Continental Flt 3407 Crash at Buffalo/Niagra Intr'nal

[e300d]

Probally one of the most ignorant and disgusting comments about this crash. I want to find this guy and slice his throat.

http://www.michael-crook.com/2009/02/13 ... mment-1925

This guy is beyond words! Slicing his throat would be too quick and easy for him! Calling this crew "murderers" is insanity at its finest. This goof needs more then an ass kicking!

Fly safe all!

http://www.youtube.com/watch?v=yQa8G3336fQ - he is a dumbass on many levels

[rampking]

If I read this correctly the report is saying the crew had the ice ref speeds selected, which in turn makes them higher. It also states they dropped the gear and called for flap 15 around 135kias........exactly how slow does this plane fly??? Whats the highest and lowest stall speed for the q400??? This seems slow, does it not?

King

[xsbank]

It seems to me to be a clue that the "initial upset" was a 30+ degree PITCH UP.... Pusher would preclude that unless it was disconnected, yet the report said it fired.... bizarre.

[big dog]

I am not an expert by any means and everyone has their own flying style. But gear extension and flap extension @ 134 kts a little late in my humble opinion. Normally 10 miles back slow to under 200kts hyd pumps on flap 5. Then around 180kts gear down/glideslope interception, under 171 kts FL 15 can be extended and if FL 35 is an option under 158kts. The power setiing is fluid but usually 15-20% does a nice job and 12-15% for the last 1000 ft to target Vref on the dot. Remember, when the REF speed switch is on 20 kts should be added to your speeds.(ie Vref of 110=120) until the crew is satisfied the control surfaces are clear of ice.

[TopperHarley]

It seems to me to be a clue that the "initial upset" was a 30+ degree PITCH UP.... Pusher would preclude that unless it was disconnected, yet the report said it fired.... bizarre.

Yes, initial pitch UP of +30, followed by pitch down of +20 I believe, followed by a bank of 110 degrees. With a tail plane stall, the a/c will pitch down (as explained in the NASA video), however this a/c pitched up quite substantially.

[Rubberbiscuit]

The stick shaker and stick pusher activated (at which point the autopilot would be disconnect at the very latest), then the airplane pitched up to 31 degrees nose up, then pitched down to 45 degrees nose down and rolled left to a bank angle of 46 degrees, then right to a bank angle of 105 degrees (already 15 degrees upside down).

As I pointed out in my previous post, it seems the shaker and pusher went off first, while still on autopilot, followed by 30 degrees nose up pitch. 30 degrees nose up at that slow a speed with lots of drag out there could probably put the aircraft into a deep stall/spin with no chance of recovery from that low of an altitude. Only speculating of course.

[goldeneagle]

Lots and lots of speculation. Lots and lots of opinion. Lots and lots of folks guessing, but, much of the guessing doesn't seem to be backed by knowledge of the physics. Yes, I am one of those nasty engineers that has worked in the field of aircraft design in the past. So, let me add some commentary in response to a bunch of the last 10 pages of posts.

- Flapless landing options. If the aircraft cannot be landed flapless, it cannot be certified. Plain and simple, end of story. With some aircraft types, it was decided long ago, actually demonstrating the flapless landing during proficiency rides added more danger than value to the excercise, so it was dropped from 'gotta show me during ppc ride' line items. Many aircraft require an initial ppc to be done in a simulator rather than airplane today for exactly that reason, there are items REQUIRED to be demonstrated, that are to dangerous to do in a real airplane, but REQUIRED for initial ppc. For most turbojet aircraft, the V1 cut is the prime and visible example. On some aircraft, the flapless landing is likely another candidate. Nobody will be flying the Q400 from the left seat without having demonstrated a flapless landing in the simulator. Left seat driver in this case transitioned from the ATR to the Q within the last 6 months, so, that initial qualification ride where the flapless landing was demonstrated would even qualify as 'current on flapless landings' by every possible definition of the word in this industry. Irrleavent of total hours in the logbook, the crew was indeed current and qualified for the type, and, between them had sufficient knowledge to assess the option of the flapless landing.

- Stalls, of all varieties. Everybody seems to have a solid understanding of aircraft behaviour in a stall. Have any of you ever done a stall, beit the traditional or the tail stall, in an aircraft loaded heavily aft? Ice accumulation on the belly for _most_ aircraft will result in a seriously aft CofG. I dunno how the Q400 accumulates ice on the belly, but, is there anybody here with FIRST HAND experience that can enlighten us a bit ? I'm going to make an educated guess, since there are increased Vref speeds on this a/c for icing conditions, some accumulation on unprotected surfaces is expected, and hence the increased speed requirements. On _most_ aircraft types, ice accumulation on unprotected surfaces will result in a heavier airplane, with an aft moving CofG. This brings up another question that I have no clue the correct answer, and, I'll ask of any Q400 qualified types in the forum, if the plane is loaded seats full, with fuel for 2 hours, how close to aft limit will your CofG be sitting normally ?

