Stall recovery procedures ATR 42

[LimpopoHippo]

My understanding of stall recovery procedures in general of any aircraft is that:
1. Reduce aoa
2. Minimize altitude loss by maximizing thrust available and reducing drag and climb as efficiently as possible while accelerating to a safe climbing speed.

I came across a recovery sequence that is specific to the ATR 42 as far I as I have been able to determine that advocates at shaker:
1. Maximum Power(Torque)
2. Check and set flaps 15( if clean at stall then extend flaps to 15 )
3. Reduce aoa

Can anyone shed some light on this sequence and enlighten me as to whether ATR 42's require special stall recovery actions?

[FougaMagister]

You won't find anything in the Emergency Recall Items or Procedures Following Failures - and I bet it's not in the QRH either (you would never have time to reach for it during a stall). It's one of these type-specific drills you have to know by heart.

You have to remember that the stall warner is designed to operate 5% above the actual stall speed, i.e. at the incipient stage; therefore, a quick application of full torque (taking care not to be too heavy-handed and go over 100%) can cleanly power you away from the stall.

As in any propeller aircraft, that creates extra lift due to the extra air flow over the wing - even before the airspeed actually starts to increase.

That would be the reason to start with the Power Levers, then looking after the config (the ATR 42 can climb away at flaps 30 on 2 engines) and only then the attitude (4. is: with a confirmed positive climb, gear.......... up). You will find that the first two actions might be more than enough to power you away from the stall, even before you reduce the attitude - but obviously good airmanship recommends completing the drill regardless!

Cheers :cool:

[Mad (Flt) Scientist]

My understanding of stall recovery procedures in general of any aircraft is that:
1. Reduce aoa
2. Minimize altitude loss by maximizing thrust available and reducing drag and climb as efficiently as possible while accelerating to a safe climbing speed.

As a philosophical point, the way that second point is phrased makes the hairs on my neck stand on end. "minimise altitude loss" first? That's a recipe for stalling again. You've got where you are by being short on energy (speed). The escape route is to add energy, which you only have two sources of - height and engine power. If you concentrate on the height maintenance aspect you just gave up one of your two sources of energy. Yes, I know people hate to deviate from their assigned level, but you'll deviate a damned sight more if you actually do stall.

Please - reduce AoA, increase speed, THEN recover altitude.

[FougaMagister]

ATR 42-300 FCOM part 2; Operating Procedures

2.02.12, P3: STALLS

(...) "Recovery of stall approaches should normally be started as soon as stall alert is perceived: a gentle pilot push (together with power increase if applicable) will then allow instant recovery. If the stall penetration attempt is maintained after stall alert has been activated, the STICK PUSHER may be activated: this is clearly unmistakable as the control column is suddenly and abruptly pushed forward, which in itself initiates recovery.

Note: the "pushing action" is equivalent to 40daN/88 lbs applied in 0.1 second and it lasts as long as angle of attack exceeds the critical value".(...)


Rather than a "gentle push" on the control column (a term which in my opinion might lead to excessive altitude loss for the heavy-handed), simply relaxing back pressure should in fact be enough (certainly if close to the ground, in the circuit for instance). The recommended procedure, while broadly similar to the FCOM technique, may actually differ slightly from one operator to another.

Cheers :cool:

[False Capture]

LimpopoHippo,
Are you sure the recovery procedure you discovered wasn't one used before the Roselawn incident involving the American Eagle ATR72?
I was taught (and continue to teach albeit to ab-initio sudents) that you SIMULTANEOUSLY reduce AoA and apply full power.
I have a real problem with retracting flap before "reducing AoA". This was one of the factors which lead to the unfortunate loss of the AA ATR72. When I was an ATR skipper we often practised icing and stalling encounters in the simulator, our recovery procedure was the one you are familiar with ie. initial actions being the simultaneous use of pitch and power.

[cajun30]

"I have a real problem with retracting flap before "reducing AoA". This was one of the factors which lead to the unfortunate loss of the AA ATR72. When I was an ATR skipper we often practised icing and stalling encounters in the simulator, our recovery procedure was the one you are familiar with ie. initial actions being the simultaneous use of pitch and power."[/quote]

False Capture, no one mentioned retracting flaps during a stall. The ATR 42/72, at least at American Eagle, calls for flaps 15 when practicing a CLEAN stall. Nothing to do with the Roselawn crash. Those guys were holding with flaps 15 making a big snow cone on the fat wing.

