Air India 787 Crashes in Western India

[Larry in TN]

Any idea if Boeings have stick pushers, and if so can they be manually overridden or are they inhibited below a certain altitude?

The 737 does not have a traditional stick pusher. In the NG, the Speed Trim System will roll in nose-down stabilizer trim to bias the nose down. The MAX uses the MCAS system to, essentially, do the same thing. MCAS is a function of the Speed Trim System. The Elevator Feel Shift system also increases hydraulic pressure from the A system to approximately double the control column forces necessary to hold the pitch at the high AoA.

The 757 and 767 also do not have a stick pusher system. Some 767s have a stick nudger system which is less aggressive than a stick pusher. The 75 and 76 also has a function which fully extends the slats, from partial to full position, as the critical angle of attack is approached.

The 787 does not have a stick pusher but an active system which limits control surface movement and provides tactile input to the pilot.

All of these Boeings have a stick shaker system which is the stall warning.

In the U.S., most pilots begin their careers in basic aircraft like Piper Cubs or Cessna 172s—getting tossed around in turbulence

A growing group of new airline pilots in the US have never flown a non-EFIS airplane or an airplane without GPS. It certainly leaves gaps in their knowledge of navigation and how the ATC system works. I once flew with a co-worker who didn't know the differences between RNAV, GPS, and FMS. He thought they all referred to the same thing. You see threads along those lines occasionally here, too.

I didn't see a stall but I did see an increase in vv down. What would your input be and the resultant vector?

It is difficult to stall a transport jet. You really have to pull hard through all of the protections. You can, however, develop some insanely high descent rates without being fully stalled. See AF447.

Losing all engines at 200' on takeoff is not going to end well no matter what you do. Keep the nose up, and airspeed dropping, isn't going to help. Maintain as much airspeed as you can and convert that into a lower rate of descent just before ground contact.

 

[SLB]

Sunday they found the CVR and DFDR more than enough time to at least get some information out. Just who has the CVR and DFDR and what are they doing with it?

It seems that India's Aircraft Accident Investigation Bureau has the boxes, but given the multitude of fingers in the pie (UK, US, Boeing, prob GE) who knows when we'll hear what's been recovered from them.

 

[35 AoA]

Survivor reported the green and white emergency lights coming on prior to the crash. This points to a more serious problem than thrust being reduced to idle.

I'm not sure that I believe this points to a problem that results in the airplane being a smoking hole in the ground, but it does tangentially bring up a good point. If they did suffer some sort of a primary power failure of the main busses, this would presumably kick off both the automation/autopilot/autothrottles, as well as the flight directors (and potentially primary displays). Depending on upbringing as an aviator, this could have introduced massive confusion on the flight deck that would probably be significant enough to fly into the ground.

 

[Larry in TN]

who knows when we'll hear what's been recovered from them

They don't just take the boxes, plug them in, and dump the data. They have procedures for how that process is accomplished so as to avoid damaging the data on recorders that have typically been through high g-forces, heat, water, fire, etc. It is a slow process.

Once the have retrieved the data, there is a process of analyzing the data, transcribing the multiple voice recorder tracks, then getting everything synced. Then subject matter experts analyze the data an draw conclusions from it.

It takes time.

I'm not sure that I believe this points to a problem that results in the airplane being a smoking hole in the ground, but it does tangentially bring up a good point.

The airplane has multiple electrical busses and multiple power sources. Each bus can be powered from multiple sources. The sources aren't generally paralleled so the process to switch from one source to another involved first disconnecting the previous source before connecting to the new source. It happens automatically, but not instantly. During that process, lights will flash. This could be what the survivor saw.

The cabin lighting is a relatively low-priority item. They are typically on one of the utility busses with other things such as galley power and recirc fans. With a loss of power, the utility busses will be shed first resulting in the emergency lights coming on.

 

[flyingcheesehead]

A growing group of new airline pilots in the US have never flown a non-EFIS airplane or an airplane without GPS. It certainly leaves gaps in their knowledge of navigation and how the ATC system works. I once flew with a co-worker who didn't know the differences between RNAV, GPS, and FMS. He thought they all referred to the same thing. You see threads along those lines occasionally here, too.

VOR/DME RNAV approaches are all gone, and there's not many planes that still have one of those RNAV units. LORAN is dead. Most newer pilots will probably only have seen the term RNAV at the top of an approach plate where it says RNAV(GPS) so I'm not surprised at the confusion there.

