NTSB Press Release Asiana 214

[PhilosopherPilot]

Which is exactly why a "gated" stabilized approach can be an inherently an unsafe practice, that allows the "I can fix it" mentality to continue down an approach but if the stars magically align right at the gate, you get to go through, even if the approach comes off the rails later on. A continuously stabilized criteria forces a pilot to constantly evaluate their approach as is, rather than shoot for a target or continue based on the false blanket that the target was met so the approach must be safe.

I'm not sure what you're getting at. If you are not stable at 1000 feet, you go around. If you become unstable below 1000 feet, you go around. It's not like the 1000' callout is the ONLY time you evaluate your energy state or your stability. It's the beginning of the time after which you MUST be stable all the way to touchdown.

 

[mshunter]

I'm not sure what you're getting at. If you are not stable at 1000 feet, you go around. If you become unstable below 1000 feet, you go around. It's not like the 1000' callout is the ONLY time you evaluate your energy state or your stability. It's the beginning of the time after which you MUST be stable all the way to touchdown.

Define unstable? Say wind is gusting 20 kts, airspeed drops 5 then shoots up 20, and you are 900 above TDZE. Do you go around? Because technically, wouldn't that be unstable?

My point is, life/flying doesn't fit in a little box all the time.

 

[PhilosopherPilot]

Define unstable? Say wind is gusting 20 kts, airspeed drops 5 then shoots up 20, and you are 900 above TDZE. Do you go around? Because technically, wouldn't that be unstable?

Yes. Our FCOM states airspeed fluctuations of -5 or + 10 knots or greater makes you unstable with the following caveat, "Speed, descent rate, lateral path or vertical path momentary deviations are acceptable if immediately recognized and corrected."

20 knots would probably cause a go-around. That's double the acceptable limit. It just depends on if it's consistent, and reoccurring.

(Just saw your last sentence.)

I get that line flying doesn't fit neatly into a box. The book is written for a little flexibility, but I've run a FOQA program before. There are a significant number who take that flexibility and go too far.

 

[seagull]

Not necessarily. In the A320 had he brought back the thrust levers to idle, autothrust would have turned off and the thrust would have stayed at idle. It seems Boeing does something similar in that 'Hold' mode except autothrottles are still armed (and ready for another use/mode). In that sense, the action of manually bringing back the thrust levers to idle should have set off a mental alarm that he was now in manual control of his own thrust.

Airbus will let you STALL the airplane with the power not coming up?

 

[seagull]

Typhoon -- there were, apparently, Asiana instructor pilots who were not aware of this issue. I am not sure whether the demonstration you discuss is something all airlines teach. Remember, not everyone follows the OEM manuals.

What about the B787 test pilot that was surprised by this? That has the same architecture (which they did not change due to wanting to keep it the same as the 757/767 architecture).

 

[typhoonpilot]

Typhoon -- there were, apparently, Asiana instructor pilots who were not aware of this issue. I am not sure whether the demonstration you discuss is something all airlines teach. Remember, not everyone follows the OEM manuals.

What about the B787 test pilot that was surprised by this? That has the same architecture (which they did not change due to wanting to keep it the same as the 757/767 architecture).

Regarding Asiana instructor pilots not aware of this issue, I would say that is a little surprising. As you know instructors run the gamut from highly experienced and highly competent to marginally experienced and marginally competent. I believe that Asiana pilots get a lot of their training from Boeing Korea. I know the head of that facility and at least one instructor there and they are the former, not the latter. But I will concede that some instructors may not be as experienced and as knowledgeable and it's possible the accident captain had one of them for his training. It's also possible the finer points of FLCH were not covered in enough detail for the student to fully understand. All transition courses cram a lot of material into them so it's quite difficult to get too detailed with every item.

Not sure about the 787 test pilot however, I've seen the interview with the chief test pilot and the chief engineer and they were both fully aware of the design and the reason why. One thing that needs to be noted is that Boeing pilots are not necessarily the most knowledgeable in all areas of operation. They have their specialties. The Experimental test pilots are really good at testing the stuff they are supposed to test towards certification. The Production test pilots are really good at doing the required items on a production test flight. Neither probably gets too involved in all the possible issues with FLCH because it's just not required as part of their tasks.

Perhaps my viewpoint gets skewed because I came from an excellent training department where this was all fully understood. I remember quite vividly a simulator demonstration of exactly what the captain did long before the accident flight. We taught an expanded demonstration to our instructors starting over a year before the accident. Our instructors could explain the written text and they knew not to use FLCH on approach and they knew to disengage both F/Ds on a visual, but until they saw some of the demos in the sim they really did not fully understand some of the possible traps. So yes, I will concede that the training could be better.