Guys, you're over analyzing this, adding power while reducing AOA and CALLING FOR FLAPS 15 TO THE NON FLYING PILOT are done simultaneously.

Why add flaps 15 to a clean stall you ask Limpopo? Who knows, maybe because the whimpy PW 120 engines can't produce excess thrust to power out of the stall without substantal altitude loss. Bottom line,follow your company's procedure.

[Miserlou]

Maybe increasing the flap setting increases lift and decreases AoA. You know, the way the aircraft stalls at a lower speed (all things being equal) with flaps than without.

[Mad (Flt) Scientist]

Maybe increasing the flap setting increases lift and decreases AoA. You know, the way the aircraft stalls at a lower speed (all things being equal) with flaps than without.

The problem is the time it takes to restabilise.

Let's say you're at shaker speed and aoa, clean configuration. If you then select a lower flap configuration, the flaps will change the lift (increase it) and pitching moment (probably nose down) but will not DIRECTLY change the AoA.

For most wings, the stalling AoA is actually lower with flaps deployed, so if you were almost stalled clean and then put the flaps down you are probably now closer to stalling in the short term due to your action.

It all depends how slow the flaps are, and what other actions you take; if the flaps move slowly and you also push (or relax force as indicated above) it may be that you get the AoA down fast enough that the margin to the stall for the new flap is maintained or even improved.

But if the flaps are quite fast and you don't do much else to get the AoA down you may end up stalling the wing BECAUSE of the flap increase.

[Miserlou]

But do remember to compare like with like, Maddy. And by the way, how is more lift created and a pitching moment exist without the AoA being affected?

It should be remembered that selecting flap is changing the coefficient of lift of the lift equation, which is not normally considered.

The AoA for one speed will not be the same as the AoA for the same wing at a different flap setting. It may, as you say, have a lower stalling AoA than when clean but as long as the speed at which this occurs is slower than the clean stall speed the net gain is in favour of a quicker recovery.

In any case, the critical angle of attack clean is reached before the that with flaps. Hence, adding flaps reduces AoA.

[Mad (Flt) Scientist]

But do remember to compare like with like, Maddy. And by the way, how is more lift created and a pitching moment exist without the AoA being affected?

The lift and pitching moment changes at constant AoA because the flaps have moved. All aerodynamic forces are a function of the angles of attack and sideslip, the speed, and the CONFIGURATION of the aircraft.

It should be remembered that selecting flap is changing the coefficient of lift of the lift equation, which is not normally considered.
The AoA for one speed will not be the same as the AoA for the same wing at a different flap setting. It may, as you say, have a lower stalling AoA than when clean but as long as the speed at which this occurs is slower than the clean stall speed the net gain is in favour of a quicker recovery.
In any case, the critical angle of attack clean is reached before the that with flaps. Hence, adding flaps reduces AoA.

You're concentrating on the steady state case - i.e. after recovery from the high AoA condition - where I agree that having flaps deployed will result in being able to fly slower safely than clean. But the immediate effect of deploying flaps is not to affect AoA except through the aircraft response to the change in pitching moment.

To illustrate (arbitrary numbers)

Clean wing : shaker at 14 degs AoA, stalls at 16
Flaps down ; shaker at 12 deg AoA, stalls at 14

Now, because the flaps down case generates much more lift at a given AoA, even though it stalls earlier (in terms of a lower AoA) you can still fly at a lower speed before stalling because the CL is much greater.

If, however, you find yourself flying at shaker speed with a clean aircraft i.e. at 14 degs AoA and you were to change the configuration quickly, without also acting to reduce the AoA, you would find (1) that you were suddenly generating more than 1'g' worth of lift and (2) almost instantaneously later, that you weren't generating much lift at all, as the wing would stall.

The 12-14 and 14-16 numbers are arbitrary, and some types may find the AoAs overlap differently. But for the general case, the AoA for stall warning and for stall get lower as flaps are deployed.

[Miserlou]

What you write cannot be true or the American Eagles guys wouldn't have died (along with the passengers) at Rose Lawn.
For the purpose of this discussion the fact that the aircraft was well iced up is irrelevant. The point is they had flap 15 selected and stalled and lost control when they raised them. They'd have had a chance if they had reselected the flap.
Ergo, the AoA of the wing with flaps extended is less then the AoA clean or it wouldn't have stalled.

