but unfortunately that is the nature of the human mind. .

When we enter adecision tunnel, sometimes it is very difficult to see outside of it. And then at 01:16:18 the armed autoland mode suddenly changed from GROUND 3 to GROUND 2 and this was announced by the captain. This fact probably gives us a clue as to where the captain was focusing his attention, but unfortunately, ACT Airlines procedures still allowed for an autoland approach to continue that particular degradation. This meant that the aircraft continued its approach, now flying over the more than four kilometer long runway as the crew happily shouted out the radial altitudes as they passed by. Both the aircraft and the first officer shouted "500", to which the captain responded "Continue".

And very shortly after that, the GPWS started screaming... - Glide slope. - Which was repeated five times, again with no reaction from the crew. - Gliding slope. - Now this is very, very difficult for me to understand. An audible glide slope warning means that the aircraft is deviating more than 1.3 points from the glide slope and in the context of a low visibility autoland approach, below 1000 feet, this should always mean a go-around, but that's not what happened and the only explanation is that the captain was suffering from some kind of extremely severe confirmation bias based on the LAND 2 indication he still saw in front of him, perhaps along with " get there" and possibly initial fatigue.

In any case, the glide slope warning finally disappeared when the aircraft descended below 200 feet radially, which was the lower threshold for that particular warning. All other GPWS warnings were also silenced at this point as the aircraft was in the vicinity of the airport and was therefore interpreted by the system as being in the landing zone. At 150 feet, the first officer shouted, "Approaching minimums," and by this time the plane had already flown over the entire runway and was now heading toward a small town on the other side. The captain responded again by telling the first officer to look out and the plane yelled "100." These were Category 2 minimums, but it still took two more seconds for the first officer to yell "Minimums," to which the captain responded, "Negative.

Turn around." And then, finally, I pressed the TO/GA buttons, but that final delay to the minimums call turned out to be the last link in this terrible chain of errors, the last in a long line of missed opportunities to regain safe flight. By waiting those two seconds with minimal calls, the plane continued to descend even further, to a height of only 58 feet, and with a landing weight of almost 275 tons, it was very difficult to change all that inertia and get the plane to climb again. The giant engine began to spin and the nose began to rise, but it was too late.

About three seconds after pressing the TO/GA buttons, the plane landed about 930 meters beyond the runway at a speed of 65 knots and then bounced slightly. The right wing hit trees just before landing, severely damaging it, and just seconds later, the landing gear and inboard engine crashed into a concrete barrier, which was part of the airport's perimeter wall. That collision caused a second landing, which also coincided with the impact of the right wing against several buildings in the town, just beyond the airport perimeter. Those impacts deflected the entire body of the plane to the right and then caused it to begin rolling as the wings detached, destroying more than 30 buildings, and it eventually disintegrated and burst into flames. (somber music) All four people on board died immediately, as did the 35 people who lived in the village.

Remember, this all happened in the middle of the night. Rescue personnel came as quickly as they could but there was very little they could do. The main cause of the accident was the lack of follow-up by the accident crew, aggravated by bad weather at the airport. It was impossible to determine why the pilots did not respond to all the warnings they received, other than the fact that psychoemotional factors must have played a role, as we discussed above. The investigation also pointed the finger at Boeing because, although the plane had performed exactly as designed, investigators felt that the complete loss of the glide slope signal should have triggered a more unambiguous warning rather than a warning to take off. to the crew of his decision tunnel. and, hopefully, force a turn back.

Several recommendations were issued as a result of this accident and if you are interested in reading them, I have created a link in the description below and for those of you who are on my Patreon team, we can discuss it at our next hangout. Have an absolutely fantastic day and see you next time, bye.