FedEx has gone through their issues with MD-11 landings. What has cured it, a new system in the aircraft or just more awareness and better training?



Typhoonpilot

 

[m77y]

What difference does any of this make? An airline pilot should be able to shoot a visual approach!

 

[dasleben]

Which is exactly why a "gated" stabilized approach can be an inherently an unsafe practice, that allows the "I can fix it" mentality to continue down an approach but if the stars magically align right at the gate, you get to go through, even if the approach comes off the rails later on. A continuously stabilized criteria forces a pilot to constantly evaluate their approach as is, rather than shoot for a target or continue based on the false blanket that the target was met so the approach must be safe.

You're oversimplifying this, talking as if we all just take that one "gate" as the only metric for stability. Obviously, that is untrue. Any competent professional is continuously monitoring the approach for stability, regardless of a 1000' ATDZE requirement to be stable.

To add: The reason the airlines have adopted 500-1000' "gates" to be stable is that they work good, last long time.

 

[mshunter]

(Just saw your last sentence.)

I get that line flying doesn't fit neatly into a box. The book is written for a little flexibility, but I've run a FOQA program before. There are a significant number who take that flexibility and go too far.

I'd like to hear more about that FOQA stuff sometime. And I'll concede that there are to many guys who do take it to farm. But setting arbitrary limits on things IMO is dangerous.

 

[Cherokee_Cruiser]

Airbus will let you STALL the airplane with the power not coming up?

If you're below 100 RA then it assumes you are landing and you won't get alpha floor protection. I don't know what the Boeing's speed was at 100 AGL. Theoretically, this could play out in an Airbus if you take autothrust off by leaving it at idle, pitch up to regain glideslope, let the airspeed go into VLS, and continue pitch up (increasing AOA) and getting into alpha prot and right then get at 100 feet AGL and pitch up even more into alpha max. Now below 100 RA, you are screwed by being well behind the power curve. Alpha floor won't save you because the airplane knows you're landing when below 100 RA in the landing configuration. That's how the crew of AF 296 got bit in the butt and pummeled through trees as infamously recorded on live television.

But you are correct that alpha floor protection is available and will prevent a stall when you're in normal law and above 100 RA.

 

[dasleben]

I'd like to hear more about that FOQA stuff sometime. And I'll concede that there are to many guys who do take it to farm. But setting arbitrary limits on things IMO is dangerous.

They're not dangerous. With very, very few exceptions, the industry-standard stabilized approach criteria work very well.

 

[PhilosopherPilot]

I'd like to hear more about that FOQA stuff sometime. And I'll concede that there are to many guys who do take it to farm. But setting arbitrary limits on things IMO is dangerous.

Well, the limits aren't really arbitrary. Among other sources such as ASIAS and ASAP, FOQA data are used to analyze stability and what limits should be set. Back then I could plot every FOQA flight on a chart and show where the long landings were, and show a correlation between approach speed and landing length, for example. When you start capturing millions of flights, the ability to really dig in and look for risk factors is stunning. MITRE/ASIAS have some simply amazing charts/data plots available. They do great work.

We've all been there, and pushed bad situations. We are pilots, and the goal is to land the plane. It's sometimes difficult to get out of that "goal" mindset and call the go-around. But to think that the limits are arbitrary just isn't true. Before something is changed it is analyzed and discussed and reanalyzed and discussed again. Sometimes the changes are rejected only to be reconsidered a year later after the data continues to show an issue.

 

[seagull]

If you're below 100 RA then it assumes you are landing and you won't get alpha floor protection. I don't know what the Boeing's speed was at 100 AGL. Theoretically, this could play out in an Airbus if you take autothrust off by leaving it at idle, pitch up to regain glideslope, let the airspeed go into VLS, and continue pitch up (increasing AOA) and getting into alpha prot and right then get at 100 feet AGL and pitch up even more into alpha max. Now below 100 RA, you are screwed by being well behind the power curve. Alpha floor won't save you because the airplane knows you're landing when below 100 RA in the landing configuration. That's how the crew of AF 296 got bit in the butt and pummeled through trees as infamously recorded on live television.

But you are correct that alpha floor protection is available and will prevent a stall when you're in normal law and above 100 RA.

They were above 100 feet. Also, above 100 feet, how slow will the Bus let you get below the target speed in normal law?

 

[seagull]

FedEx has gone through their issues with MD-11 landings. What has cured it, a new system in the aircraft or just more awareness and better training?