And I wasn't talking about steady state cases. All parameters will begin to change during the flap travel until reaching the new steady state case values.

[Mad (Flt) Scientist]

Im not 100% aware of the chain of events at Roselawn, but unless the aerodynamic behaviour of the ATR72 wing is very unusual, then the general statement that the AoA at which one stalls is lower with flaps deployed is generally true.

Since the ATR was, if memory serves, lost due to seperated flow ahead of the ailerons - only - and the wing was essentially clean ahead of the flaps, I have some doubt as to the validity of the assertion. In fact, did not the crew at one point regain lateral control following the flaps 0 selection, but at low angle of attack, only to then pull back up into the stall? In which case the dynamics of the accident shed little light on the relationship between stall AoA flaps 0 and down (15?) since the stall occurred after the transition had occurred some time before.

Or, if you wish to go to a different source...

Contrary to popular belief, the stalling angle of attack when flaps are extended is less than when flaps are retracted.

and

Extending the flaps also increases the angle of attack of the wing for a given pitch attitude.

from High Lift Devices By Dave Esser (http://www.erau.edu/er/newsmedia/articles/wp1.html)

which together means that you get a higher actual wing aoa (though the aircraft aoa doesnt move of course) and lower stall aoa when you extend the flaps initially. hence, closer to the stall.

[FougaMagister]

The ATR 72 was lost at Roselawn, Indiana, because it stalled. Then again, it stalled in the first place because the aircraft was flown in the hold, in freezing rain (i.e. severe icing), a condition which the crew should have tried to leave asap, with the wrong flap setting and without having selected level 3 anti-icing. In other words, they got it seriously wrong...

Within a few weeks, ATR sent an ATR 72 to Edwards AFB where at the FAA's request it was subjected to severe icing flight testing - including behind a KC-135 which hose was spraying water in icing conditions on the ATR's airframe. While flight qualities were in some cases affected by ice accretion, the aircraft continued to handle as advertised - and the FAA was satisfied enough to lift the restrictions it had placed on operation of the ATR following the Roselawn accident.

It is worth considering that the ATR would not be such a popular choice in, say, Canada, Sweden, Switzerland, Italy, Germany, Denmark, New Zealand, UK, Poland, etc. if it had been proved to be less than safe in icing conditions.

As with all turboprop aircraft, it does requires careful handling in icing (if only because unlike jets, it may have to operate in the icing level as opposed to flying over it), but provided the icing procedures are strictly adhered to, it will handle as advertised.

Cheers :cool:

[False Capture]

cajun30,
You said "False Capture, no one mentioned retracting flaps during a stall."
In the recovery procedure which LimpopoHippo is questioning, the second step of the procedure is:
"Check and set flaps 15( if clean at stall then extend flaps to 15 )"
If you had flap30 selected then this procedure requires you to select flap15 (ie. retract flaps) before reducing AoA. In this case you would be retracting flaps during a stall.
Most of the discussion on this thread has dealt with the clean scenario and the pro and cons of extending flaps to the flap15 position. I was thinking about the flap30 scenario and I should have made that clearer - sorry.:uhoh:
At the end of the day, I disagree with the three step recovery procedure which LimpopoHippo has brought to our attention.

[Stuck_in_an_ATR]

As for flaps 30 stalls, we are taught to retract flaps to 15 when speed reaches the "yellow bug" (the minimum speed for flaps 15) during recovery. The "Max power, check power, flaps 15" procedure is for clean config only... Re. the stall AOA with and without flaps, the ATR FCOM states stick shaker activation for flaps 0 at 11.6 deg AOA and for flaps 15 at 11.3 deg AOA - not much difference there and the additional lift generated by flaps moves you immediately 10-20 kt from the stalling speed. And last, but not least - I had the opportunity to practice this tehnique in the a/c yesterday (of course it was a training flight, no pax on board! :} ) and it works at advertized - you can recover with no altitude loss at all. As one of my former instructors said: "It's better to clear the trees at 80kt, than hit them at 100kt" :)

[Clandestino]

The ATR 72 was lost at Roselawn, Indiana, because it stalled. Then again, it stalled in the first place because the aircraft was flown in the hold, in freezing rain (i.e. severe icing), a condition which the crew should have tried to leave asap, with the wrong flap setting and without having selected level 3 anti-icing. In other words, they got it seriously wrong...


Actually the one who got it wrong is you, FougaMagister, but don't worry you're in good company of 99% of general public and at least 70% of pilots who remember the Roselawn.

