AVweb

« Back to Full Story

3407 in Buffalo: FAA Better Pray It's Not Icing

  • E-Mail this Article
  • View Printable Article
  • Text size:

    • A
    • A
    • A

In covering the Continental Connection Flight 3407 crash in Buffalo, we skimmed dozens of web sites to see who was reporting what. Typically, these sweeps reveal reporting of various detail and quality, some of which is painful to read. I clipped this clinker from an AP report: "If a midair de-icing system is not working, guidelines from the Aircraft Owners and Pilots Association Air Safety Foundation say that pilots can take a number of steps, including changing speed, pulling the nose up or down, or trying a 180-degree turn to rid the plane of ice."

That's the sort of sentence that's in the right hemisphere, but about 60 degrees off course. Not that I'm bashing my colleagues in the work-a-day press. You try reporting on a subject you know nothing about and writing to a 10-minute deadline. I'm quite certain I've made similar misinterpretations when I've reported on subjects outside my expertise, which all journalists must occasionally do.

But another comment from one of the leading lights in the industry, ABC's aviation expert, John Nance, caught my eye. ABC's site quoted him as saying he would be surprised if icing caused the crash. Icing, said Nance, "is usually something that this type of aircraft can handle very well…and it's a brand new aircraft."

It's too soon to say if icing was the probable cause of this crash, but Nance's comment struck me as myopic and rooted in the not entirely unreasonable notion that FAR Part 25 certificated aircraft are capable of flying in known icing conditions. After all, they do it every day, day in and day out. Except when they don't.

Before the Roselawn ATR crash in 1994 and the Detroit Comair/Delta Connection crash in 1997—-an Embraer EMB-120RT--I would have shared Nance's view that this type of aircraft can handle icing. And by "type," I'm thinking more of market type: pneumatic-boot equipped turboprops. Because of the routes they fly and the markets they serve, these airplanes spend a lot of time slogging around in bad weather. On a scuzzy, icy northeastern winter day, a crew might see three or four approaches to minimums before lunch. If there's the kind of large droplet icing out there that brought down the Roselawn ATR, they're likely to find it.

The NTSB revealed on Friday that Flight 3407—a Dash 8 Q400—experienced wild pitch and roll excursions immediately after the flaps were extended. This sort of departure at a configuration change is strongly suggestive of wing or tailplane icing, which, according to the CVR, the crew had discussed. However, reports seem to indicate a pitch up after autopilot disengagement, which isn't consistent with a tail stall. Further, the wreckage was contained in a small area and the aircraft appears to have impacted in a flat attitude 180 degrees off its inbound course for the ILS 23 at Buffalo. That's another fingerprint of a stall mush or spin and a further suspicion of icing-induced stall issues. Reports over the weekend indicated that the FDR showed stick shaker activation, so I'm sure the NTSB will look at a garden variety stall not related to icing. CG could be an issue, too.

Nothing of substance is known now, of course, but I think of lot of people in the industry are starting to place their bets. If it turns out to be icing, the FAA will have hell to pay. For years and following the Detroit crash, the NTSB has faulted the agency for not doing more to improve ice protection requirements and crew training, such as disseminating approach speed and configuration charts for approaches in known icing conditions. If icing is implicated here, I predict the NTSB will announce a preliminary conclusion sooner rather than later. And it will ratchet up the pressure on the FAA to respond, as well it should.

Meanwhile, I'll part company with John Nance. I'll be surprised if this crash doesn't have icing all over it. And by the way, for any regional pilots reading this blog, let us know if your training included a review of the Detroit Comair accident.

Comments (55)

Thirty years ago I was flying a commuter aircraft and landed with heavy ice build-up. We had another leg and were short on time and I de-iced the left wing entirely, but left a little on the right. The truck was out of de-icer. It was such a small amount, the weather was cold but clear, and I knew I had flown into the airport with much more ice on both wings. Upon takeoff the aircraft began to roll and it took full aileron deflection to keep it level. A teensey bit more and it would have been all over. The ice on one wing and the clean other wing produced differential lift, and this was so strong it took full aileron deflection to bring the aircraft around for an emergency landing. When I hear this aircraft began an uncontrollable turn and roll, I think about differential lift secondary to an ice-compromised airfoil.

Posted by: Tom Pinto | February 15, 2009 4:38 PM    Report this comment

I agree that the misleading and often downright incorrect reporting that attempts to inform the public is criminal. Trying to explain the difference between de-icing boots and bleed air that heats the leading edge and never lets the ice form in the first place is beyond the ability of any reporter.