I have seen so many arguments among pilots about what is or isn't a "real" FMS that I'm not surprised there is some confusion as well. Technically, a G1000 with the GFC700 is an FMS...

 

[AvNavCom]

There's no question that automation enhances safety, but there’s real value in a training progression that starts with minimal technology and builds up to a fully automated cockpit.

O/T - I'm somewhat persuaded by the opposite school of thought: start with maximum automation and then slowly strip it away down to fundamentals. Sort of like a helicopter student only getting to practice with one axis at first and slowly being given more and more responsibility. I certainly trained the old-school way in the days before GPS and I have a love of both pilotage and stick-and-rudder skills, but you'll pry my GNX-375 out of my cold, dead fingers and I'll add an autopilot as soon as my street-drug side hustle takes off.

 

[Larry in TN]

VOR/DME RNAV approaches are all gone, and there's not many planes that still have one of those RNAV units. LORAN is dead. Most newer pilots will probably only have seen the term RNAV at the top of an approach plate where it says RNAV(GPS) so I'm not surprised at the confusion there.

In 2014 I was furloughed from a job flying 767s. I briefly worked for a regional airline flying CRJs before getting my current job at one of the big 3. The CRJ-200 has a single FMS with DME/DME updating and GPS updating via a single GPS.

During boarding, our GPS failed. The Captain quickly contacted maintenance and did a flight crew deferral of the GPS. When he got off the phone, he said, "We'll have to go 'green needles'." On the CRJ, "green needles" means displaying the raw data VOR on the screens and tracking VOR radials. I said, "No. Once we get a thousand feet, or so, we'll have good geometry on several DME stations, ANP will drop and we'll be good for terminal and enroute RNAV. We just won't be able to do RNAV procedures." He had no idea what I was talking about.

To that Captain, RNAV, GSP, and FMS were all synonyms. If the GPS wasn't working then we didn't have RNAV. He'd learned that the system had DME/DME updating in ground school but he didn't understand that it could provide RNAV without GPS.

Those 767s that I was flying before that job had no GPS. They had DME/DME/IRU FMS systems. I crossed the Atlantic Ocean with that system, though not on the NAT tracks as the RNP was too low for DME/DME/IRU RNAV.

RNAV procedures (SID, STAR, and approaches) will sometimes have a note that says "DME/DME/IRU RNAV-0.3 N/A". RNAV procedures that do not have that note can be flown with non-GPS DME/DME/IRU united if the applicable RNP can be maintained.

So, yeah. GPS, RNAV, and FMS all have distinct and separate meanings. The information is all in the AIM but many don't read it.

 

[Larry in TN]

I'm somewhat persuaded by the opposite school of thought: start with maximum automation and then slowly strip it away down to fundamentals.

That's de-automating. The ability to appropriately de-automate is a skill that is taught, practiced, and learned.

When things aren't going right, de-automate, as needed, to maintain aircraft control. i.e. LNAV/VNAV not giving the desired result? Switch to HDG and Vertical Speed modes. Auto-throttles not doing what you need? Manually set thrust. etc.

 

[AvNavCom]

That's de-automating. The ability to appropriately de-automate is a skill that is taught, practiced, and learned.

Seems like a solid skill for seasoned pilots. What do you think of it as a tool for teaching instrument approaches or holds to students? Show them a perfectly-flown, fully-coupled version with full automation, and then slowly remove the assistance until they're flying the inevitable partial-panel, no-gyro, aircraft-on-fire version.

 

[Doc Holliday]

Sunday they found the CVR and DFDR more than enough time to at least get some information out. Just who has the CVR and DFDR and what are they doing with it? This is not a First Generation Foil Recorder, Today all Boeing current-production models use DFDR systems, which will store 64 12-bit words per second (wps) over a 25-hr period in electronic memory. Also Boeing uplink technology allows real-time remote access to aircraft data parameters via the ACARS digital datalink system, they should have some information on the timeline of the crash.

Not going fast enough for you?

 

[Rick A]

Losing all engines at 200' on takeoff is not going to end well no matter what you do. Keep the nose up, and airspeed dropping, isn't going to help. Maintain as much airspeed as you can and convert that into a lower rate of descent just before ground contact.

Good advice, regardless of what you're flying...

 

[Larry in TN]

What do you think of it as a tool for teaching instrument approaches or holds to students? Show them a perfectly-flown, fully-coupled version with full automation, and then slowly remove the assistance until they're flying the inevitable partial-panel, no-gyro, aircraft-on-fire version.