Contrary to popular misconception, ATR lost at Roselawn never stalled!! Neither was severe icing present, it was light to moderate. Although with hindsight it seems that crew dug out a trap for themselves, you can't blame them for crash because supercooled large droplets (SLD) that contributed to crash were completely unknown at the time and neither manufacturer nor certifying authorities were avare of effect of very unusal ice accretion pattern on airplane controllability.

ATR-72 was bound for KORD and was holding at FL100 in cloud with some light accretion. Levels 2 (electric heating of props and flight control surfaces horns) and 3 (boots on wing and tailplane) de-icing were turned on. During hold capt. felt that deck angle was too high and airplane "wallowed" in turns so he selected flaps 15° to decrease it. Now the airplane was flying at low AoA and speed through area of SLDs which hit the wing and then froze behind the deicing boots (and out of crew sight), forming ridge of ice. All went well untill they were cleared to FL80. During descent overspeed warning sounded, flaps were retracted and control wheels snapped immediately to full right-wing down position. It seems that ice caused aileron hinge moment reversal - that is instead of centering they moved to full deflection. It was estimated that force of 50daN was required to center it. ATR knew that reversal occurs at 26° AoA but as the wing would stall before they didn't think it was significant or that it can cause problems. Plane rolled rapidly right and dived. During recovery tail and outer wings separated at 350+ kt and 4G.
If the crew didn't select flaps15°, if they didn't retracted it during descent or if ATR had hydraulic ailerons they would probably be still around but then we'd know nothing about SLDs and I suspect ATR pilots are the only ones informed about them.

After crash, wings boot were modified on all ATRs so now they cover area were ridge formed. Also two sets of sppeds were introduced - for normal conditions and conditions susceptible to icing. Second set is higher so airplane is operated at lower AoAs. That brings us to original question and the answer is: flaps 15 is ATR standard stall recovery procedure but for clean configuration and icing acretion conditions alone.

I stongly reccomend Peter Garrison's "A question of perspective" article, dealing with both Roselawn crash and accident reports made about it. It was published in Flying magazine somewhere about 1997. but I'm sure it can be found in his "Aftermath" book.

[FougaMagister]

Clandestino - thanks for the details of the Roselawn accident.

I had forgotten to mention that while the FAA and ATR couldn't reproduce the conditions of the Roselawn accident in flight tests, the de-icing boots were subsequently extended in all new-built airframes (and retroffited to previously built a/c if I'm not mistaken).

Also, those who have flown the ATR will appreciate that one of its (few) design shortcomings is the under-sized ailerons; while that does not affect handling during normal operations, it certainly does on one engine (white-knuckle time!) and also probably did during the sequence of events at Roselawn (with the bank angles you mention).

Cheers :cool:

I hope Peter Garrison's book is not the one I perused through ages ago and which blamed 100% of the crash on the ATR's design with a "if it's not made in the USA it's got to be rubbish" approach...

[Stuck_in_an_ATR]

By the way - I've read about the Roselawn accident, that when the aircraft started rolling, the crew selected flaps 15 again, but the MFC logic didn't authorize flap extension due to overspeed. Does anyone know if this overspeed protection has been removed from the MFC since?

[Dick Whittingham]

Just asking. Could the 15 flap be to get the extra nose down trim that comes with flap extension? If the ATR has a stick pusher this means that it has undesirable basic stall characteristics and needs a pusher for certification.

Dick W

[Mad (Flt) Scientist]

I don't think you'd be able to take credit for a flap configuration change in demonstrating acceptable stall characteristics for certification, unless you actually automated it to be part of the pusher system somehow. Certainly I wouldn't try that with my authority!

I suspect it's simply that the L/D is still pretty good so the drag penalty isn't much worse than at zero, the AoA penalty seems to be small, and it also isn't too fast a translation for the flap, so by the time you get to flaps 15 you've dealt with the immediate AoA issue and it now gives you a bit more lift/margin to play with.

[hocuspocusfidibus]

At school they tought me to:

decrease aoa;
increase power;
roll wings level.

Why didn't anybody of you mention wings level??

Excuse me for the question?

Hocus Pocus Fidibus

[rhovsquared]

It's not the stall i'd woory about on the ATR is the ensuing spin:} :} :}

[SIDSTAR]

On the A320 the stall rfecovery sequence (if stalling clean) is TOGA Thrust, flaps 1. Slats run out quite quickly and VLS (lowest selectable speed) drops by about 8 - 10 kts due to effects of slats. Hold the attitude and you lose no height.