- Data recorders suggest 134cas prior to the incident. Again, not personally familiar with the type, but, I do have on the order of 30 years experience (part time, but, enough flying to put more than 5K hours in the book, most/all on the west coast), of which 1/3 of the time at least is flying in weather that is typical west coast winter. I certainly wouldn't let my 421 slow to 134 with a load of ice on during approach to yvr, so, that number seems a little slow to me. Slow airplanes present a higher angle of attack, and, that's conducive to more buildup of ice on the belly, makes the airplane heavier, and NOT by a small amount.

- Agree wholeheartedly with doc. Experience does teach when it comes to ice. One learns to recognize the arse munching the seat, and say to oneself 'I really dont want to be here, but, now that I am, what am I going to do different today because of it?' The vast majority of my exposure to ice is flying single pilot, so, I dont normally have the benefit of a conversation with partner to help clarify thoughts on the conditions. But, I do have long conversations with 'self' in these conditions.

- CVR data initial 'first listen' suggest ice buildup was 'significant' according to crew conversation. In this context, significant is a function of the crew past exposure to icing conditions. If they have seen it a lot, significant is a fair amount of ice. If they only see ice rarely, what they call significant may indeed be 'minor' to those who see it a lot. Was the rate of catch high or low ? None of us know, yet. BUT, the CVR data analysis so far is just 'a quick listen to what the crew was saying at the time'. Detailed analysis will take some weeks, or months, but rest assured, there will be a techie somewhere that will analyze the heck out of that recording, and one of the things they will be looking to pull out of the background noise will be a detail that answers that question. The CVR will have at least 'faint' recordings of ice being shed from the propellers, and hitting the fuselage. It's going to take some number of weeks to pull that out of the recording with some certainty, but it's there. When that comes out, we will have an answer as to wether the catch rate was light, moderate, or severe in the moments leading to the upset.

- FDR data suggests a violent pitch up initially at the upset. All the talk here up to now has been regarding ice on the tail as the prime potential trigger cause. What if all the surface de-ice was working just fine, but, there was a failure in the ice protection somewhere in the pitot-static system ? Careful cross correlation of the radar tracks, with the FDR can validate the static data recorded. That same correlation of radar data and the aircraft position information from the FDR can be used to validate pitot pressure readings. This is another of those post-analysis details that will take a few weeks/months. gotta wonder tho, partial blockage of a pitot system due to ice, autopilot trying to nail a speed, what's auto gonna do to the controls ? If auto is trying to nail 135, there is a little blockage in the pitot, 120 gonna sneak up real fast before anybody realizes it, unless ofc 'anybody' has a hand on the column and is feeling that 'this is not normal, something is different today' feeling that makes the arse start munching the seat.

So, now I have an open question to all the Q400 drivers on this forum, and please, no responses from anybody that doesnt' have at least one dash-8 variant listed on thier license. What's that airplane gonna do in the following conditions. You blasted off seats full, fuel for a 1 hour trip with an alternate relatively close to destination, I'm thinking you are loaded to target maximum landing weight on arrival. On the descent you get a moderate rate of catch on all unprotected surfaces, and, you have a partial failure in the pitot heat, so, it's not fully blocked, but, it's got a partial blockage, resulting in a 30 to 60 second lag on the asi. You select gear and flap with auto driving. On that airplane, does auto target an airspeed, or an angle of attack? Is the shaker triggered by an airspeed, or an angle of attack ?

I'm _NOW SPECULATING_ that if auto targets an airspeed, and the asi system is lagging, then, your airplane is gonna get awful slow, awful fast, until the shaker / pusher which monitors aoa kicks in. Now, if that thing gets awful slow, and there's a whole pile of ice on the belly, what's going to happen next ?

[FICU]

As I pointed out in my previous post, it seems the shaker and pusher went off first, while still on autopilot, followed by 30 degrees nose up pitch. 30 degrees nose up at that slow a speed with lots of drag out there could probably put the aircraft into a deep stall/spin with no chance of recovery from that low of an altitude. Only speculating of course.

I see it that the AoA increased first, hence the positive nose pitch, to activate the shaker followed by the pusher which dropped the nose.

Possibley the heavy pilot workload and extra drag and weight of the ice caused the PF to get behind the power curve after the gear and flaps were selected and the speed wasn't recovered.

