Mad Jock post at 1158 hits the nail on the head.

Tim Allan's observations / article about loss of power must now be critical reading and teaching to turbo prop crews.

I found, if memory serves, that zero thrust on Saab 340 is about 30 TQ.?
No doubt DB6 will keep me right !

Therefore any torque less than this zero thrust setting will create DRAG, and must be dealt with without delay.

Must be included in the simulator as a teach and practice. Auto feather failure is difficult to deal with even when you know it is on the agenda. Control and secure practice is essential.

Unfortunately not, you can get it while decending as well. It doesn't matter to be honest what your doing if your going down or up as soon as your speed comes back it bites.

The larger the angle of bank the greater the speed as well. Once you go past about 10 degrees the Vmca increases rapidly in a none linear manner.

Unfeathered Prop
 

I fly the Caravan - with an unfeathered prop it flies as if you have just jettisoned the wings - the rate of descent is horrific. Feathered, the thing flies on forever !

My type won't NTS or auto feather what ever you want to call it until it senses negative torque.

Which is 10-15% below the zero thrust setting/feathered

All transport category aircraft are certified to fly after a failure of one engine.

Your brother's CASA must have been too heavy for the airport temperature and altitude on that particular day. That is a performance calculation problem. It is not the fault of the aircraft.

There are many factors which can limit the performance of an aircraft. (overloaded, too hot, too high, improper Center of Gravity, etc) Some of these factors could be at play with this ATR-72 crash as well.

One of the most difficult things a pilot trains for is an engine failure right at liftoff. Some pilots are better than others at this maneuver.

The investigation may prove otherwise but my gut feeling tells me the pilot did not handle the aircraft properly following an engine failure.

thing is....

I have never heard of a turbine turbo prop engine failure that is anything like what you get in the sim.

I have had first hand storys from pilots not just reading the reports.

They have had rear compressor failures metal coming out the back EGT's through the roof.

They have had the spline couples go between the prop and gear box.

They have had the compressor suck a bird and start surging.

They have FEDEC's go nuts.

Engines that start cycling between 0 and 100% torque.

prop governors not governing.


But never a bang at V1/Vr and then the engine winds down all very well behaved.

Not much left of the aircraft forward of the wing, indeed even the wing is pretty much non existent. However this could be in the way they strapped the aircraft to get it out of the water

Surprised anyone survived the crash.

in flight Lo Pitch or ATPCS failure
 

I would like to present these two possible scenarios with ATPCS failed compounded with Eng Flame out.If not reacted to properly in the second segment climb..performance would be severely diminished, this added to posiible mishandling and eroding speed below Vmca...possibly leading to the loss of control we saw depicted in the video sequence.58 people is not a full load on an ATR 72 and for performance to be diminished like that Lo Pitch and or ATPCS might have played a critical role.The assymetric drag experienced during LO pitch if prop not feathered is worse than during engine flame out.It is a Memo item! ATPCS provides for(apart from torque boost on the live engine) auto feather on the engine suffering power loss.So if ATPCS failed...you take matters in your hands..and deal with it swiftly or brace!

its not even that warm there just now.

Min 14 deg max 18 in the last 36 hours.

Wind under 15 knots.

Just For info, if -600 is the same as -500 you don't need negative torque on the ATR but Tq< 18%, and prop isn't going to feather if you don't have the pwr management selector on T/O... like when you are above minimum acceleration altitude...

Regarding CRM: the pilot (according to news reports) issued a Mayday, therefore managed to graduate to the Communicate part of ANC.

Bearing in mind the timescale, apart from Sullenberger, is there a record of this happening in similar circumstances?

I-NAVV what is your zero thrust torque?

Crappy aircraft
 

In my view the ATR is a huge compromise. Underpowered and poorly designed. How they got the "double pull" procedure approved is really beyond strange.
The only benefit of the ATR is that it gets 70 ish people into the air as cheap as possible. Compared to the SAAB 2000:s 4152 SHP the ATR 72-600 is, with it's 2475 SHP, quite underpowered. Bear in mind that they have roughly the same MTOM (ATR 200 kg lower).

Some ATR driver might also shed some light to the engine problems the ATR fleet has had in recent times. Something about massive propeller vibrations?
Maybe the cause of this accident?

For simulated oei ops in manuals is reported only a Power lever angle 39* not a torque, but if you consider that for a 3* glide slope you fly around 22 Tq roughly (not sure, I don't follow the Tq a lot like other colleagues) I don't consider a TQ above 18% an high drag setup...

More than likely because it isn't and they chose that number for the auto feather to trigger because to go below that you would be into the drag zone.

Whats flight idle?

I know nothing about the ATR, but, can these engines experience a runaway torque conditon like the PT-6 in a King Air? Maybe the left engine was OK, while the right side was screaming past redline? That scenario in King Air sim is one of the hardest to recognize and handle properly.

I-NNAV
 

If you have flown the ATR you will understand the meaning of Lo Pitch fail and ATPCS failure impact on take off performance.Swift manual feathering would have been required to comply with second segment climb.It might be whilst trying to do that and arrest descent rate they eroded speed and fell below Vmca..that is the scenario am eyeing upon and hope when reports are released to confirm or cofound it.Any ATR 72-600crew in here please can you shed any ligh t on this.

