I guess I was thinking that if the tail control surfaces were stuck in the right position, it would cause the nose to rise. Maybe the icing situation is a red herring and something else was at fault here?
It’s looking much like they had a straightforward, uncomplicated stall.
There is a transcript of the cockpit voice recorder here (PDF). It’s interesting in that it covers the whole flight. It’s also apparent they were taken by surprise.
Eerie - never read one of those before… 4 min before impact, the FO says:
I’ve never seen icing conditions. I’ve never deiced. I’ve never seen any—
I’ve never experienced any of that. I don’t want to have to experience that
and make those kinds of calls. you know I’dve freaked out. I’dve have like
seen this much ice and thought oh my gosh we were going to crash.
Wow…
The thing that boggles my mind is that, apparently (and this is subject to change if there is more information later) when the stick shaker went off the flying pilot pulled back on the yoke. That is exactly the wrong thing to do.
The media seems to be making a big deal out of the captain not being specifically trained on stick-shakers. Well, OK, it probably wouldn’t be a big deal to add that to testing, but even though I haven’t had the training of a passenger-carrying airline pilot, and I’ve never been in a cockpit with a stick shaker, even I know that if the stick shakes you’re about to stall and don’t haul back on the stick/yoke! I started learning stall recoveries my second or third hour of flight training, presumably the pilots involved in this accident did as well, and one practices stall recoveries as long as one flies. Or should.
Me too. My pilot training in college was brief, but stall recovery was the second thing to learn after how to take off.
I heard the pull back thing from a pretty reliable source, too. According to a friend, the Q400 also will auto dive (for a second or 2) in stall on top of the stick-shake. That might account for him pulling back (ie - if it confused him). Unfortunately, if that’s the case, he did everything he could to counterract the plane’s failsafes.
The FO was a cute, youngish girl and the captain was a guy my age (not young). It sounds as if he were making the sort of chit chat that a lot of guys about my age do with younger women when they’d really kind of like to see them naked, but have pretty much (absolutely) no chance of that happening.
I’d read recently that it’s typically the second chair flying the plane and the captain should be ready to jump in and override. I wonder which of them actually took control at the point stuff started to go tits up? I’m also underwhelmed with the FO’s reaction (screaming). That seems not the most productive thing to do there, although I’m not sure they weren’t already at the point of no return.
Judging by the timestamps, it looked to me like the scream happened at the moment of impact or damned close.
Ya, OK, but the FAA promotes “sterile cockpit” rules which, if I recall correctly, would mean that everything that happens under 10,000 feet altitude should be strictly the business at hand, including cockpit conversations.
(This doesn’t apply to me, since I seldom fly ABOVE 10,000 feet, but even us little plane pilots are encouraged to stay strictly business as “sterile cockpit” during any phase of landing, including from early in the descent/approach stage. In other words, they were screwing up procedures typically followed even by student pilots these days.)
You want to flirt/BS with your flight crew? OK, either do it one the ground outside the airplane, or during cruise. You don’t do it during an important phase of flight like landing, and you especially don’t do it in bad weather!
Actually, standard procedure usually is for them to switch off, so the left seat guy does one landing, the right seat guy does the next. Should be a 50/50 chance either particular one is flying.
My guess is the one who didn’t scream. Typically you only scream when there is nothing more you can do. If you’re trying to recover from a dive you’re doing someting.
It is not the most common reaction, no. The most common last utterance of a pilot is actually most likely to be “Oh, shit”, at least in English speaking countries. However, screaming does occur, in men as well as women, and typically it occurs at the point absolutely all hope is lost.
My estimation (with usual disclaimers) is that by the time she started screaming they were toast and she knew it.
Yeah, the scream happened just before impact (he had just said something like “we’re in”). I think it is quite understandable to scream when you are about to die. Note that prior to screaming she was trying to fix things, she raised the flaps and asked if he wanted the landing gear up, she seemed to remain relatively calm up to the point she realised she was about to die. Incidently raising the flaps was a bad thing to do, while raising the gear was a good thing to do. For the bloke’s part, stalling was a bad thing to do, pulling back when the stick shaker went was also a bad thing to do. Together they crashed the aeroplane.
ShibbOleth, as Broomstick said, it is normal for the FO and captain to take turns at flying each leg. The person whose leg it is is known as the pilot flying (PF) while the other pilot is the pilot not flying (PNF). There are other acronyms to describe the same jobs but essentially one pilot flies the aeroplane, or controls the autopilot if it is engaged, while the other pilot does checklists, makes radio calls, and, most importantly, monitors what the pilot flying is doing.
On this flight the captain was the pilot flying and the first officer was the pilot not flying. Hence she makes all of the radio calls and does checklists and other actions on the command of the pilot flying. So it was the guy who was physically responsible for getting them into the situation and getting them out of it.
The Pylon, you are correct, the Dash 8 400 (as well as the 300) has a stick pusher that will push the control column forward. This initiates stall recovery if the pilot hasn’t already taken action and is usually fitted to aircraft that don’t have particularly nice stall characteristics. The captain on this flight over-rode the stick pusher which would suggest he didn’t fully understand what was happening, possibly due to a failure in his training.