[Canoehead]

So, now I have an open question to all the Q400 drivers on this forum, and please, no responses from anybody that doesnt' have at least one dash-8 variant listed on thier license. What's that airplane gonna do in the following conditions. You blasted off seats full, fuel for a 1 hour trip with an alternate relatively close to destination, I'm thinking you are loaded to target maximum landing weight on arrival. On the descent you get a moderate rate of catch on all unprotected surfaces, and, you have a partial failure in the pitot heat, so, it's not fully blocked, but, it's got a partial blockage, resulting in a 30 to 60 second lag on the asi. You select gear and flap with auto driving. On that airplane, does auto target an airspeed, or an angle of attack? Is the shaker triggered by an airspeed, or an angle of attack ?

I'm _NOW SPECULATING_ that if auto targets an airspeed, and the asi system is lagging, then, your airplane is gonna get awful slow, awful fast, until the shaker / pusher which monitors aoa kicks in. Now, if that thing gets awful slow, and there's a whole pile of ice on the belly, what's going to happen next ?

Not a Q400 driver, but the -100 uses lift transducers and the -300 uses AoA vanes. I'm pretty sure the -400 also uses AoA vanes. As for what the A/P does in the approach mode (assuming that's what you mean), it simply follows the command bars of the flight director. On an ILS glide slope, it flies a pitch mode to follow the horizontal command. As for speed (both vertical and forward), pilot controls that with power settings. All the flight director knows is that it's supposed to keep the command bars pinned in the center, and it will do whatever it has to do to maintain it. If the power setting isn't sufficient, things eventually get exciting.
On the DH8 there are no auto-throttles (even on the -400).

[lucky37]

Not a Q400 driver, but the -100 uses lift transducers and the -300 uses AoA vanes. I'm pretty sure the -400 also uses AoA vanes. As for what the A/P does in the approach mode (assuming that's what you mean), it simply follows the command bars of the flight director. On an ILS glide slope, it flies a pitch mode to follow the horizontal command. As for speed (both vertical and forward), pilot controls that with power settings. All the flight director knows is that it's supposed to keep the command bars pinned in the center, and it will do whatever it has to do to maintain it. If the power setting isn't sufficient, things eventually get exciting.
On the DH8 there are no auto-throttles (even on the -400).

Canoehead is right in regards to the -100 series in that there is no auto-throttle incorporated into the A/P. The pilot maintains speed through manual power lever selection. I am quite certain the Q400 does not have an auto-throttle as well. Does the ATR the left-seat pilot was flying before moving to the Q400 have auto-throttle?

..you have a partial failure in the pitot heat, so, it's not fully blocked, but, it's got a partial blockage, resulting in a 30 to 60 second lag on the asi.

I don't know if it's possible to have a partial failure of the pitot heat without an associated warning light on the overhead annunciator panel. On the -100 you have pitot heat failure indications that illuminate "when the associated pitot head-heater is not receiving power, either due to power supply failure or because the related switch is off" (taken from AFM). I wouldn't think the Q400 would be any different, but I may be wrong..?

[Valhalla]

- Stalls, of all varieties. Everybody seems to have a solid understanding of aircraft behaviour in a stall. Have any of you ever done a stall, beit the traditional or the tail stall, in an aircraft loaded heavily aft? Ice accumulation on the belly for _most_ aircraft will result in a seriously aft CofG. I dunno how the Q400 accumulates ice on the belly, but, is there anybody here with FIRST HAND experience that can enlighten us a bit ? I'm going to make an educated guess, since there are increased Vref speeds on this a/c for icing conditions, some accumulation on unprotected surfaces is expected, and hence the increased speed requirements. On _most_ aircraft types, ice accumulation on unprotected surfaces will result in a heavier airplane, with an aft moving CofG. This brings up another question that I have no clue the correct answer, and, I'll ask of any Q400 qualified types in the forum, if the plane is loaded seats full, with fuel for 2 hours, how close to aft limit will your CofG be sitting normally ?

The CofG limit on the Q400 (like all Dash-8s) is wide and forgiving. With a full load of passengers, you are almost always close to the center and in balance, assuming properly loading baggage. And in icing conditions, ice will accumulate on anything sticking out like antennas, probes, the nose, but never on the belly itself. Plus, the Q400 has tons of spare power to overcome any weight increase caused by ice.

- Data recorders suggest 134cas prior to the incident.

This is VERY slow for a Q400 that is not yet in flap 15 landing configuration - very close, if not already at the point of stick shaker and pusher with the increase ref switch on.