Posted by: Dan Unger | February 15, 2009 4:50 PM    Report this comment

there is a report of a dash 7 accident and there seems to be some similarity and a Dash 8 report:

http://www.tsb.gc.ca/eng/rapports-reports/aviation/2005/a05a0059/a05a0059.asp

There may be a trend here. Icing and the lack of stall/slow recognition may be the straw that broke the camels back

Posted by: Christopher Basham | February 16, 2009 5:12 AM    Report this comment

It is a real shame we still have to learn from things like these. I used to fly for a commuter in the Northeast which lost a Jetstream 41 due to exactly this...autopilot remained engaged during an approach while accumulating ice. We changed our SOP's to say that the autopilot would not be used if ice is being encountered. I wonder what their SOP's dictated during icing conditions regarding autopilot use?

Posted by: Graeme Lang | February 16, 2009 8:31 AM    Report this comment

For a NASA video which appears to have great relavance to this accident go to http://video.google.com/videoplay?docid=2238323060735779946

Posted by: Jack Frost | February 16, 2009 10:43 AM    Report this comment

"ice bridging" should be part of the discussion, especially with recent conclusions on the subject that were not entirely uncontroversial.

Posted by: Robert Stewart | February 16, 2009 12:26 PM    Report this comment

In what way should "Ice Bridging" be part of the discussion? I wasn't involved with the previous discussion so I don't know whether the controversial conclusions here refer to NASA's finding that, if consistent with the POH, modern boots should be used early and often?

Posted by: Tom Tyson | February 16, 2009 1:44 PM    Report this comment

Tom, thanks for your comment but it concerns me, and should others. Lack of de-icing before takeoff is perhaps more dangerous then any talk of inflight anti, or de-icing. Just what do you mean "I left a little on the right (ice)"? I must say that I hope you learned a lesson? Worse than what you suggest, having to use the controls as you describe, would most certainly have gotten the hindered wing closer to stall then ever by your very reaction (aileron correction increases AOA [maybe to stall]). If in doubt about tipstalling...use rudder! What would make you think there was any differential icing (hence lift) here? That makes no sense. Your mistake was made on the ground by not fully deicing one of the wings. If you stall..chances are reasonable that it might not be straight ahead (unless your in a Cub). That is: one wing will drop..thens yous spins!

Posted by: eric hanson | February 17, 2009 4:42 AM    Report this comment

Despite the drama between the NTSB and the FAA it seems like icing accidents for Part 121 aircraft are pretty rare. If this turns out to be an icing accident the crew's relative inexperience will no doubt be scrutinized. With a trend toward greater government intervention I won't be suprised to see a push to decertify all turbo prop aircraft from Part 121 operations in icing conditions. The logic being that anti icing systems on jets don't have inflight icing accidents. Of course the implications of this on airservice and cost would be significant in a negative way.

It is true that the various reports (bloggers/journalists)about this accident are laughable. My favorite was that the reason the plane was pointing away from the airport was because the FAA recommends making a 180 in order to deice the airplane..hence the pilots must have executed a 180 turn at the marker (without notifying the tower)in an attempt to remove ice from the aircraft. Brilliant!

Posted by: Mark Hangen | February 17, 2009 9:20 AM    Report this comment

NASA also offers free online training for pilots describing both the ice-contaminated wing stall and ice-contaminated tail stall. From http://aircrafticing.grc.nasa.gov/courses.html select "A Pilot's Guide to In-Flight Icing", then scroll down to the end of the course to "In An Emergency" for descriptions of the aerodynamics, cues and different recovery.

Posted by: Judy Van Zante | February 17, 2009 11:27 AM    Report this comment

I agee with Jim Hall regarding the certification requirements for transport category aircraft. I have experience in piston powered and turboprop powered transport aircraft both with pneumatic boots and with heated leading edges. The level of safety that is expected today requires the anti-icing performance of heated wing and tail leading edges. The trade-off is in fuel and weight. It is worth the trade-off.

Posted by: Steve Hooley | February 18, 2009 9:21 AM    Report this comment

As is common for any aviation accident, it's just too soon to tell. I can understand what Mr. Nance said, as Mr. Bertorelli mentioned, since a sudden pitch up would not seem to indicate an icing problem. I don't like Mr. Bertorelli's comment regarding journalism. If one does not have knowledge, one should not be blindly reporting. There are enough aviation isssues prevalent that any responsible reporting agency should have at least one knowledgeable person on staff. The "cringe factor" is terrible and horribly misleading to the general public.

Posted by: Frank Loeffler | February 18, 2009 9:32 AM    Report this comment

Frank, seems to be that in icing conditions the autopilot would slowly trim the aircraft nose up to maintain altitude. I wouldn't be surprised to see an immediate pitch up when the autopilot let go followed by a stall and then a secondary stall/spin when the pilots saw the ground coming at them. The latest NTSB info seems to show that the pitch up was commanded by the pilots and that the aircraft was inidicating at least 20 knots above stall speed on final. Assuming no problems with the stick shaker, what, other than icing, causes it to activate? On the other hand, how does an airplane that is on the verge of stall reach a 31 degree nose up pitch? Wouldn't it stall way before that? What would cause an a professional crew to yank back on the yoke at just 1800 AGL under ANY circumstances??