On the topic of the chit-chat, it has been suggested in various places that he was flirting. He may have been flirting, he may have been chatting more than he would with a male co-pilot, on the other hand, some people are just chatty. The conversation seems very normal for a crew doing a routine flight. They discuss methods of filling out paper work, the crusty old captain tells a few tales, they discuss their ambitions, and they discuss their work. It’s all pretty normal except that the chatter doesn’t die off as they get busy flying, and they seem to allow it to distract them. A professional crew doesn’t need a sterile cockpit rule because they will naturally kill their conversation as their workload increases, unfortunately this didn’t happen on this flight.
I found the CVR quite difficult to read. They chatted so much that by the end of the tape, you felt like you’d got to know them a bit.
Ummmm… I looked again - does anyone else get the same sick feeling I do from this? Just after takeoff…
21:32:13.0
HOT-2
they didn’t do a twenty four hour ice protection test.
21:32:15.4
HOT-1
yeah I just did.
21:32:16.5
HOT-2
you did?
21:32:17.2
HOT-1
yup.
21:32:20.5
HOT-1
that’s why I was looking back on all those pages.
21:32:22.8
HOT-2
yeah.
21:32:23.2
HOT-1
it said it was miss— er uh you know.
21:32:27.3
HOT-2
yeah.
21:32:28.0
HOT-1
there uh there was one page there wasn’t anything on it. so it’s like they
they tore it out in error.
21:32:34.1
HOT-2
oh yeah.
21:32:35.4
HOT-1
so.
21:32:58.1
HOT-2
just twenty four ice test complete? or write—.
21:33:02.0
HOT-1
yeah I’d— I’d put twenty four hour uh check’s complete.
Just a confirmation here from the transcript:
(Note: HOT-1 is the captain’s mic, HOT-2 is the first officer’s mic, CAM is the cabin mic.)
I cut out some material from the start of trouble for brevity - the apparent problem starts at timestamp 22:16:26.0. (Translating to the 12-hour clock, that’s 10:16 pm, 26.0 seconds.) The material I quoted above doesn’t even get to 10:17 pm, running from the 35.4 to 53.9 second marks. The first officer screams less than 2 seconds before the recording cuts off.
(Side note: the “#” is in place of an expletive, and the “*” indicates an unintelligible word.)
Yeah, a lot of bad things happened here. These guys did indeed need to be sterile below 10k, but they weren’t, and it’s my guess that neither of them had really ever paid heed to that rule. It’s almost ironic that in their chatiness, they brought up the fact that they had icing, yet never discussed potential actions which may have been required due to such icing. At the very least, I think the icing may have contributed to the plane slowing down more than usual on approach, which is what caused the stall. Whatever was the cause, the pilot’s scan was poor, so he didn’t notice the airspeed drop.
The pilot also seemed very inexperienced with CRM (cockpit resource mgt) and direction. He never seemed to really take charge, verbally, when things began to go wrong. Never directed the FO to raise the flaps and only responded to the gear when asked by the FO.
The bizarre thing to me is that both pilots would think pulling back on the stick was ok. Neither one had mentioned anything about icing on the tail being a concern, so in my mind, neither one figured it was a stab stall. So if it’s clearly a wing stall, what pilot in the world pulls back on the stick? The way that nose pitches up in the animation screams to me kneejerk panic. And stick shakers have been around for many, many years. I’m not a commercial pilot, but I know what they are and how I’d react if I ever felt it. So for the life of me, I can’t understand why two (arguably) qualified pilots would feel stick shakers and, I’m guessing, the buffet, and hopefully glance at the airspeed, and then think pulling on the stick is proper.
I’m not crazy about monday-morning quarterbacking, but this is not much different than flying a healthy plane into a building on a clear day.
This is reaching a bit, but…
Pulling back on the stick and retracting the flaps would be the correct actions if it was a stall brought on by tailplane icing.
That’s a very specific situation, hard to identify, and the corrective actions go against everything pilots are taught about stalls. I doubt very much the pilots would have recognized that situation and reacted properly that quickly. But you never know…
Like I said in my post above, they had not discussed anything like that. It was obvious that they were both horribly inexperienced with icing in general (especially the FO, who was the one that raised the flaps without being ordered to), and had not discussed anything about icing, stalls, and different recovery techniques.
Unless there is a specific, immediate, tactile sensation that totally differentiates a tail stall buffet from a wing stall buffet, there’s no way the pilot would have known. And even if he somehow did know, pulling up on the stick without raising the flaps would be wrong (although I don’t have the aircraft-specific emergency procedures right in front of me to verify that). From the transcript, it’s apparent that while the pilot was busy fighting the plane, the FO took action on her own and raised the flaps.
If I’m the FO in the right seat and the pilot does something totally unexpected in a critical phase of flight (like pulling on the stick to counter a stall buffet or stick shakers when nothing else has been discussed) , I’m yelling, “WHY ARE YOU DOING THAT??” and if I don’t get a quick answer, I’m going to try and physically intervene.
