Colgan 3407 Down in Buffalo

[Saab guy]

I Wanted To Show My Solidarity For All My Former Colgan Brothers In This Time Of Sorrow. Prayers To All The People Involved And To The Crew's Families.

Peace And May God Rest Their Sole's

 

[ksu_aviator]

Just out of curiosity, does the Q400 use hydraulics to control the elevator?

Now, this may not be possible, but I'm still trying to figure out why an elevator would move to full nose up when the flaps are extended. The only common link the two systems may have is hydraulics.

So I was wondering if maybe the hydraulic line developed a leak and it just happened that by extending the flaps, the primary hydraulic reservoir was emptied leaving no hydraulic fluid to provide elevator control. Then, because the autopilot trimed for a nose up condition, the trim tab took control of the elevator, causing the elevator to go to the nose up stop and thereby causing the erratic pitch up that led to the stall/spin.

Ok, ok, before anyone lays into me, I know its wild/off the wall, but it bugs me that 2 pilots can be flying along in normal flight and 26 seconds later they are dead. So I'd like to find a logical explination.

 

[Sig]

So I'd like to find a logical explination.

Thank god! Why didn't someone else come up with this idea sooner?

What's an explination, and what does logic have to do with it? When you find it, present it to the nearest NTSB member and smile for the cameras. Make it look like one of those novelty oversized checks, and have your name writ large across the top!

I, for one, am holding my breath until the explination appears. Keep working, this is a noble cause. Hopefully others will join in your datamining quest.

 

[Minimaniac]

Just out of curiosity, does the Q400 use hydraulics to control the elevator?

Now, this may not be possible, but I'm still trying to figure out why an elevator would move to full nose up when the flaps are extended. The only common link the two systems may have is hydraulics.

So I was wondering if maybe the hydraulic line developed a leak and it just happened that by extending the flaps, the primary hydraulic reservoir was emptied leaving no hydraulic fluid to provide elevator control. Then, because the autopilot trimed for a nose up condition, the trim tab took control of the elevator, causing the elevator to go to the nose up stop and thereby causing the erratic pitch up that led to the stall/spin.

Ok, ok, before anyone lays into me, I know its wild/off the wall, but it bugs me that 2 pilots can be flying along in normal flight and 26 seconds later they are dead. So I'd like to find a logical explination.

While possible, I find your theory unlikely for one simple reason (and seeing as I know nothing about the Q400, there may be an even more simple reason)... A hydraulic system failure or loss of fluid to that magnitude should have been annunciated to the crew. This would have been very obvious on the FDR and/or CVR recordings, and would have been released at the same time the aircraft's final attitude and motions were released. In absence of such, I think it is safe to say that no alerts were given to the crew beyond the stick shaker and pusher.

It seems as though this will be a crash as a result of a stall condition. Perhaps it was a tail stall, or perhaps it was a wing stall that resulted from too low an airspeed given the additional weight and drag of the ice on the aircraft. I remember there still being some bumpy rides that night. A decent gust or bump could have been the difference between a safe outcome despite the ice, and the outcome we have.

 

[ms6073]

I, for one, am holding my breath until the explination appears.

Sure you didn't mean to say esplanation as in 'Luuuuccccccy, you got sum esplainin todo'?

Naturally everyone is curious - most for professional reasons - but we all want answers. Unfortunately, in the absence of anything definitive from those tasked with investigating the accident, forum members are left to speculate, postulate, and theorize as to possible causes. There is no reason to get sanctimonious as I doubt anyone intends disrespect towards family and friends of those affected by the crash.

 

[Captainzero1]

My thoughts and prayers go out to the crew. One observation I see is that the lead investigator for the NTSB has caused some of these problems because I don't think he explains himself very well. The recommendation concerning icing is one point. No determination has been made wether or not the crew was flying through severe icing so why mention the AP shouldn't be on or off. Many airplanes went through that area that night and none reported severe icing so for them to have the AP on was perfecrtly acceptable.
Monday morning QB, I think it was a tail plane stall. I did one of those in the sim once and it wasn't pretty. As soon as I called for final flaps off I went. The recovery is opposite and when you are 800 feet above the ground, are you really going to remember. This accident in my opinion will serve as a wake up call to all pilots who operate in the northeast during the winter. As soon as your airplane starts to behave funny with a configuration change, bring it back to its original config and evaluate. If its a no/partial flap with the speed up landing then so be it. I realize some of us operate from some limited strips with performance issues, then go to a nice big one. I know easier said than done, but no one is going to tell me how to fly my airplane.
Have the Horizon guys who fly the Q across the cascades day in and day out chime in on how this airplane handles ice?
Fly safe everyone.