I am also struck by the fact that the Captain only had 100 hrs in this type of aircraft. How do members of AVWEB feel about sitting in the back with such a low time (in type) Captain, given the flight conditions? If you walked on an aircraft and the relative times (total, instrument, time in type, currency etc) were posted would you elect not to take the flight given the situation (weather, night etc)?

Posted by: Mark Hangen | February 18, 2009 12:57 PM    Report this comment

Re my above comment, apparently the ref speed for the stick shaker was set higher due to icing conditions, hence the stick shaker/pusher activated when the flaps were set and the aircraft slowed to the ref speed. The reason they were able to pull the nose up 31 degrees is that the aircraft was nowhere near stall speed. But it surley stalled at 31 degrees.

Why did the crew yank the nose up? Perhaps the crew was convinced a tail plane stall was likely and had a response prepared? Retracting flaps and pulling the yoke back would be consistent with a tail plane stall recovery. Maybe they were just behind the airplane, didn't realize what was about to happen and reacted forceably and unfortunately incorrectly. Ref comments above on time in type.

Posted by: Mark Hangen | February 18, 2009 1:28 PM    Report this comment

I was not so specific but take a look at the accident report DHC Dash 7-102, VP-CDY. These pilots were test pilots and if they were fooled when attempting stall tests -- a stall spin accident is the result of any incorrent action. Simply, icing and the lack of stall/slow recognition may be the straw that broke the camels back

Posted by: Christopher Basham | February 18, 2009 11:39 PM    Report this comment

The FDR (2-18 news release) indicated that the stick pusher was activated. A “snatch forward” of the stick is the likely aircraft response to an ice induced tail plane stall and recovery requires a very forceful pull on the stick to overcome the snatch forward. Quite possibly, the pilot mistook the automatic stick push for a tail plane stall and yanked back on the stick. Unlike us armchair flight jockeys, the pilot only had a second to make a call and it would have been a tough call to make.

View the NASA video referenced by a previous commenter; it really explains what is going on with so-called tail plane stalls.

Posted by: Byron Jones | February 20, 2009 2:21 PM    Report this comment

When I was learning to fly, my instructor impressed on me, since you have NO idea what your plane now weighs, the iceing will cause the stall speed to , increase, so increase your airspeed, lest you fall out of the sky, I only saw one reply which said , the aircraft, was 20 mph over stall speed, may not have been enough, but the unfortunate aspect , was that autopilot, was still engaged until seconds before. Sorry, but looks like too slow/ too low.

Posted by: Leighton Samms | February 20, 2009 5:30 PM    Report this comment

An aspect not yet commented on is the possibility of tail icing (run back behind the boots?) which was not yet to the degree of tail stall. This type of icing will occur on only one side of the tailplane and if far enough aft can significantly change the elevator hinge moments. It could cause a change in hinge moments such that stick forces were heavier than normal in one direction and lighter than normal in the opposite direction. This is a classic setup for PIO.

A change in handling would have been masked by the autopilot and therefore would appear as a "sudden" change to the crew.

It is possible that the PIC applied normal back pressure to recover from the stick pusher and the result was an abnormal pitchup due to lower than expected stick forces. But pushing forward might have resulted in less than expected pitch change.

Or the reverse may have happened and the PIC over compensated pulling back.

In either case the PIC would have immediately graduated to the ranks of test pilot with no warning. The fact that they apparently had the aircraft stabilized at impact says they were still flying it to the end. Just to late for a full recovery. While the use of the autopilot in the level of icing experienced was approved, one has to question the continued use during configuration changes where an autopilot authority is challenged in the best of conditions.

Posted by: Snow Man | February 21, 2009 10:47 AM    Report this comment

Assuming a properly operating pusher, pulling back against the stick pusher defeats the purpose of the stick pusher. It is a blunt signal to the pilot that any increase in angle of attack will probably cause a stall and loss of control. Resisting it will cause a dramatic increase in drag and possible stall. In certain windshear recovery techniques the airplane is flown out with the pusher "nibbling" on the control column. The pilot "recovers" from the stick pushers by yielding the control column to the direction it is pushing. Back pressure on the control wheel "normal" or otherwise is reduced to recover.

Posted by: Steve Hooley | February 21, 2009 12:01 PM    Report this comment

Steve,

I think you are confusing the stick shaker with the stick pusher. When the shaker is active you are approaching stall and no change in pitch is commanded by the system. And your right "flying the shaker" is a practiced technique.

However when you reach stall (or what the system thinks is a stall) the pusher commands a nose down pitch change. If the pilot does not believe he is in a stall situation or believes a less aggressive pitch over is required, he must exert a force to pull back on the stick to counter the pusher command.

If stick forces have lightened due to tail ice, a normal input force could result in a larger than expected pitch changes leading to PIO.