The captain killed them because he had poor leadership, a bad cockpit environment, a poor scan of his instruments, and took the wrong actions in response to an emergency. The FO killed them because she didn’t have the experience to know what was going on, and have immediate, constructive input. Colgan killed them by poor training and low standards. I’m pretty sure that when NTSB’s final report comes out, that’s what you’ll see.
It boggles my mind, in a document as sober and important as this, that they feel the need to put hash marks in place of expletives. The extent of American prudery never ceases to amaze me.
On the issue of tail-plane stalling, I gather that the Dash 8 does not suffer tail-plane stalls (which explains why I haven’t been trained on them.) Earlier speculation about tail-plane stalling, including from me, appears to have been a red herring.
The icing may have slowed them down a little bit, but really, the main thing slowing them down would be their configuration change. Like other turbo-props, the Dash 8 slows down very quickly. Put the condition levers to max (props to fine pitch), pull the power back to flight idle, and you’ll be slowed down from cruising speed to approach speed in a couple of miles. There is a real need to increase the power as the condition levers go to max on approach, to compensate for the massive air-brake effect the props have.
mhendo, I don’t think censoring of expletives is peculiar to the USA. I’ve seen similar treatment of recordings from other countries. I guess they consider the exact words to not be pertinent to the investigation.
The Pylon, I’m not sure what you’re getting at with the CVR excerpt about the icing checks. The 24 hour icing checks are a check of the de-icing equipment done as part of the 24 hour checklist. As the name suggests, they’re done once a day. In our operation the crew does them. It is no big deal if the checks haven’t been carried out, it’s really just a matter of running the de-ice system. If the checks haven’t been done and you then encounter ice, you effectively do the checks by turning the system on. If there are any malfunctions, you’ll get a master caution light to alert you. My understanding of that conversation is that the de-ice checks haven’t been signed off as completed, so the captain did the checks himself, that’s really just a minor issue and is unlikely to have had any impact on the accident.
Broomstick, I’m not sure how other companies operate, but ours doesn’t have any recurrency checks for stall recovery. I have done stall recovery three times in the last four years, once when doing the Dash 8 endorsement, once when being check out on the slightly larger 300 series Dash 8, and once as part of my command upgrade. Prior to flying the Dash 8 it was always something that was done as part of initial type training, but not on subsequent checks. I think the assumption is that stall recovery is a basic skill that does not need to be practiced or assessed on an ongoing basis. For the vast majority of pilots, that’s valid. Whether that thinking should be changed in response to this accident is open for debate.
In speaking with ATP level professional pilots I’ve heard frequently that they do get to do stall recoveries. Indeed, you’ve had it three times in the last four years. Even if not explicitly part of a schedule of recurrency training your own experience shows that pilots do get stalls periodically in training. Perhaps it should be included for those who haven’t had it for a long while but given that regional pilots typical haven’t and aren’t flying the same aircraft for decades at a time they should encounter stalls in training and in getting endorsement. For such a basic skill no, you shouldn’t need to review it frequently but it should show up occasionally.
True, though I’d call it sporadic rather than periodic. If I were to stay working for this company flying Dash 8s till I retire, I could conceivably never do a stall recovery again. Personally I don’t find intentional stall and recovery to be very valuable. It’s too artificial. I think it is more valuable doing unusual attitude recovery where the risk of stall is increased. We do those at each six month check. During most of those you won’t get close to the stall but every now and then the stick shaker will go and when it does it’s a far more realistic, dynamic situation than a straight and level stall and recovery. It’s also very easily fixed with standard recovery technique. In fact, because the shaker activates sometime before the stall occurs, all it needs is a handfull of power.
ultimately the simulator time is limited and aside from sequences mandated by the regulator, it is up to the company’s training department to make the best use of the time. There will always be some things that are neglected in favour of something else.
Colgan claim that their pilots get stall training and retraining, but something’s obviously gone amiss here.
I’ve been in a similar situation to the accident crew except it was in the simulator and we had a failed engine. Like the accident crew we were distracted (by a high workload rather than chat, but the effect is the same), and like the accident crew we had the de-ice gear on which artificially increases the speed at which the stall warnings activate. We were much lower than Colgan, we were turning finals in bad weather at about 500 feet and the stick shaker activated due to poor speed management by the captain and inadequate monitoring and support by the FO (me.) Unlike the Colgan captain, my captain said “oops” and increased the power a suitable amount. The speed increased and the stall was avoided. It saddens me to think that such a simple action could have saved the Colgan crew and their passengers.
I’m finding this discussion very interesting and would like to thank our flying Dopers for their insights.
To help me follow the discussion, can you let me know if I’ve correctly understood three points:
-
a stall is triggered when the plane’s speed drops below a certain level, so the airflow over the wings is no longer sufficient to provide lift? and,
-
the best way to recover from a stall is to put the nose down somewhat (by pushing the stick forward), to allow gravity to increase speed quickly, but obviously avoiding a nose-dive? and,
-
by putting the nose up (by pulling the stick back), the danger from the stall is increased because the airspeed drops even more, decreasing the lift further?
Thanks!