[Stuck_in_an_ATR]

There's an urban legend at our outfit, that in the early days of the ATR some guys tried stalling it with the stall protection systems off during a test flight. To cut the long story short, they had to change their pants after landing :}

[Clandestino]

Maybe I didn't make my point clear enough the first time around:

For purpose of stall recovery, flaps 15 are to be extended in ice accretion conditions only. It serves to increase stall margin.

Also spracht ATR FCOM

[error_401]

Just to make sure that everybody gets the message about the Roselawn:

Instead of bitching around you all should rather go to the source (the finest IMHO which is the NTSB database and get the reports straight)

BTW: It clearly states that they encountered water droplets the size and bla bla which clearly exceeded the certification standards bla bla - In short it means they probably were in SEVERE ICING - quote.

Now the references:

Factual report - here:
http://www.ntsb.gov/ntsb/GenPDF.asp?id=DCA95MA001&rpt=fa

Probable cause - here:
http://www.ntsb.gov/ntsb/brief.asp?ev_id=20001206X02420&key=1

Accident brief - here:
http://www.ntsb.gov/ntsb/GenPDF.asp?id=DCA95MA001&rpt=fi

I'll be back with the full narrative and if I can find it again with the know-how they got from the tests at the NASA icing branch.

error_401

[Nick NOTOC]

My two pennies worth:

First of all stall (exceedence of AoA) is the first and foremost danger, recover that by reducing your AoA. Then you will find yourself with your seccond problem and that is altitude loss, for this we may (or may not if we're really unlucky) have engine power, this power (energy) is then converted into speed. To make this more effective let's get rid of unnessesary drag, so flaps to 15 degrees.
The speed at which you perform these tasks depends on your situation. If you have lots of altitude, use it in your recovery, if not I would be more power minded.
Last but not least, icing: No aircraft is certified to fly into icing conditions for a prolonged period of time whenever the supercooled droplets exceed 40 microns. So since I cannot measure the droplets, my way of operating when picking up ice: GET OUT OF THERE!
Use anti and de-ice to buy time, but either climb (3000 often is enough to get above ice accretion) or descent (if safe enough).

Good luck to all ATR drivers in the Wx

Nick

[Clandestino]

It clearly states that they encountered water droplets the size and bla bla which clearly exceeded the certification standards bla bla ...is what the NTSB and Bureau d'Enquettes-Accidents said, but.. In short it means they were in f:mad: SEVERE ICING is your and BEA's interpretation, with which I can not possibly agree.

According to FAA, during severe icing the rate of ice accumulation on an aircraft is such that de-icing/anti-icing equipment fails to reduce or control the hazard. Immediate diversion is necessary. So was it present at Roselawn? No - boots and heaters dealt with ice quite nicely, the ridge of ice that proved to be fatal formed on unprotected area behind the boot.

Hindsight is always 20/20 but does anyone think that crew should have considered routine hold with some icing as potentialy dangerous, when even designers and certifying authorities didn't know it was at the time?

[error_401]

Clandestino,

That is exactly the danger with the ATR because the anti- and de-icing equipment is doing a terrific job. Nevertheless in icing conditions more than light you will start to accumulate ice all over the plane including the underside of the belly the pack inlets antennae etc. I'd like to know how much icing the boots on the ATR could really take... and make you believe everything is fine.

I have not been able to find your reference on FAA severe icing elsewhere than on the NOAA page.

The ridge that formed behind the boots on the Roselawn ATR may be strong evidence that they were flying in heavy icing conditions.
The real danger lies exactly in this equipment being so efficient that the crew may think they are safe. But runback ice is always a problem. So I fully agree with your last statement that at that time they probably could not have known.

I'm happy that ATR had to change the design of the boots and other measures which make the plane even safer. What probably led to the fact that icing incidents and accidents have nearly been eliminated on the ATR may be the specific procedure changes calling for: Leave severe icing conditions immediately.


Nick NOTOC,

I fully agree and actually do the same. As long as you see ice accumulate get out of it as long as you still can. Unfortunately ice has the tendency to build up on the window frame, wipers, ice evidence probe to quite remarkable sizes but will also brake off from time to time suggesting less icing than actually present. I usually start a timing at first ice accretion and after 10 minutes I'll get nervous if the accumulation is still strong.