Posted by: Snow Man | February 21, 2009 12:54 PM    Report this comment

I still contend that the aircraft was no where near stall speed when the stick shaker/push went off. If so the crew could never have pulled the nose up 31 degrees. I still think the ref speed for shaker was set to 20 knots faster than stall speed and that deplying the flaps reduced th airspeed to that point. Either the crew was surprised and yanked back on the yoke (doubtful as they had enough experience to know that was wrong) or the PIC was anticipating a tailplane stall and initiated a "recovery" when the stick shaker activated. The clue is that they retracted the flaps when the stick shaker occurred. This is consistent with a tail plane recovery. Seems unlikely they would have had the presence of mind to retract the flaps if they were caught off guard and in a panic.

Posted by: Mark Hangen | February 21, 2009 3:59 PM    Report this comment

Snow Man, With all due respect, I have plenty of time in jet and turboprop aircraft and simulators. I know the difference between a "shaker" and a "pusher". If the pilot responds appropriately to the "pusher" there is very little change in pitch attitude. Light stick force has nothing to do with it. The appropriate response is max. power and don't fight the pusher. One can fly the airplane with the "shaker" on constantly and the "pusher" just lightly tapping forward on the control column.

Posted by: Steve Hooley | February 21, 2009 9:15 PM    Report this comment

Addendum: If the pilot overcontrols the pusher that is when the airplane attitudes can become dramatic.

Posted by: Steve Hooley | February 21, 2009 9:19 PM    Report this comment

Gentlemen:

You confuse the tests that were done with a DHC-6 twin Otter versus a Dash 8. The Twin Otter is essentially a big Cessna with two turbine engine. It is just as easy to fly as 172. As such, the Twin Otter has no stick shaker or stick pusher. The advantage for testing is simplicity low cost and no systems interfering with the aerodynamics.

Both the Twin Otter’s stall warning systems and the Dash 8 are based on angle of attack. Both aircraft’s system will warn of approaching the critical angle of attack. In the case or the Twin Otter a buzzer/warning light and the Dash 8’s stick shaker. Triggering either indicates a wing stall.

Posted by: Christopher Basham | February 21, 2009 11:55 PM    Report this comment

I have a commercial instrument rating, and degree from a respected institution in our field. Rah rah, you say..well I must just say that in all of my training, there was no differentiation on "tail plane" stall versus any other stall. I submit that any talk of the pilot considering a stall on the tail versus the wing as ridiculous. Period. No pilot I know of, worries, or reacts specifically to a tailplane stall (from icing or anything else). To think that this or any pilot is reacting to this "tailplane stall" makes me want to say that anyone that submitted such jiberish is not a pilot. I agree with anyone who said anything of this kind: "firewall power" or alternatively: "full military power". Maybe someone has flown a plane here with "shakers, and or pushers", but I submit that we are over analyzing on all of this. Furthermore, the facts aren't in, and I support the pilots (slightly less than capn'Sully). Let's end this thread until we find out facts. Thank you

Posted by: eric hanson | February 22, 2009 5:02 AM    Report this comment

Eric, no disrespect for your commercial pilot's license and degrees, but have you read Advisory Circular 91-51A - especially appendix 2?

As for speculation, yes that is what it is what a BLOG is for. Both you and I can put our opinions and experiences, right and wrong on a subject of interest.

The NTSB report will likely be two years before it is released. In the meantime, the former NTSB chairman, who should know better, has called for the grounding of all highwing turbo props. I don't know his sources but high wing or low wing a tail stall is caused by the same mechanism - ice overload.

Posted by: Christopher Basham | February 22, 2009 7:54 AM    Report this comment

Eric

LOL. You should watch this video. Perhaps you seek a refund from your "respected institution". :-)

http://video.google.com/videoplay?docid=2238323060735779946

Posted by: Mark Hangen | February 22, 2009 8:39 AM    Report this comment

Mark, thanks for your response. With all due respect, you miss my point entirely. I watched the great video, and thanks for that. I simply state that practical test standards that I am aware of do not include this supplementary information. The information may very well be available to pilots in a supplementary way, however the PTS are the FAA's method of standardized training. As such, I simply state that I am unaware of FAA training for potential "tailplane stall" recovery. Again; period.

Posted by: eric hanson | February 23, 2009 2:14 AM    Report this comment

Eric

Not having your level of expertise, I assume that there is no rating or course training standard for flight into known icing conditions.

Certification standards exist for flight into known icing conditions for the airframe only and likely varies from FAR 23 to FAR 25.

I assume that the Dash 8 is a FAR 25 standard for aircraft (transport aircraft design). I also, assume that the Dash 8 has the same level of testing as anyother transport aircraft. I also assume that if your aircraft is certified for known icing and the pilot holds a current instrument rating then you are free to go when icing is forecast (likely IFR). As such, you must be aware of advisory circular AC 91-51A which states the recovery procedure for both roll and pitch upset.