 

[surplus1]

I agree with the last thing you said...

Dash8s do not have, at least not that I know of, a "less than stellar" history of prop problems.

I don't think that the Swedes (launch customers for the Q-400) would agree with you. Hopefully the overspeed problems they experienced have been eliminated.

I believe the type you refer to in "D" was the Brasilia? Weren't they Hamilton Standard props that were shedding blades? At least one of those events was traced to MX.

Yes, the type I was referring to was the Brasilia and yes the props were by Hamilton Standard. However the single instance of sheeding a blade was NOT what I was referring to. The problem I meant to refer to was propeller overspeed. The cause was a defect in the govenor design. To my knowledge it has been corrected but not before it killed a lot of people. The problem was not corrected as a reult of any NTSB report. It was corrected because dedicated airmen who flew the type took the time out of their own lives to solve the problem on behalf of themselves and their fellow airmen. They were not NTSB technicians. They were just pilots who cared.

The speculated and conjectured and investigated on their own, until the finally found the cause. Then they had to force it on a manufacturer who fought bitterly to deny it.

Overspeeds in the Q-400 haven't killed anybody as yet and hopefully never will, but they've come too close for my comfort. All the more reason to make other pilots aware of them and hope they will study those events with a view to avoidance. I have no loyalties whatever to DeHaviland, Bombardier, or Dowty-Rotol and no regrets at not having experienced the wonders of the Dash8. My friend there are no perfect airplanes just as there are no perfect pilots.

The other Dowty-Rotol powered aircraft to which I made reference was the F-27/227 series (Fokker 500). It's propeller control system was a Rube Goldberg from my perspective and I don't regret at all getting out of it or telling anyone why I think it sucks.

Too many people, even Jim Hall (ex-NTSB chair who really oughta both a) know better than to think one accident, due to causes yet unknown, in an a/c type with no history of similar incidents is cause for grounding the fleet, and b) keep his yap shut) are in too much a rush to pin the blame for this crash on someone or something. How about we all wait for the facts?

Speculating is one thing - we all do it and it can produce some useful results. Rushing to judgement is another beast entirely, and it does no one any good while potentially damaging the innocent.

With all due respect I think that we are ALL waiting for the facts and I don't believe anyone has suggested grounding the aircraft or pinning the 'blame' on anyone.

In my own case, I have repeatedly stated that the scenarios I've posted are all hypothetical.

Talking about what may have happened is a natural thing among pilots and there is nothing wrong with it. What we say may all turn out to be hog wash in the final analysis but nevertheless, intelligent speculation doesn't hurt anyone and just might help many to avoid scary situations.

I don't think I've 'rushed to conclusions' at all, nor have I been critical of any individuals.

What I have said so far may not be the cause of this event and I know that. That is why it is hypothetical. On the other hand I'm not a FOX or CNN newscaster and everythigng I've said so far has a technical basis that is factual. You're free to shoot it down if you can, as is everyone else.

Yes, it is speculation but it is also based on knowledge and careful analysis, not hearsay (unless of course you put information released by the NTSB into the category of hearsay).

When the members of the Board are finished they will render an opinion based on a vote. It may be unanimous as to probable cause or it may not be. The majority will carry the day, but that will NOT mean they are correct. It will only mean that is the best they can do.

No one will ever know exactly what happened on that night, including the NTSB. They will collect data, conduct analysis and then make a decision based on a vote. Often they get it right. Sometimes they do not. That's the name of this game.

Keep that in mind as well.

 

[Sig]

There is no reason to get sanctimonious as I doubt anyone intends disrespect towards family and friends of those affected by the crash.

That wasn't sanctimony, that was acrimony.

 

[MALSR]

Just out of curiosity, does the Q400 use hydraulics to control the elevator?