Combination of learning about the kinematic issue I explain here, with training and HUD guidance.

 

[Cherokee_Cruiser]

They were above 100 feet. Also, above 100 feet, how slow will the Bus let you get below the target speed in normal law?

Depends if AT is on. If it is engaged, then it shouldn't drop below VLS. The thrust will increase to keep at or above VLS. If autothrust is off, then you are in manual control and you can drop below VLS, below alpha prot, and into alpha max at which point alpha floor with TOGA power will engage. The alpha max is the max AOA, so in the case of approach to landing, the assumption is you're pitching up for glideslope while the speed continues to decay and your pitch up increases your AOA. That alpha max (solid red tape on airspeed indicator) speed depends on different variables like aircraft weight.

 

[Roger Roger]

Combination of learning about the kinematic issue I explain here, with training and HUD guidance.

I have read that before and found it absolutely mind blowing that there is a school of thought alive out there that says it's a good idea to decrease pitch attitude in the flare in the manner you described.

 

[gotWXdagain]

I'm not sure what you're getting at. If you are not stable at 1000 feet, you go around. If you become unstable below 1000 feet, you go around. It's not like the 1000' callout is the ONLY time you evaluate your energy state or your stability. It's the beginning of the time after which you MUST be stable all the way to touchdown.

You're oversimplifying this, talking as if we all just take that one "gate" as the only metric for stability. Obviously, that is untrue. Any competent professional is continuously monitoring the approach for stability, regardless of a 1000' ATDZE requirement to be stable.

To add: The reason the airlines have adopted 500-1000' "gates" to be stable is that they work good, last long time.

I agree with a continuous evaluation, the problem is as new pilots are coming in and being taught a target-oriented stabilized approach, the mindset is less on achieving a continuously stabilized approach and more on getting through the gate, their evaluation of their approach is based on that goal which-I've seen firsthand in evaluating when I was a check instructor- can lead to some seriously unstable approaches that are continued much later than they should have been, in many cases to the runway, because they somehow managed to make it through the gate and therefore give themselves a false safety net. The mindset has become "i've made it through the gate, therefore I am now invulnerable" with sometimes expensive consequences.

 

[PhilosopherPilot]

I agree with a continuous evaluation, the problem is as new pilots are coming in and being taught a target-oriented stabilized approach, the mindset is less on achieving a continuously stabilized approach and more on getting through the gate, their evaluation of their approach is based on that goal which-I've seen firsthand in evaluating when I was a check instructor- can lead to some seriously unstable approaches that are continued much later than they should have been, in many cases to the runway, because they somehow managed to make it through the gate and therefore give themselves a false safety net. The mindset has become "i've made it through the gate, therefore I am now invulnerable" with sometimes expensive consequences.

Really? I haven't seen that at my airline. We are pretty clear about our unstable approach criteria though. 1000 feet AGL isn't a "gate." It is the point after which you must remain within fairly tight parameters or else you must go around. Maybe the problem is the way it is being taught at your carrier? That seems more like a training issue than a stabilized approach concept problem.

 

[dasleben]

I agree with a continuous evaluation, the problem is as new pilots are coming in and being taught a target-oriented stabilized approach, the mindset is less on achieving a continuously stabilized approach and more on getting through the gate, their evaluation of their approach is based on that goal which-I've seen firsthand in evaluating when I was a check instructor- can lead to some seriously unstable approaches that are continued much later than they should have been, in many cases to the runway, because they somehow managed to make it through the gate and therefore give themselves a false safety net. The mindset has become "i've made it through the gate, therefore I am now invulnerable" with sometimes expensive consequences.

I guess I really haven't seen this at my present and former carriers. No doubt every carrier has its pilots who will try to salvage a bad approach below 1000', but it's very uncommon to see. Even less common for the other pilot(s) to allow it all the way to the runway.

Edit to add: We used to toss in final flaps at 200' or so in the C402 and still be stabilized, but that's not something I'd consider good practice in anything with mass.

 

[seagull]

Depends if AT is on. If it is engaged, then it shouldn't drop below VLS. The thrust will increase to keep at or above VLS. If autothrust is off, then you are in manual control and you can drop below VLS, below alpha prot, and into alpha max at which point alpha floor with TOGA power will engage. The alpha max is the max AOA, so in the case of approach to landing, the assumption is you're pitching up for glideslope while the speed continues to decay and your pitch up increases your AOA. That alpha max (solid red tape on airspeed indicator) speed depends on different variables like aircraft weight.

But can it be "on" but not "armed" so it won't capture the speed?