When I was teaching instrument flying, I first taught basic attitude instrument flying. How to use the flight instruments and flight controls to achieve the different performances that you need in instrument flying. I separately taught instrument procedures outside of the airplane. I used navigation simulators to teach the concepts of navigation and how the procedures were flown. Once skills were developed separately, I brought them together in the airplane.

IMO, far too many CFIIs start having their students fly instrument procedures, that they don't fully understand, while they are still struggling to made the airplane do what they want solely by reference to the instruments.

I think we're straying rather far from the topic of this thread.

 

[Tarheelpilot]

There are no documented cases of any of this detection tech failing/making mistakes? Would hate to have this super ubiquitous tech that has worked perfectly for the past 10 years incorrectly identify a condition and apply the wrong remediation on an airliner.

And what exactly is the remediation if birds are detected past V1 / just at liftoff?

Oh. Lots of examples. Teslas famously don’t see motorcycles. Lots of motorcycle riders being turned into pancakes on Tesla front bumpers.

 

[40YearDream]

They don't just take the boxes, plug them in, and dump the data. They have procedures for how that process is accomplished so as to avoid damaging the data on recorders that have typically been through high g-forces, heat, water, fire, etc. It is a slow process.

Once the have retrieved the data, there is a process of analyzing the data, transcribing the multiple voice recorder tracks, then getting everything synced. Then subject matter experts analyze the data an draw conclusions from it.

This is the recorder unit in the 787. I was the lead mechanical designer for this unit. There's 2 of them in the aircraft. They're designed to be a 'combi' unit, with both voice and data; but could be configured differently, I suppose. With it is a backup-power source and a cockpit microphone. The protected memory is the only part that needs to, or is expected to, survive the crash & fire.

https://www.geaerospace.com/sites/default/files/enhanced-aircraft-flight-recorder-3254F.pdf

 

[EdFred]

Oh. Lots of examples. Teslas famously don’t see motorcycles. Lots of motorcycle riders being turned into pancakes on Tesla front bumpers.

Maybe they do and are programmed to ignore them.


Anyone have an idea on what the timeline would be for releasing what they do know from the black box recovery? 2 days? 2 weeks? 2 months?

 

[Flying Keys]

The 737 does not have a traditional stick pusher. In the NG, the Speed Trim System will roll in nose-down stabilizer trim to bias the nose down. The MAX uses the MCAS system to, essentially, do the same thing. MCAS is a function of the Speed Trim System. The Elevator Feel Shift system also increases hydraulic pressure from the A system to approximately double the control column forces necessary to hold the pitch at the high AoA.

The 757 and 767 also do not have a stick pusher system. Some 767s have a stick nudger system which is less aggressive than a stick pusher. The 75 and 76 also has a function which fully extends the slats, from partial to full position, as the critical angle of attack is approached.

The 787 does not have a stick pusher but an active system which limits control surface movement and provides tactile input to the pilot.

All of these Boeings have a stick shaker system which is the stall warning.


A growing group of new airline pilots in the US have never flown a non-EFIS airplane or an airplane without GPS. It certainly leaves gaps in their knowledge of navigation and how the ATC system works. I once flew with a co-worker who didn't know the differences between RNAV, GPS, and FMS. He thought they all referred to the same thing. You see threads along those lines occasionally here, too.


It is difficult to stall a transport jet. You really have to pull hard through all of the protections. You can, however, develop some insanely high descent rates without being fully stalled. See AF447.

Losing all engines at 200' on takeoff is not going to end well no matter what you do. Keep the nose up, and airspeed dropping, isn't going to help. Maintain as much airspeed as you can and convert that into a lower rate of descent just before ground contact.

True in general, but the definition of “just before ground contact” when it comes to deadsticking a 500,000 lb airliner in urban terrain may be a little fuzzy. To wit: this could be interpreted as “just before ground contact.”

 

[Larry in TN]

Teslas famously don’t see motorcycles

Mine does. Don't confuse Autopilot with FSD. Autopilot is an advanced lane-keeping function combined with active cruise control. The state of FSD improved dramatically, for both HW3 and AI4 cars, just four months ago with the release of versions 12.6.4 and 13.2.8. Many things that FSD couldn't do six months ago it can do now.

the definition of “just before ground contact” when it comes to deadsticking

Transport jets generally begin their flare in the 30' to 20' RA range.

 

[Lindberg]

They don't just take the boxes, plug them in, and dump the data. They have procedures for how that process is accomplished so as to avoid damaging the data on recorders that have typically been through high g-forces, heat, water, fire, etc. It is a slow process.