Now, this may not be possible, but I'm still trying to figure out why an elevator would move to full nose up when the flaps are extended. The only common link the two systems may have is hydraulics.

So I was wondering if maybe the hydraulic line developed a leak and it just happened that by extending the flaps, the primary hydraulic reservoir was emptied leaving no hydraulic fluid to provide elevator control. Then, because the autopilot trimed for a nose up condition, the trim tab took control of the elevator, causing the elevator to go to the nose up stop and thereby causing the erratic pitch up that led to the stall/spin.

Ok, ok, before anyone lays into me, I know its wild/off the wall, but it bugs me that 2 pilots can be flying along in normal flight and 26 seconds later they are dead. So I'd like to find a logical explination.

The elevator is controlled by three separate hyd. systems during the approach phase. Loosing hydraulic control of the elevator is not very likely.

 

[firstthird]

One observation I see is that the lead investigator for the NTSB has caused some of these problems because I don't think he explains himself very well.

Chealander is a loose cannon. He is the same guy that showed up at the SWA 1294 press briefing after the MDW accident and proceeded to treat it like it was a fact finding mission when in fact it was to announce the findings. He made some comments about the pilots that were not at all supported by the text of the NTSB report.

I think that he is WAY out of line wrt this Colgan accident. Speculating and proposing theories to the press. I've never seen an accident investigation handled this way and find it disturbing and unprofessional. He needs to find another line of work and the NTSB needs to get back to their job of determining the causes of accidents and away from the job of feeding the press monster with regular, poorly thought out speculations.

 

[surplus1]

I have never seen a turboprop autopilot that could adjust the aileron trim, and I doubt that the Q400's will either.

I don’t necessarily disagree with you. Although I don’t know the Q-400 I too doubt that its auto pilot moves aileron trim tabs. If you got that impression that’s my fault for not being clear enough. Let me try another way to explain what I wanted to say.

As you read this please do not think that I’m trying to give you flying lessons or aerodynamic instruction. I do NOT mean that at all. I’m just trying to clarify what I said in the previous posts – and only because some of it caused misunderstanding. I’m not an expert – this is just an explanation of what I wanted to mean in the hypothesis. I’m just sharing my thoughts on this most difficult matter.

As a function of controlling bank (in a turn) or keeping the wings level (in level flight) the auto pilot must have some type of input to the ailerons, or spoilers, or combination of both to be able to control the aircraft about its lateral axis.

I made the presumption that the primary control used for roll in the Q-400 is conventional ailerons. If that is NOT the case, please let me know.

If it IS the case, then the autopilot must have input to the ailerons. As far as I know, this can be done in two ways: a) the autopilot displaces a tab on the aileron, which in turn moves the aileron, or b) the autopilot moves the aileron itself.

Since I agree that control by trim tab is unlikely in this a/c type - For purposes of the hypothesis I presumed the latter to be the case, i.e., the autopilot moves the aileron itself. It does not move an aileron trim tab. If this is wrong, please correct me.

Now flying on autopilot, let’s maintain an altitude and make a turn by commanding a heading change from 270 to 180. The autopilot will raise the left aileron and lower the right aileron to induce the turn. Displacement of the aileron will continue until the bank angle reaches the ‘normal’ as programmed in the autopilot. This is usually twenty-five (25) degrees in T-category aircraft.

At the same time the autopilot will apply back pressure to maintain the altitude and alpha will increase.

When the 25 deg bank angle is reached (on most aircraft) the ailerons will be returned to neutral. In a machine with good stability there will be no overbanking tendency at this bank angle and the wings will not try to ‘level’ themselves. If either one of those things does occur, small adjustments will be made by the autopilot to keep the bank angle as constant as possible.

If anything (such as lots of ice on the left wing) tries to make the bank angle increase, the autopilot will raise the aileron on the opposite wing to prevent it. If the force that is causing the left wing to seek a greater bank angle continues or increases, the autopilot will continue to apply ever increasing aileron displacement to prevent it.

Eventually, the autopilot – not the aileron – will exceed its limits and will no longer be able to apply sufficient aileron to stop the overbanking. If this happens the autopilot will disconnect.