So, now I have an open question to all the Q400 drivers on this forum, and please, no responses from anybody that doesnt' have at least one dash-8 variant listed on thier license. What's that airplane gonna do in the following conditions. You blasted off seats full, fuel for a 1 hour trip with an alternate relatively close to destination, I'm thinking you are loaded to target maximum landing weight on arrival. On the descent you get a moderate rate of catch on all unprotected surfaces, and, you have a partial failure in the pitot heat, so, it's not fully blocked, but, it's got a partial blockage, resulting in a 30 to 60 second lag on the asi. You select gear and flap with auto driving. On that airplane, does auto target an airspeed, or an angle of attack? Is the shaker triggered by an airspeed, or an angle of attack ?

The Q400, like all modern airplanes, is very computerised. A paritially blocked or failed pitot tube (there are two on the Dash) would cause a mismatch between the two air data computers, and would set off a whole bunch of caution lights, plus automatically disconnect the autopilot.

And with regards to triggering stall protection, there are two Stall Protection Modules on the Q400 that drive the shaker and pusher. They take info from the AoA plus airspeed, flap position and engine torque. They operate independantly of each other and must "agree" to make the system work.

[vortac]

From http://www.cbsnews.com/stories/2009/02/ ... 5800.shtml here is summary of the crew's flight experience.
-------------
Capt. Marvin D. Renslow finished training on a Dash 8 just two months ago, reports the Buffalo News. He had amassed 110 hours in the turboprop's cockpit, logging the bulk of his 3,379 total flight hours on the much smaller Saab twin-engine turboprop. First officer Rebecca Lynne Shaw had around 770 hours flying the Dash 8 and 2,244 total hours. -------------

I won't judge the crew regarding their experience because many things can come of this investigation, BUT I would like to say I was very surprised to read the info.

On the DH8, the SOP's I'm familiar with, the min clean (gear up, 0 flap) flying speed is 150kts in any circumastance, icing or not. Furthermore, there is an instantaneous fast/slow speed indicator beside both pilots AI which will indicates speed relative to the current Vso of the aircraft. I would imagine that in clean configuration at 134kts, even with no ice on the aircraft, would indicate below 1.3 Vso.

[Boobie Toucher]

Capt has 110 hrs on type.
134Kts too slow in the clean config even when not iced up.
Aircraft pitched up because stall warning shocked/scared him and he prob over pitched and got a secondary stall that flipped him.
Pilot error is a much more realistic argument than anything at this point. the dash has two systems for everything such as pitot-static and flight computers, auto-pilot controllers etc.

Any of us dash 8 guys would be screaming "AIRSPEED!!!" at 134kts clean at night in icing.

The PNF didnt apparently. The call at 134 would be "Airspeed" Not "gear down". And that would be in the 100. The 300 is even worse. The 400 is worse than that...

[2milefinal]

Flight Data Show Response to Loss of Speed Resulted in Deadly Stall That Downed Plane


more in Business »
By J. LYNN LUNSFORD and ANDY PASZTOR

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...


http://online.wsj.com/article/SB1234929 ... malertNEWS

[fanspeed]

Yikes.

[Q400 Driver]

Capt has 110 hrs on type.
134Kts too slow in the clean config even when not iced up.
Aircraft pitched up because stall warning shocked/scared him and he prob over pitched and got a secondary stall that flipped him.
Pilot error is a much more realistic argument than anything at this point. the dash has two systems for everything such as pitot-static and flight computers, auto-pilot controllers etc.

Any of us dash 8 guys would be screaming "AIRSPEED!!!" at 134kts clean at night in icing.

The PNF didnt apparently. The call at 134 would be "Airspeed" Not "gear down". And that would be in the 100. The 300 is even worse. The 400 is worse than that...

This seems to be the sad truth ... remembering the calulated stall speed with flap 5 on a Q4 is approx. 110-112 knots turn on the increase Ref speed switch adds 20 knots, so now we're around 130 -132 ish knots .. now add the ice build up and bumpy air.

[Stinky]

I think C of G will come into this also, they had 20+ empty seats, I know on the 300 a dozen or so empty seats can easily throw you out of limits.
Couple that with the low airspeed and ice.
As the previous poster stated, maybe after stick pusher the pilot iinstictively pulled back creating a secondary stall.

[V1RotateV2]

For those of you speculating out there, some facts that may help:

Q400 loaded with 45 pax and 5000 pounds of remaining fuel will weigh below or around 57,000 pounds. Stall speed for flap 0 at this weight is approx 121 kias. Flying in icing conditions the pilot should add 20 kts, so the new "buffered" stall speed is 131 kias. If they were flying at 134 kias, they were only 3 knots over the red stall cue. Extremely slow for a clean Q400.