Once the have retrieved the data, there is a process of analyzing the data, transcribing the multiple voice recorder tracks, then getting everything synced. Then subject matter experts analyze the data an draw conclusions from it.

It takes time.

It's obvious that you're part of the conspiracy to keep this information out of the hands of those who really need access to it: POA members.

 

[flyingcheesehead]

Oh. Lots of examples. Teslas famously don’t see motorcycles. Lots of motorcycle riders being turned into pancakes on Tesla front bumpers.

Uh, yeah they do. Motorcycles, bicycles, pedestrians, garbage cans, etc are all correctly rendered on the screen which indicates that the FSD software sees them appropriately.

Construction barrels will sometimes show as cones, but they're effectively the same thing.

 

[Flying Keys]

Mine does. Don't confuse Autopilot with FSD. Autopilot is an advanced lane-keeping function combined with active cruise control. The state of FSD improved dramatically, for both HW3 and AI4 cars, just four months ago with the release of versions 12.6.4 and 13.2.8. Many things that FSD couldn't do six months ago it can do now.


Transport jets generally begin their flare in the 30' to 20' RA range

Coming from Vref calculated for a specific landing weight, configured to land on an unobstructed runway. Fuzzy.

 

[Larry in TN]

Fuzzy.

Losing all thrust at 200' on takeoff will not produce a good outcome. Keep a nose-high attitude won't improve that.

 

[nauga]

It's obvious that you're part of the conspiracy to keep this information out of the hands of those who really need access to it: POA members.

Yeah, I'm not sure where this apparent sense of entitlement to data from foreign operator operating a foreign registered airplane on a flight between foreign airports through foreign airspace and being extracted and analyzed by a foreign government agency comes from. Maybe we should show them our certs and let them know we're standing by to help.

Nauga,
who will keep it on file

 

[nauga]

Don't confuse Autopilot with FSD.

And please, PLEASE don't confuse software considered suitable for automotive applications with anything remotely suitable for flight-critical aviation use unless you've got much more than an owner's manual and marketing materials. Different applications, different environment, different standards, and different requirements.

Nauga,
from the other side of the fence

 

[AvNavCom]

Yeah, I'm not sure where this apparent sense of entitlement to data from foreign operator operating a foreign registered airplane on a flight between foreign airports through foreign airspace and being extracted and analyzed by a foreign government agency comes from.

I seem to recall hearing a podcast with a senior (now retired) NTSB official that discussed such situations. IIRC, the fact that it is a Boeing will necessarily involve a USA (read: NTSB) rep on the investigation team by treaty, since the country of manufacture is a stakeholder. Not to mention that I imagine the NTSB are probably some of the best aircraft accident investigators on the planet. I may be incorrect in my recall, but I would find it very odd if the NTSB is not already involved in this investigation.

 

[nauga]

I seem to recall hearing a podcast with a senior (now retired) NTSB official that discussed such situations. IIRC, the fact that it is a Boeing will necessarily involve a USA (read: NTSB) rep on the investigation team by treaty, since the country of manufacture is a stakeholder. Not to mention that I imagine the NTSB are probably some of the best aircraft accident investigators on the planet. I may be incorrect in my recall, but I would find it very odd if the NTSB is not already involved in this investigation.

I would bet on that, but it will most likely continue to be led by the Indian government's equivalent. That still doesn't give anyone any entitlement to the data. It will likely be released before or upon publication of the final report, but on their schedule, not ours.

ETA: In addition to the NTSB, there will also be Boeing and GE reps (among others) involved. The comment about entitlement still applies.

Nauga,
boarded

 

[AvNavCom]

I would bet on that, but it will most likely continue to be led by the Indian government's equivalent. That still doesn't give anyone any entitlement to the data. It will likely be released before or upon publication of the final report, but on their schedule, not ours.

Indeed, and well-said. I'm content to wait on the official preliminary and final reports.

 

[Matthew]

Losing all thrust at 200' on takeoff will not produce a good outcome. Keep a nose-high attitude won't improve that.

Nose high, tail low.

It looks like the nose cleared the first building, but the tail did not.

 

[Hang 4]

I would bet on that, but it will most likely continue to be led by the Indian government's equivalent. That still doesn't give anyone any entitlement to the data. It will likely be released before or upon publication of the final report, but on their schedule, not ours.

ETA: In addition to the NTSB, there will also be Boeing and GE reps (among others) involved. The comment about entitlement still applies.