As soon as that occurs the ailerons will neutralize with the slip stream – whatever was causing the over banking will still be there – and the aircraft will roll hard to the left (due to the sudden removal of the right up aileron that was holding the bank angle. [There’s your 46 deg bank angle the initially occurred according to NTSB]

Note. It really doesn’t matter if autopilot disconnect was triggered by something else. As soon as it disconnects the aircraft will roll to the left. The only exception is if it was in level flight at the time of the disconnect.

In a case where the autopilot has moved the aileron by a trim tab, the tab would remain deflected and the bank angle will not increase immediately or by much. The pilot will have to re-trim the ailerons by whatever method is provided in that aircraft type – when he no longer wishes to continue the turn..

That is what I was trying to say earlier. Sorry if I used the term “trim” in the wrong place. If I said that the autopilot was trimming the ailerons I did not mean to imply that it was moving a trim tab on the aileron.

 

[FMEpilot]

Does anybody know if the Non-rev Colgan Capt was riding in the jumpseat?

 

[surplus1]

Since we’re here, let’s move to the elevator which has also caused some problems (because of what I said) for some people.

The Q-400 either has a conventional horizontal stabilizer with hinged elevators, stabilators, or an adjustable horizontal stabilizer.

Normally, transport aircraft do not have stabilators (all-flying tailplanes). Those are usually found on military aircraft capable of supersonic flight, which can’t work with hinged elevators. Of course there is always an exception and one such case that I know of was the L-1011. It uses stabilators for pitch control. There are a couple of light aircraft such as the Piper Cherokee that use stabilators balanced by servo-tabs. Stabilators are controlled by the stick or yoke.

Some aircraft (most commercial jets) have adjustable stabilizers while others have a combination of elevator and adjustable stabilizer. Adjustable stabilizers are moved by the trim system. Elevators are moved directly either by mechanical or hydraulic linkage.

I believe that the Q-400 has hinged conventional elevator moved by hydraulics. Please correct me if I am wrong. The hydraulics part I got from two sources: a) comments in this thread, and b) Bombardiers web site which shows the tailplane with an area of 180 sq ft., which I guessed would make the elevator too heavy to move without boost.

If we have a hinged elevator that’s heavy enough to require hydraulics, then we probably have elevator trim tabs. The autopilot will either move the elevator via its trim tab or by using the hydraulics. I chose the former (tab) as being the most likely. If this is wrong please correct me.

In the scenario – if the aircraft levels at a fixed altitude maintained by the autopilot and the power to hold that altitude is less than required, the autopilot will trim the elevator nose up continuously until balance is achieved, i.e., power required = power available. Airspeed will decrease progressively until it matches the other two. Equilibrium is essential to sustained level flight.

If in that process the airspeed results in entering the impending stall regime (whether normal or reset for icing) – it will trigger the stick shaker and the autopilot will disconnect.

If that happens – what does the elevator do?

If the autopilot used the hydraulics to displace the elevator – will something in the system relieve the pressure and allow the elevator to return to the slip stream?

If the autopilot used the trim tab to displace the elevator – will something in the system reset the tab to neutral and allow the elevator to return to the slip stream?

If the answer to either of those questions is YES – the nose will drop and we go about partial stall (shaker) recovery.

If on the other hand the answer to either one of those questions is NO – the nose will pitch up - further increasing AOA and the pusher will activate. Depending on how much elevator up trim was input by the autopilot prior to disconnect – the pitch up could be quite rapid and extreme.

I believe, in my hypothetical scenario, that the answer was NO and that is why the nose pitched up violently initially. The pusher may overpower the up elevator (I’m uncertain if it can) but elevator position wont change until the pilot inputs manual or electric nose trim to neutralize the elevator.

If the PF doesn’t change pitch trim he will be left to fight the up elevator and the pusher at the same time. In my opinion, that is the most probable of the two. If this had happened to me, I seriously doubt that I would think about trimming right away. That would leave me with a bucking bronco on my hands.

Once again, that’s all my theory of what might have occurred. Educated speculation and no more.

 

[surplus1]

This is an intentional duplicate post. I put it in the wrong thread before.


Read this article

http://online.wsj.com/article/SB123492905826906821.html

This is an excerpt of the first two paragraphs. Reportedly, the source is 'the investigators' (NTSB)

"Investigators examining last week's Continental Connection plane crash have gathered evidence that pilot commands -- not a buildup of ice on the wings and tail -- likely initiated the fatal dive of the twin-engine Bombardier Q400 into a neighborhood six miles short of the Buffalo, N.Y., airport, according to people familiar with the situation.