Flap 5 stall for that same weight is approx 112 kias, again plus 20 in icing.

It takes a lot of power and a few seconds to accelerate a Q400 that has slowed too much, due to the plane's mass. It has tons of power, but is also weighs in at 63,000 lbs.

The Q400 does not use 1.3 Vso as the ref speed, but 1.23 Vso, plus 20 kts for flap 15 or 15 kts for flap 35. This is to account for the icing in unprotected areas of the wing. If the pilots can see the wing and it "looks" clean prior to landing, the increase ref switch can be turned off and revert to normal Vref speeds.

Q400 autopilot will not hold speed during an approach. Will hold altitude, vertical speed or track the glideslope / FMS. Speed will be a function of the pilot's power setting at any time. Autopilot should be used (except in severe icing) to reduce workload.

If the wing, windshield, prop or tail deice system stop working or malfunction, there are a bunch of indications for the pilot. It will not go unnoticed. Even one valve not opening for one boot will trigger an alarm.

From Q400 drivers: with over an inch of ice on the unprotected areas (lower windshield, spigots, etc.), she flies like any other day.

BUF did not have anything over light to occasional moderate icing that night, according to the other a/c reports and the station's weather. On the Q400 the indication of severe icing is noting ice on the cockpit's side windows. No pilot that I talked to could remember ever seeing that.

[Canoehead]

From http://www.cbsnews.com/stories/2009/02/ ... 5800.shtml here is summary of the crew's flight experience.
-------------
Capt. Marvin D. Renslow finished training on a Dash 8 just two months ago, reports the Buffalo News. He had amassed 110 hours in the turboprop's cockpit, logging the bulk of his 3,379 total flight hours on the much smaller Saab twin-engine turboprop. First officer Rebecca Lynne Shaw had around 770 hours flying the Dash 8 and 2,244 total hours. -------------

I won't judge the crew regarding their experience because many things can come of this investigation, BUT I would like to say I was very surprised to read the info.

Why does that surprise you? We are talking about a US regional carrier. In fact, the F/O's time wouldn't be uncommon for a skipper. As for the 110hrs on type, likely only 50hrs of line indoc required before being released.

As for the other comments about being so slow with nothing hanging out, I'd have to agree. I don't know what the speeds are like on the -400, but the -100 and -300 have some pretty low Vlo and Vfe speeds as it is... such that we are always making the selections during deceleration through the speeds. I don't recall hearing any speed restrictions given out on the ATC audio file, so given the captain's TT on type, flying on a dark and dirty night... let's just say if it were me, I'd have been configured long before they seem to have been.

There. I just speculated/arm-chaired. I'm such a hypocrite . I'll just shut up now... (but I still think Jim Hall is on dope )

[Hedley]

I didn't post anything to this (now 12 page) thread,
even though I had nothing to contribute.

[Canoehead]

http://blogs.wsj.com/middleseat/2009/02 ... irmanship/

I'd say that after reading the article referred to in the link (and posted above), it seems like the WSJ has some of their $%^T together. I actually didn't cringe when reading it. Other news outlets could take some guidance from these folks.

Aside from that, sounds like rough days ahead for the people at Colgan regarding training. I bet upset/stick-shaker recovery will become more than the 5 minutes currently allotted to us in the sim. I hope that is the case...

One major factor in the two extremes of airmanship mentioned in the link: TIME.

I didn't post anything to this (now 12 page) thread,
even though I had nothing to contribute.

I wish I had your stamina

[55+]

Interesting item - not the visual graphic on the left.

http://www.nytimes.com/2009/02/19/nyreg ... sh.html?hp

[Highflyinpilot]

That visual graphic on the left is pretty creepy, seing the house intact and all.

[crazy_aviator]

In the "good" old days ,,, the Captain was god and the co-pilot handed out the coffee and did the radios etc ,, NOW, we have a condition wherin there is equality rights, quotas, discrimination ,, and I wonder if the "captain" was overly agressive and outspoken. being paired with a co-pilot who was humble and intimidated by the captain ? CRM, human factors?? Why didnt she speak up about the combination of LOW speed AND gear and flap actuation ?

[Chuck Ellsworth]

In the "good" old days ,,, the Captain was god and the co-pilot handed out the coffee and did the radios etc ,,

This sort of statement always makes me wonder when exactly those " good old days " were.

I do not recall those days but then maybe I am to new to the industry to have experienced that era.

Can you tell me how long ago it was you were flying when you had to work with that mindset crazy_aviator?