Being a lowly "unkowledgable" VFR pilot please correct me if I am providing jiberish,i.e., wrong.

The question remains was there a real tail stall or did the pilot, as a result of the advisory circular, create the event. This will spill over into either certication or flight standards. There are big bucks on liability and the lawyers will be licking their lips either way who say their only concern is $afety.

Posted by: Christopher Basham | February 23, 2009 5:00 AM    Report this comment

Christopher, your assertions are reasonable. Aircraft are certified for known icing. Also, there are Op Specs (usually a company manual). I simply am saying that I am unaware of "tailplane stall" recovery techniques per se. If the tail were to stall first, this actually is a "self stall recovery" in effect for the aircraft (the A/C naturally will decrease alpha, or angle of attack on its own). My point is that pilot certification does not require some knowledge or specific training for a "tailplane" stall per se. That is ridiculous to propose, and is what I refer to in my criticism. Yes, that is jiberish to presume pilots are required to recover from a "tailplane stall" per se as opposed to any more general stall condition. You yourself mention an advisory circular. It is just that sir....ADVISORY.

Posted by: eric hanson | February 23, 2009 5:24 AM    Report this comment

Eric: Maybe some help for you here. As a CFI, I can tell you that the practical test standards have little to do with real flight proficiency. They are merely the most basic training guidelines for piloting certification. Pilots who hope to survive flying in challenging conditions--weather being one--soon learn that they have to go far beyond the PTS. Tailplane stalls are poorly understood in the general pilot population but have been implicated in a number of accidents, including at least two turboprop accidents. Anyone who flies in ice needs to understand the implications. AC-91-51A and the NASA film provide this. Whether it's advisory or required is irrelevant. It's basic survival knowledge. --Paul Bertorelli

Posted by: Paul Bertorelli | February 23, 2009 7:44 AM    Report this comment

Eric, you are probably unaware of it because certification requirements don't allow for more than one recovery technique. The stall margin must be high enough to allow one identical recovery technique for both icing and non-icing conditions. Reference: FAR 25.207(e)&(f). In other words, you are right--nobody should be pulling back on the stick in a Part 25 airplane during stall recovery if the airplane was properly tested and certified. Personally, I find it easier to believe the pusher 'pulled' rather than that the pilot out-muscled the pusher (79 pounds of nose-down force) to raise the nose up 30 deg.

Posted by: David Tuuri | February 23, 2009 9:07 AM    Report this comment

I see a change in SOP's and some focused simulator training coming for them, if it has not already.

Posted by: Graeme Lang | February 23, 2009 10:13 AM    Report this comment

Per the informative video provided my Mark, above, there clearly are actions a pilot might take. Also, aircraft feedback may give an indication of ice induced tailplane stall. My institution did teach that reduced flap or zero flap may be indicated on approach with actual airframe ice. I did conduct an approach with some minimal ice, and followed the advice using zero flap and appropriately higher airspeed. Obviously this was succesful as I am here writing. The point I have been making, is that pilots are not generally trained to recover from a tailplane stall. I think they should be for these operations at least, and we don't know if the accident here discussed involved a pilot so trained. As Paul suggests, and I agree, the PTS shouldn't be considered limiting as to education. I imagine most pilots conducting operations similar to this accident flight are educated on tailplane stalls. Keep in mind this A/C was said to have "pancaked in", which would be indicative of a stall of the primary lifting surface and possibly a subsequent spin.

Posted by: eric hanson | February 23, 2009 9:25 PM    Report this comment

I agree that it was a stall/spin accident and that the stall was likely a main wing stall. It seems like the stall was pilot induced base on the NTSB feedback that the 31 degree pitch up was pilot directed. I wouldn't be surprised to see that the final analysis shows that ice was a contrbuting factor only to the extent that it caused the pilot to react as he did because (ironically) of his prior training and what he assumed was likely to happen. Ultimately it will come down to pilot training and experience. Given his prior experience in a SAAB 340 I would speculate that he was well aware of tail plane icing. The question that will be raised is if he was so trained why did he leave the autopilot on as this is a no no if tail icing is suspected. My guess is that Dash 8/Commair SOP is to leave the autopilot on in icing conditions. I really feel for the Captain. He had so little time in the DASH 8 and yet a lot of time in the 340 which is more prone to tail plane ice. Seems like he may of had "one foot in each boat". This will raise all sorts of questions about transition training and time in type required to be PIC. Of course the NTSB may find something totally different caused this accident.