The commuter plane slowed to an unsafe speed as it approached the airport, causing an automatic stall warning, these people said. The pilot pulled back sharply on the plane's controls and added power instead of following the proper procedure of pushing forward to lower the plane's nose to regain speed, they said. He held the controls there, locking the airplane into a deadly stall, they added."

Now look closely at the second paragraph.

My opinion: A) The first sentence (2nd paragraph) is plausible. B) I do not believe the second sentence. I think the NTSB's assumption is wrong.

This is telling us that when the shaker activated and the autopilot disconnected - the pilot pulled back hard on the yoke, added power and then held the back pressure throughout.

To put it simply, I think that's unadulterated BS!

I believe the NTSB (at present) is misinterpreting the pitch -up data (which aggravated the stall) as being caused by or resulting from incorrect pilot input.

If true this is damning.

It is also the complete opposite of training and I just don't believe it.

In critical situations pilots (initially) automatically react in accordance with their training. When the shaker goes off - you push not pull.

I think that whatever 'investigator' said this is overlooking the nose-up trim induced by the autopilot prior to the disconnect.

It may turn out that mistakes were made by the crew prior to the upset - but I just don't buy that one.

For now I'll stick to my theory of elevator trim as the cause of the pitch up and the progression from shaker to pusher - the initial stall of the wing.

After the pusher took effect and pitched the nose down - then yes - the pilot applied back pressure. But not before.

These folks were flying on instruments. You just can't convince me that any pilot who looks at the ADI and sees a servere pitch-up is going to try to correct it by pulling on the yoke - no matter what noise the shaker might be making.

If this really came from an investigator - then they're doing a lot more "speculating" than anyone in this thread.

 

[DiverDriver]

Surplus you've been around a long time and have lots of time. But this wouldn't be the first time a pilot reacted by pulling in a stall rather than pushing. And it wouldn't be the first time a full nose up trim condition lead to a extreme nose pitch up with no override by a pilot trying to bring the nose down. People do funny things under real stress despite training. But I agree that we should not castigate the crew before a thorough investigation is complete and this investigator does seem to be out of line so far.

 

[q100]

This is an intentional duplicate post. I put it in the wrong thread before.


Read this article

http://online.wsj.com/article/SB123492905826906821.html

This is an excerpt of the first two paragraphs. Reportedly, the source is 'the investigators' (NTSB)

"Investigators examining last week's Continental Connection plane crash have gathered evidence that pilot commands -- not a buildup of ice on the wings and tail -- likely initiated the fatal dive of the twin-engine Bombardier Q400 into a neighborhood six miles short of the Buffalo, N.Y., airport, according to people familiar with the situation.

The commuter plane slowed to an unsafe speed as it approached the airport, causing an automatic stall warning, these people said. The pilot pulled back sharply on the plane's controls and added power instead of following the proper procedure of pushing forward to lower the plane's nose to regain speed, they said. He held the controls there, locking the airplane into a deadly stall, they added."

Now look closely at the second paragraph.

My opinion: A) The first sentence (2nd paragraph) is plausible. B) I do not believe the second sentence. I think the NTSB's assumption is wrong.

This is telling us that when the shaker activated and the autopilot disconnected - the pilot pulled back hard on the yoke, added power and then held the back pressure throughout.

To put it simply, I think that's unadulterated BS!

I believe the NTSB (at present) is misinterpreting the pitch -up data (which aggravated the stall) as being caused by or resulting from incorrect pilot input.

If true this is damning.

It is also the complete opposite of training and I just don't believe it.

In critical situations pilots (initially) automatically react in accordance with their training. When the shaker goes off - you push not pull.

I think that whatever 'investigator' said this is overlooking the nose-up trim induced by the autopilot prior to the disconnect.

It may turn out that mistakes were made by the crew prior to the upset - but I just don't buy that one.

For now I'll stick to my theory of elevator trim as the cause of the pitch up and the progression from shaker to pusher - the initial stall of the wing.

After the pusher took effect and pitched the nose down - then yes - the pilot applied back pressure. But not before.

