This reminds me of one of the lessons I learned from a college professor who was also a former fighter pilot that I used when I started flying professionally and still use when I'm doing anything safety sensitive. He said that once you notice that you've made a mistake on some small, less important detail it was a sign that you need to take a minute and go back and check your work on the big, important, safety critical, things. Often when we're fatigued, distracted, or rushed we naturally forget little things or cut corners. But we have to recognize the latter as a sign of the former. In this case it seems like the forgotten pitot cover and failure to set the V-speeds was a very clear sign of being fatigued or rushed. That means it's time to go back and double check your work on the important safety items. I guess we'll find out how it all factored into the accident when the full report comes out.
"He said that once you notice that you've made a mistake on some small, less important detail it was a sign that you need to take a minute and go back and check your work on the big, important, safety critical, things."
Really excellent point, which I've used purposefully in aviation and have extended to other fields.
In the original post (as mentioned), I avoided giving my own "WTF???" reactions. But:
- If you missed the pitot cover, what else did you miss?
- If you missed V1, similarly what else did you miss?
- If you didn't shut down *both* engines and restart the avionics, what error messages might be lurking there?
- Was it time to tell the (very powerful and company-owning) passengers, "We need to think this over."
- If you got an "autopilot holding nose down" message, how would you prepare for the next step?
We have all made mistakes in aviation, and life. Those of us interacting here are fortunate enough to have survived them. (I can think of two or three aviation mistakes I have made that could have gone the other way.) But this seems like quite a sequence of them.
And a reminder to check and check and check again.
Jim, if you imagine yourself in the left seat, and the QR checklist is saying "AP off" but you have EICAS saying "AP HOLDING NOSE DOWN", doesn't that say quite loudly to you, "be ready to grab this thing, it may get sporty"? Perhaps another case where treating the checklist as a "do-list" can bite.
Those pilots were having a very, very bad day. Missing the pitot tube cover was bad all by itself, and suggests they were way behind the airplane even before engine start. It's not as if it's easy to miss if one of them had actually gone over to the starboard side of the aircraft.
The failure to set the bugs also suggests that they were flustered before the first take-off attempt (presumably the bugs hadn't been set for that one either unless they were electronic and had reset to nominal values after the first attempt). And that they didn't use the checklist for the second take-off attempt either.
The ADS-B graph actually suggests that they reduced their climb rate from 3K fpm to 2k fpm about 45 seconds before the incident; perhaps as a reaction to the EICAS warnings. If so, this probably made the problem worse, assuming that the trim motor was no longer working. Slowing the climb, unless they reduced power, would have increased airspeed and thus the amount of nose-down force required, which of course would make the aircraft's attempt to regain its trimmed airspeed all the more abrupt when the autopilot came off when the stabilizer trim switch was turned off (which probably had no effect on the trim motor which likely had already stopped working), disconnecting the autopilot and thus abruptly releasing all the nose-down force.
Something like this happened to me once in a Piper Saratoga SP I'd rented. The trim motor was not working, which is hardly a no-go item for a single-engine piston aircraft. But there was a warning to not use the autopilot if the electric trim didn't work. Of course, in my Sky King desire to use all the shiny toys available, I blithely disregarded that and used it anyway. At some point, I decided to disengage the autopilot, but the aircraft was badly out of stabilator trim by that point. As soon as I hit the switch, there was a very sharp jolt as the nose popped up. I hand-flew the rest of the flight and the return leg. It was good practice (and several hours of hand-flying at night), and a better lesson.
Very sad accident. I wonder if something about the screwed-up preflight caused all of the trim system issues. It seems unlikely that the aircraft was certified with a potential catastrophic fault lurking in the trouble-shooting checklists.
Thank you. And excellent point about the shift from 3000 fpm to 2000 fpm shortly before the upset.
There are some things about this that seem plain, starting with the preflight. As you say, it will be crucial to understand whether other problems stemmed from the beginning of the flight — and if not, how this system could have these vulnerabilities.
I've also had the Whoa!! moment of disconnecting an autopilot and "discovering" how the trim had been set. You would think that the "AP Keeping Nose Down" message would be a heads-up — but of course it's easy to say that in retrospect.
In the small plane world, we're taught about "get there-itis" and "how will this look in the NTSB report?" I know that circumstances are different when you're flying for hire — and flying the company CEO and his family in the company-owned jet. This example may stick in people's minds.