Posted by: Mark Hangen | February 23, 2009 10:08 PM    Report this comment

Good thought Mark. The best pilots may make an incorrect conclusion in a moment of instant crisis. Perhaps the mechanisms (stick shakers and movers), actually confuse a pilot's natural reactions. Most definitely the A/P should have been deactivated, this is well accepted after a renouned icing accident during a hold. That, as I am aware of, resulted in a catastrophic stall/spin which went down very hard over nine years ago. This was the famous ATR crash. I don't think you meant to be as harsh as you sounded in your post, Mark, vis-a-vis my respected institution. If they didn't emphasize potential "tailplane stalling", I doubt other institutions would have either. As far as getting my "money back"...geez, I wish (not due to poor education)! Thanks for your input here! My thoughts were based on the prevalent FAA training for all pilots in general, you understand. I think the FAA would be reticent to teach pilots to do an opposite recovery method for stalls (which tailplane stall-due to icing apparently requires). Certainly, they would think more lives are saved through conventional stall indication/recovery methods. Perhaps aircraft certification and pilot certification in known icing need revisiting. I do respect all the input here!

Posted by: eric hanson | February 24, 2009 12:43 AM    Report this comment

Eric and Mark -- I will wager that the flight manual calls for autopilot to be off in sever icing conditions (what every that is). The conditions leads to our mental exercise of guess the conditions with minimal inputs. Yesterday's flash stated that "the pilots simply overreacted when the autopilot disengaged on a slow approach and the stick would dropped the nose at a higher than normal speed. That would be how the stick pusher should behave when the icing system is activated." If this is so -- then the pilots will be faulted for inadvertent stall and icing is a non issue.

Pitch attitude for a full power stall in a 172 is about 30 degrees pitch up and my aircraft a decathlon is about 20 - 25 degrees.

The plot thickens!

Posted by: Christopher Basham | February 24, 2009 1:56 AM    Report this comment

That, as I am aware of, resulted in a catastrophic stall/spin which went down very hard over nine years ago. This was the famous ATR crash.<<

Just to correct the record, the Roselawn ATR accident (1994) was not a stall/spin event, but rather an uncontrolled roll and subsequent high speed dive (more than 300 knots) and impact. It was caused by runback icing blocking air flow over the ailerons.

Posted by: Paul Bertorelli | February 24, 2009 4:38 AM    Report this comment

Eric and Marc-

Have either of you actually felt 80 pounds of force from a stick pusher? Maybe you can find a pilot who can show you how it feels and then test your theories again? I would find it more likely the pilot would be lucky, if overriding it was the intent, that a nose-down attitude would merely be lessened, not raised way up to 30° above the horizon.

Posted by: David Tuuri | February 24, 2009 5:37 AM    Report this comment

Let's be clear: I have no theories stated on the cause of this accident. Mr. Bertorelli does bring up clarification on the '94 ATR accident though, and I remember "aileron reversal" being mentioned also(that alone should bring this thread into the next century). I cannot, and have not, commented on the automatic "pusher" systems. Paul, with all due respect "run-back icing" or not, uncontrolled rolls, e.t.c. in many cases involve an aerodynamic condition we all understand as a stall (almost by definition)again; in most cases. That said, the ATR crash was more complicated (as Paul notes). In any event, the A/P was on and that IS germain to the current event perhaps. David, as I said above I haven't specifically stated a theory here to be "tested". Thanks for your input none the less!

Posted by: eric hanson | February 24, 2009 6:30 AM    Report this comment

David:

You seem to be saying that the crew would be incapable of pulling the nose up against the 80lb force of the stick pusher. Couple of thoughts...it was the NTSB who stated that analysis of the black boxes showed that was exactly what happened. Additionally, the NASA video on tail plane icing indicated that the test pilots had to overcome 175lbs of stick forward force in some cases..so i suppose it is possible. Lastly despite 6000 hours of PIC (3000ish in turbines) I have never flown an aircraft with a stick shaker. Couldn't I just go to the gym and snatch set s rowing machine to 80lbs and see how difficult it would be to accomplish?

Posted by: Mark Hangen | February 24, 2009 6:56 AM    Report this comment

I'm not finding the facts released, so far as I know them, to support a conclusion that the pilot pulled the stick back at all.

1) Did the pitch attitude decrease at all--as a result of the pusher action and as it surely would have been an unexpected surprise?

2) Does adding full power comport with a tailplane stall recovery technique, if that's what the pilot was trying to accomplish?

3) Even if a pilot can muster enough strength to override 80 lbs of nose-down force, would he then be inclined to pull way high into a stalling attitude--and hold it there, while bucking and rolling, all the way to the ground?

On the other hand, if the stick pusher uses the autopilot servo (and I don't know what it uses) to effect a nose-down 'push', isn't it possible an electrical short, say because of condensation, could cause the pusher to pull the nose up with 80 lbs of force and cause a stall? Until I see more facts, I'll go with the latter and grant the deceased pilot the benefit of the doubt.