These folks were flying on instruments. You just can't convince me that any pilot who looks at the ADI and sees a servere pitch-up is going to try to correct it by pulling on the yoke - no matter what noise the shaker might be making.

If this really came from an investigator - then they're doing a lot more "speculating" than anyone in this thread.

Actually, and I've been concerned about this for years, each of the airlines I've been at (including 2 Dash8 operators) trains stall recoveries with the emphasis on holding attitude and powering out. The reduction in angle of attack comes from increased forward speed, not from change in pitch. Completely bass-ackward from the primary flight training we all received where reducing pitch attitude is the first step to recovery. Now I know the theory behind the "airline method" is that the recovery is from the first indications of a stall, but...

The first indication of a stall is usually the stick SHAKER. The pusher might never be experienced in a training environment, unless recovery is slow/delayed. To my knowlege - correct me if wrong - SAAB 340s do not have a stick pusher. It may very well be that the crew had never encountered one in action before. I'm speculating here - only those familiar with Colgan's Q400 program can say if this would have been the case or not.

To go one step further, lets say the crew of 3407 was attempting this maneuver against a stick pusher. To maintain pitch attitude and altitude would require enough back pressure to overcome the pusher. When it breaks away, all of a sudden there's 80 pounds or so of back pressure on the yoke. That could lead, I would think, to a 30 degree pitch up and an accelerated stall rather easily.

Pure speculation I know. But, if it turns out that this was just a case of a tired and distracted (by ice, new aircraft type, unfamiliar airport, high workload approach, whatever) crew getting too slow and then attempting to recover from the stall indications as per simulator training, then I think we pilots need to take a long hard look at stall recovery training and return the emphasis to the most important thing in any stall recovery, reducing the angle of attack.

As a former Dash8 100 driver, can any of you Q400 folks tell me if your stall recovery training in type involved the pusher? Or was it as described above?

Q.

PS: I am not blaming the crew - I don't know the facts of what happened. But if the scenario I laid out did actually take place, I think we have an industrywide need to re-evaluate stall recovery training.

 

[WSurf]

Actually, and I've been concerned about this for years, each of the airlines I've been at (including 2 Dash8 operators) trains stall recoveries with the emphasis on holding attitude and powering out. The reduction in angle of attack comes from increased forward speed, not from change in pitch. Completely bass-ackward from the primary flight training we all received where reducing pitch attitude is the first step to recovery. Now I know the theory behind the "airline method" is that the recovery is from the first indications of a stall, but...

Its because you are training in a 121 Commerical environment. You're not doing full stalls on the commerical level like you did in the underpowered Piper or Cessna.
Its the first indication of stall, the shaker or buffet you are instructed to recover then.
Added power and either leveling your pitch or maybe lowering a bit should get you out of the impending stall easily.

So no, the 'airline method' is not a$$ backwards.

 

[BirdCatcher]

To my knowlege - correct me if wrong - SAAB 340s do not have a stick pusher.

Saab has a stick pusher

 

[q100]

To my knowlege - correct me if wrong - SAAB 340s do not have a stick pusher.

Saab has a stick pusher

thx - next question: would it ever be encountered in training?

 

[LJ45]

from the WSJ:

According to the plane's flight recorders, Flight 3407's descent into Buffalo was routine until roughly a minute before impact, when the crew lowered the landing gear, followed by the command to extend the wing flaps, which enable the plane to fly at slower speeds.
Almost immediately, these people say, the plane's air speed slowed rapidly, causing a stall-warning device known as a "stick-shaker" to cause the pilots' control column to vibrate. This was followed by a "stick-pusher," which automatically forces the stick forward.
At this point, the captain appears to have pulled back with enough force to overpower the stick-pusher and shoved the throttles to full power, according to people familiar with the matter. Safety board officials said the nose pitched up to a 31-degree angle. Already at a dangerously low speed, the wings immediately stopped generating lift. The plane whipped to the left and then entered a steep right turn, losing 800 feet of altitude in less than five seconds. At one point the right wing was perpendicular to the ground, according to information taken from the flight data recorder.
The pilots continued to fight with the controls almost all the way to the ground, and in the final moments, "it appeared that they were beginning to make headway when they ran out of altitude," said one person who looked at the data.