Disturbing, despite the minuscule US air fatalities. As the airlines misplace mountains of luggage, I am skeptical about their other ‘fail safe’ systems.
Yes — this is still overall an extremely safe part of modern life. But the failures are horrifying, when they occur. And they should or may direct attention to larger training / regulation / accountability issues.
Excellent re-cap of the NTSB Preliminary Report, Mr. F. What we see so far is first, the kind of mistake no pilot wants to ever make; the failure to complete a proper preflight inspection, with consequences. Figuring out the rest of the event will involve lots of moving parts; whether (and/or under what circumstances) the Flight Management Computer (FMC) input data drives the speed bugs in normal operations, the origin of the stab / mach trim fail annunciations and the degree to which the airplane (and its auto-flight system logic) was at odds with pilot inputs after the rejected takeoff with a failed pitot tube and more. My new-tech experience was limited to the Boeing 737 NG models. With no knowledge of the Challenger systems I'll have to wait for the full report, but this new conflict between pilot inputs and mechanical control of the aircraft amplifies elements of the "appropriate level of automation" discussion that's been ongoing since the earliest Airbus technology, and continues through the Max80 problem and beyond. I think we'll soon begin to see some pressure to revisit the fundamentals of crew / aircraft interface, for better or worse.
Yes, again, thanks — from an obviously authoritative point of view. Appreciate your distinction between what seems clear, pitot tube onward, and what is going to require more analysis of system failures and vulnerabilities.
Although this failure was different in its origin from the recent runway-incursion close calls, maybe collectively they will direct attention to crew / automation / etc interactions. As you say.
Truly one of the most shocking NTSB reports I've ever read. Many lessons here, but the most obvious one you mention at the top: be extremely skeptical of initial reports about such events. I saw the headlines about "severe turbulence kills passenger" and thought that sounded quite unlikely. If you just start from a purely Bayesian perspective, about 75% of aviation accidents are due to pilot mistakes. In the absence of any other information, that's a pretty good place to start. Sadly, it was true even in this example.
This reminds me of one of the lessons I learned from a college professor who was also a former fighter pilot that I used when I started flying professionally and still use when I'm doing anything safety sensitive. He said that once you notice that you've made a mistake on some small, less important detail it was a sign that you need to take a minute and go back and check your work on the big, important, safety critical, things. Often when we're fatigued, distracted, or rushed we naturally forget little things or cut corners. But we have to recognize the latter as a sign of the former. In this case it seems like the forgotten pitot cover and failure to set the V-speeds was a very clear sign of being fatigued or rushed. That means it's time to go back and double check your work on the important safety items. I guess we'll find out how it all factored into the accident when the full report comes out.
"He said that once you notice that you've made a mistake on some small, less important detail it was a sign that you need to take a minute and go back and check your work on the big, important, safety critical, things."
Really excellent point, which I've used purposefully in aviation and have extended to other fields.
In the original post (as mentioned), I avoided giving my own "WTF???" reactions. But:
- If you missed the pitot cover, what else did you miss?
- If you missed V1, similarly what else did you miss?
- If you didn't shut down *both* engines and restart the avionics, what error messages might be lurking there?
- Was it time to tell the (very powerful and company-owning) passengers, "We need to think this over."
- If you got an "autopilot holding nose down" message, how would you prepare for the next step?
We have all made mistakes in aviation, and life. Those of us interacting here are fortunate enough to have survived them. (I can think of two or three aviation mistakes I have made that could have gone the other way.) But this seems like quite a sequence of them.
And a reminder to check and check and check again.
Jim, if you imagine yourself in the left seat, and the QR checklist is saying "AP off" but you have EICAS saying "AP HOLDING NOSE DOWN", doesn't that say quite loudly to you, "be ready to grab this thing, it may get sporty"? Perhaps another case where treating the checklist as a "do-list" can bite.
In writing this I was trying to be very careful just to quote the NTSB, and not saying "Well, it seems to me."
But, yes, what you say is how it seems to me.
Those pilots were having a very, very bad day. Missing the pitot tube cover was bad all by itself, and suggests they were way behind the airplane even before engine start. It's not as if it's easy to miss if one of them had actually gone over to the starboard side of the aircraft.