Posted by: David Tuuri | February 24, 2009 7:27 AM    Report this comment

Chris:

I listened to Buffalo approach / tower for the 30 minutes during which this accident occurred. Not much could be gleened from the exchanges between ATC and the first officer as she was mostly acknowledging ATC headng and altitude commands. Routine is how I would descibe it. What was noteworthy was that there was little mention of icing by any of the aircraft on the freq. My experience is that when icing is moderate (let alone severe) that pilots tend to mention of it. My initial expectations were to hear chatter about ice...so it was surprising to hear so little in that regard. Combined that with the fact that they were apparently able to get the aircaft to a 31 degree nose up attitude and you'd be hard pressed to conclude icing was the culprit.

Posted by: Mark Hangen | February 24, 2009 7:32 AM    Report this comment

David:

Check this out:

http://www.nytimes.com/2009/02/19/nyregion/19crash.html?_r=1&scp=1&sq=3407&st=cse

While it is possible the crew panicked and pulled the yoke full aft and held it there...(the CVR transcript will be helpful in determining) I, like you, would like to believe that they were good enough pilot's to realize how wrong this would have been. Hence my speculation that the PIC theorized that the aircraft was in a tailplane stall in which case, flaps up and full aft would be appropriate.

Posted by: Mark Hangen | February 24, 2009 7:43 AM    Report this comment

Mark,

Although I appreciate the link, the article does nothing to change my mind. The main gist was this, too: "...investigators said they were still not ready to cite human error for the crash.

Among the issues they are exploring is whether the stick pusher activated appropriately."

The article doesn't say how many degrees the nose pitched down, therefore I must assume it didn't. If it did, NTSB should have specified how far down it went down before going up (which WAS specified).

Posted by: David Tuuri | February 24, 2009 8:22 AM    Report this comment

Mark, reading that fairly well written article, I note that one could surmise that the pilot had been trained for tailplane icing and applied that recovery technique(counter to traditional stall recovery). He had been in an aircraft noted for potential tailplane stall (the Saab 340). If he did apply an abrupt back pressure (as the NASA video suggests may be necessary for tailplane stall), he may have gotten the thing into a traditional stall spin (as wreckage seems to support). This is precisely why tailplane stall recovery is not widely disseminated in my opinion! It is not proven to be a "first course of action" when experiencing any subtle indication of a stall of any kind. Again, what is the training here? If we have a pilot familiar with these operations speculate..then, and only then, am I comfortable with the conversation here.

Posted by: eric hanson | February 24, 2009 8:23 AM    Report this comment

Mark:

I am not sure if you are agreeing with me or I have confused things. I stated the investigation reports from yesterdays flash indicated that icing was not the current line of investigation. I agree with you that if there is a tail plane stall the reaction is a rapid pitch down of the Dash without any control input. I looked at your link to the animation. Other than the high pitch attitude, it looks like an executed incipient spin entry.

Another interpretation would be a botched snap roll where incorrect control inputs were made resulting in an accelerated spin entry. There are many hypotheses here such as deep stalling the aircraft which is, some aircraft, not recoverable. Some reports indicated that the crew were, in the last moments, executing a recovery.

Simply too little room to effectively recover.

Posted by: Christopher Basham | February 24, 2009 8:33 AM    Report this comment

The following was clipped from an article provided above by Christopher Basham (a Canadian Dash-8). Note here that their control inputs included "holding the control column aft for 22 seconds without increasing power(sounds like this dreaded tailplane stall recovery technique to me). Control was subsequently restored only AFTER traditional stall recovery techniques were applied! Following is the article in part: "..The standard technique for a stall recovery is to immediately and simultaneously advance the control column to reduce the angle of attack, apply maximum power, and then level the wings when the aircraft has exited the stall. An ice-induced stall requires a recovery technique in which the control column is moved forward aggressively (altitude permitting) to reduce the angle of attack and trade altitude for airspeed. FDR data show that, after the stall, power remained unchanged and the control column moved aft of its pre-stall position for about 22 seconds. The aircraft exited the stall when the control column was subsequently moved forward."

So here is evidence of pilots losing 4200 feet with stick aft and no power increase! Sound familiar? Yes, the NASA video (and their tailplane stall recovery technique)! These guys almost screwed the pooch because they didn't follow the "basic minimum knowledge provided in the PTS". Put that in yer pipe and smoke it!. Arrogant disregard for "basic" training kills more pilots than a lack of reading Advisory Circulars.

Posted by: eric hanson | February 24, 2009 9:07 AM    Report this comment

Eric:

Indeed it all comes back to training. A couple of interesting points...to my knowledge virtually nothing was known about tail plane icing until the NASA research which was the result (I believe) of the Roselawn ATR crash and the subsequent crash a year or so later in Detroit. Amazing that such a phenomenon took so long to be recognized and researched. It is generally known that ice accreates faster on smaller surfaces and that the tail plane is always smaller than the main wing (thickness). For planes that are likley to be at risk of tail plane icing does this mean than the tail will generally stall before the main wing? If so then the standard recovery would always be a tail plane recovery. In which case the 3407 PIC's move to the DASH 8 (apparently less susceptible to tail plane stalls) would be a bit like an American driver having to react suddenly while driving on the left side of the road in England. Not pretty. Training and experience being the key.