Nauga,
boarded

If people get frustrated with bureaucracy in the US, they need to spend time with the folks that moved it to a high art, India. If a task needs 2 people, 10 will be involved. I would not be holding my breath for much useful information to show up in the near term. The first deputy director of the information directorate will need to hold a meeting with 20 people, arguing over whether "happy" or "glad" best conveys what they are trying to say, then 10 more layers will need to approve that.

 

[Kenny Phillips]

And please, PLEASE don't confuse software considered suitable for automotive applications with anything remotely suitable for flight-critical aviation use unless you've got much more than an owner's manual and marketing materials. Different applications, different environment, different standards, and different requirements.

Nauga,
from the other side of the fence

Yeah, the flight software is easier. Much less stuff to hit once you are airborne.

 

[StraightnLevel]

If people get frustrated with bureaucracy in the US, they need to spend time with the folks that moved it to a high art, India. If a task needs 2 people, 10 will be involved.

I've heard it said that Romans invented bureaucratic red tape, the English codified it in their legal system, and the Indians have perfected it as a cultural practice.

 

[nauga]

Yeah, the flight software is easier. Much less stuff to hit once you are airborne.

What's the equivalent of DO178 for automotive software and would *any* software in a car be DAL A?

Nauga,
assured

 

[Tarheelpilot]

Mine does. Don't confuse Autopilot with FSD. Autopilot is an advanced lane-keeping function combined with active cruise control. The state of FSD improved dramatically, for both HW3 and AI4 cars, just four months ago with the release of versions 12.6.4 and 13.2.8. Many things that FSD couldn't do six months ago it can do now.


Transport jets generally begin their flare in the 30' to 20' RA range.

Well I can’t tell which FMS load is in the Tesla behind me so I just avoid all of them when I’m riding.

 

[Kenny Phillips]

What's the equivalent of DO178 for automotive software and would *any* software in a car be DAL A?

Nauga,
assured

Missed the point. Short version: Cars have a problem self-driving, and that's mostly humans, some walking, some driving.

 

[Bill]

Oh. Lots of examples. Teslas famously don’t see motorcycles. Lots of motorcycle riders being turned into pancakes on Tesla front bumpers.

Uh, yeah they do. Motorcycles, bicycles, pedestrians, garbage cans, etc are all correctly rendered on the screen which indicates that the FSD software sees them appropriately.

Construction barrels will sometimes show as cones, but they're effectively the same thing.

Certain Teslas will not recognize certain motorcycles at night, and yes motorcyclists have died in these accidents.

 

[Jdm]

BBC- Under international rules set by the UN aviation body ICAO, a preliminary investigation report should be released within 30 days, with the final report ideally completed within 12 months.

 

[Omalley1537]

Certain Teslas will not recognize certain motorcycles at night, and yes motorcyclists have died in these accidents.

I believe the one local to me was in broad daylight.

 

[Piperonca]

What's the equivalent of DO178 for automotive software and would *any* software in a car be DAL A?

Nauga,
assured

ISO 26262 and maybe not but should be, if a catastrophic outcome could be involved. Funny story: a friend went to paramedic school and on his first call in real life went to a badly damaged car. A young woman had been driving and her body was still behind the wheel. It took a moment for him to realize her head was in the back seat. He quit his job that day and later became, ironically, a software tester.

 

[traumamed]

Lots of people asking about FDR and CVR data already. Indian bureaucracy and the number of hands in the cookie jar aside, no aviation authority is ever going to risk publicly releasing data prematurely ever again.

 

[Kenny Phillips]

ISO 26262 and maybe not but should be, if a catastrophic outcome could be involved. Funny story: a friend went to paramedic school and on his first call in real life went to a badly damaged car. A young woman had been driving and her body was still behind the wheel. It took a moment for him to realize her head was in the back seat. He quit his job that day and later became, ironically, a software tester.

Having worked in jet engine certification for a couple of decades, there are some events the software won't be able to mitigate (same with flight software, or autonomous driving software.) The only "work-around" is to ensure that the probability of those unfixable events occurring is extremely remote. Or at least convince the authorities of that.

 

[Nico_490]

Losing all engines at 200' on takeoff is not going to end well no matter what you do. Keep the nose up, and airspeed dropping, isn't going to help. Maintain as much airspeed as you can and convert that into a lower rate of descent just before ground contact.

That's it.

"Fly the airplane as far into the crash as possible." - Bob Hoover

But in the heat of the moment, I've seen pilots who intentional pull the nose in the pattern, even when the stall warning is barking.