The failure to set the bugs also suggests that they were flustered before the first take-off attempt (presumably the bugs hadn't been set for that one either unless they were electronic and had reset to nominal values after the first attempt). And that they didn't use the checklist for the second take-off attempt either.
The ADS-B graph actually suggests that they reduced their climb rate from 3K fpm to 2k fpm about 45 seconds before the incident; perhaps as a reaction to the EICAS warnings. If so, this probably made the problem worse, assuming that the trim motor was no longer working. Slowing the climb, unless they reduced power, would have increased airspeed and thus the amount of nose-down force required, which of course would make the aircraft's attempt to regain its trimmed airspeed all the more abrupt when the autopilot came off when the stabilizer trim switch was turned off (which probably had no effect on the trim motor which likely had already stopped working), disconnecting the autopilot and thus abruptly releasing all the nose-down force.
Something like this happened to me once in a Piper Saratoga SP I'd rented. The trim motor was not working, which is hardly a no-go item for a single-engine piston aircraft. But there was a warning to not use the autopilot if the electric trim didn't work. Of course, in my Sky King desire to use all the shiny toys available, I blithely disregarded that and used it anyway. At some point, I decided to disengage the autopilot, but the aircraft was badly out of stabilator trim by that point. As soon as I hit the switch, there was a very sharp jolt as the nose popped up. I hand-flew the rest of the flight and the return leg. It was good practice (and several hours of hand-flying at night), and a better lesson.
Very sad accident. I wonder if something about the screwed-up preflight caused all of the trim system issues. It seems unlikely that the aircraft was certified with a potential catastrophic fault lurking in the trouble-shooting checklists.
Thank you. And excellent point about the shift from 3000 fpm to 2000 fpm shortly before the upset.
There are some things about this that seem plain, starting with the preflight. As you say, it will be crucial to understand whether other problems stemmed from the beginning of the flight — and if not, how this system could have these vulnerabilities.
I've also had the Whoa!! moment of disconnecting an autopilot and "discovering" how the trim had been set. You would think that the "AP Keeping Nose Down" message would be a heads-up — but of course it's easy to say that in retrospect.
In the small plane world, we're taught about "get there-itis" and "how will this look in the NTSB report?" I know that circumstances are different when you're flying for hire — and flying the company CEO and his family in the company-owned jet. This example may stick in people's minds.
Disturbing, despite the minuscule US air fatalities. As the airlines misplace mountains of luggage, I am skeptical about their other ‘fail safe’ systems.
Yes — this is still overall an extremely safe part of modern life. But the failures are horrifying, when they occur. And they should or may direct attention to larger training / regulation / accountability issues.
Excellent re-cap of the NTSB Preliminary Report, Mr. F. What we see so far is first, the kind of mistake no pilot wants to ever make; the failure to complete a proper preflight inspection, with consequences. Figuring out the rest of the event will involve lots of moving parts; whether (and/or under what circumstances) the Flight Management Computer (FMC) input data drives the speed bugs in normal operations, the origin of the stab / mach trim fail annunciations and the degree to which the airplane (and its auto-flight system logic) was at odds with pilot inputs after the rejected takeoff with a failed pitot tube and more. My new-tech experience was limited to the Boeing 737 NG models. With no knowledge of the Challenger systems I'll have to wait for the full report, but this new conflict between pilot inputs and mechanical control of the aircraft amplifies elements of the "appropriate level of automation" discussion that's been ongoing since the earliest Airbus technology, and continues through the Max80 problem and beyond. I think we'll soon begin to see some pressure to revisit the fundamentals of crew / aircraft interface, for better or worse.
Yes, again, thanks — from an obviously authoritative point of view. Appreciate your distinction between what seems clear, pitot tube onward, and what is going to require more analysis of system failures and vulnerabilities.
Although this failure was different in its origin from the recent runway-incursion close calls, maybe collectively they will direct attention to crew / automation / etc interactions. As you say.
Truly one of the most shocking NTSB reports I've ever read. Many lessons here, but the most obvious one you mention at the top: be extremely skeptical of initial reports about such events. I saw the headlines about "severe turbulence kills passenger" and thought that sounded quite unlikely. If you just start from a purely Bayesian perspective, about 75% of aviation accidents are due to pilot mistakes. In the absence of any other information, that's a pretty good place to start. Sadly, it was true even in this example.
John, thank you. Well put. Appreciate your reading and weighing in.
Thank you, Jim!