Posted by: Mark Hangen | February 24, 2009 9:11 AM    Report this comment

Lessons I take away from this discussion are that total awareness is critical. I believe that the PIC somehow overlooked the fact that the deicing system would result in a stick shaker at an airspeed that was as much as 20 knots faster than normal. Was this a training curriculum problem? Poor assimilation of training or just problem of workload and distraction? Who knows? A second lesson is that reaction without clear understanding of the problem you are trying to solve is almost never good. Perhaps a moment or two of reflection was in order here before reacting. I know they were close to the ground but the result of a pause could not have been any worse in retrospect.

I recall a low approach to KMCI on the ILS at night in a Turbine Commander. I expected to see the rabbit leading me to the runway. Instead my first visual clue was that I somehow was flying perpendicular to the rabbit. Instaspaital disorientation. First instinct to this "surprise" grab the yoke and yank the aircraft around (at about 500ft AGL)to line up with he ILS. Hand on yoke I stopped because it just didnt make sense..how could the autopilot make such a mistake? Turns out the "perpendicular rabbit" was actually traffic on the highway which so happened to run perpendicular to the runway. If I was an inexperienced instrument pilot I am not sure I could have resisted my impulses.

Posted by: Mark Hangen | February 24, 2009 9:16 AM    Report this comment

Chris:

I am agreeing with your post that ice wasnt the issue.

Posted by: Mark Hangen | February 24, 2009 9:19 AM    Report this comment

Eric wrote, "Put that in yer pipe and smoke it!."

Not sure who's pipe or what smoke, but that bit about the 22 seconds delay doesn't sound like THIS accident, rather an upset in a Canadian incident earlier. In THAT incident the deicers weren't activated, the wing was probably contaminated and caused a stall at a higher than normal speed. If deicers weren't activated, the 'increase Vref' switch obviously wouldn't be either--if the Dash 8 even was so equipped (dunno), so the IAS got slower than in THIS accident. Also, there was no mention of an installed 'stick pusher'. If one was installed, though, it could explain the fore/aft motion of the control column, since during THAT 4,200 foot descent the plane did stall three times and the pusher (if it had one) would have been expected to kick in and out--which would support my hypothesis more than yours.

So there. :)

Posted by: David Tuuri | February 24, 2009 9:51 AM    Report this comment

David, my comment was not meant for anyone specifically (..smoke it). The article Christopher links to indicates, possibly, an attempted tailplane stall recovery which lost 4200 feet. The article as I mention, says that the aircraft never recovered until the yoke was brought from its held position of 22 seconds (aft), and moved forward. Please reread my post(s) if necessary before commenting. We cannot kill pilots by confusing established stall recovery methods and introducing an extra thought process within a crisis moment. If an aircraft requires other than standard stall recovery, it should be recertified such that it complies with normal aircraft behaviors. Everyone commenting here needs to read these links to follow the conversation appropriately. This is not a competition amongst hypothesis here David. I have not submitted a hypothesis. I am merely critical of some proposed methods. We don't have enough facts to comment further here in my opinion!

Posted by: eric hanson | February 25, 2009 12:00 AM    Report this comment

Gentlemen be nice.

This discourse is an academic exercise with the most basic of information. We are all right in our approaches. Furthermore, we all have to wait the investigation report. There could be a significant change in either design or operations as a result. None of us know the outcome but we well all be impacted.

I have learned many things through the discourse about an environment that I will never be in (willingly). IFR means to me Infinite Funds Required. If I upgrade; it will be to a Pitts.

I recall the investigation into the Legionnaire’s disease. I remember a Doctor who went on TV to air his view of the cause. The old buzzard was our better and a survivor of world war 1 gas attacks (phosgene). He related that the observed symptoms are identical and the gas would be produced by leakage of Freon onto hot electrical parts. However, he did not have access to autopsy results or the cultures. He was wrong. Nevertheless, I will bet that the investigators were listening and tested for the gas.

Posted by: Christopher Basham | February 25, 2009 12:50 AM    Report this comment

Nice, comments Christopher. I apologize for any inflamatory comments of my own. I do think perhaps that an aircraft should be devoid of this tailplane stall charachteristic. Perhaps instead of endangering the majority of pilots on ambivalent stall recovery practices, we could consider improved aircraft certification particularly for known icing certification. If one needs to apply after some discerning of which type of stall it is, a recovery, than let it be known we have set back safety. I understand that money is involved, and FAA has a dual charge of promoting and regulating aviation, but we shouldn't allow operations of aircraft that are poorly suited to known icing.

Posted by: eric hanson | February 26, 2009 1:38 AM    Report this comment

Add your comments

Log In

You must be logged in to comment

Forgot password?

Register

Enter your information below to begin your FREE registration

« Back to Full Story