Dispatch with ATPCS u/s or off requires:
Bleeds off

Increase V1 limited by Vmcg by 5kts
Incr Vr by 2kts
Increase Vmca by 3kts
Decrease max 2nd segmet weight by....3600kg in case of ATR 72
For Approach increase Vmcl by 3 kts

So if any of the above were breached in an ATPCS unavailable scenario you can see the performance impact.

recovery from a stall or abnormal roll control
 

Control wheel..Push firmly
*if flaps zero(extend flaps to 15)
*if flaps 15 (likely case):
Pwr management ...TO
CL/PL......................MCT
ATC........................notify (May day! in this case...cose to the ground).

Note..This procedure is valid whatever the LDG GEAR position (down or up..in this case UP)

Note that the narrative is based on ATR 500 series.I wonder what the diff would be with 600 series checklist.

340drvr and prop overlimit
 

There is a short in flight troubleshoot procedure after retyrding TQ to about 84% and see if the Np increased is below 101% (in whic case continue flt normally

Or if Np remains. Above 101% and conditions permit..shut down the affected engine and follow S.E. ops QRH procedure.

*red limits must not be deliberated exceeded.

I don't think this was the scenario.

Auto feather tq triggers
 

For PW121A eng. It is <18%
For PW127E/127M is is <21%

FI PL position is 37 degrees PL angle. The thrust will depend on Atmospherics and EEC regulation that moment.

Once you have past the acceleration altitude and set the power management from TO to Climb you do not have anymore the protection of the ATPCS.
On the -600 you just move the rotary switch from TO to Climb and that's it the climb sequence is over.
Should you loose an engine just at this moment it can be very tricky as you need to recognise it quickly enough to have time to feather it manually.
Furtermore the -600 has in my opinion some interface ergonomy can be confusing and with stress I am not sure if the crew could not be confused by some PFD indications instead of reverting to "basic airmanship".

We will see what it comes out from the inquiry....

The ATR is quite underpowered, single engine in optimum conditions is as requested for certification.
In actual OEI conditions, especially with a windmilling prop, I have some doubts if it can maintain level flight and speed.

Tracking shows the aircraft climbed to at least 1250 feet.

On the initial takeoff it was still on rwy heading and by the end of the rwy at 300 feet and 113kts.

At one point after the rwy end it makes a sharp right turn while climbing to 1250 feet where the speed is now at 83kts!

It further drops to 81kts while descending and eventually gets to 85kts before the impact.

For those familiar with the ATR 600, what is the Vmca, flaps up or in takeoff position?

The wing stalled....that seems pretty obvious to many here. It was certainly not a deliberate roll to the left, the aircraft simply isn't capable of such a rate of roll.

For those who favour the VMCA theory I strongly disagree. Reducing IAS through VMCA would never produce such a high rate of roll. For certain one wing stalled, and quite likely very shortly after so too did the other one.

A tragedy, and for those fortunate enough to survive, also miracle.

@ deefer dog...
 

Well a Vmca departure could also lead to a wing stalling if you try to keep the nose up and try to counteract with ailerons plus spoilers.

I know that with full power on one engine only if you allow the speed to go below Vmca on a Dash 8 it will roll into the dead engine pretty fast.

Of course if you keep bleeding the speed back then the wing will stall most likely the one with the failed engine first because of the prop wash on the other wing.

In this case it could be that they initially thought they could make the opening on the other side of the highway but at the last minute saw they could not and added power (if it wasn't always at max power) and were too slow and hit Vmca.

Anyway with the black boxes retrieved, they will know.

Obviously an aircraft not under control both in direction, speed and altitude.

http://i17.photobucket.com/albums/b7...pskxwc2wew.png

I am trying to find online that article of Tim Allan's. Some is asking for it.

I think everyone up north was pushed towards reading by a certain exam callsign and Flight ops inspector.

It really is a superb article for all multi engine prop pilots.

It was in the operator safety magazine and it was written when Tim was still a test pilot for them which I reckon was maybe around the 1995-2000 period.

Now the perfect thing would be someone nice from BAe regional that reads this to get authorized for public viewing and online somewhere. It really is a superb training document and would aid flight safety to all.

I think my copy of it went to the wind a few years back with a hard drive failure. And anyway I wouldn't post it on line as understandably its copyrighted.

I have been on the ground for most of the day at the crash site. I'm not going to offer any speculation as to what caused the accident, but I'd like to say that the Taiwanese emergency services have done an absolutely stellar job today with the rescue and recovery operation.
It has been a bad 12 months, particularly for Asia, for aviation. I am hoping that this year will see an end to this peculiar run of aviation related tragedies.

MJ,

copy of the article here. It should be compulsory on each type rating. Starts pg 9

http://www.availableaircraft.co.uk/I..._-_Issue_1.pdf

https://www.facebook.com/video.php?v=1053519784674205

USe that video on HD and full screen. Its not feathered and still turning.

Regarding the screen shots on page 1:

Is it perspective/elevator position, or is there a piece missing from one of the horizontal stabilisers?

If so: debris from failed engine taking out control surfaces?

EDIT: it's most visible on the first, the second has contamination from the impact and light gantries, making it artificially appear that the lower side is longer.
EDIT2: I suspect the length difference is a perspective issue due to the position of the stabilisers on the T-tail. Not sure if this totally accounts for the difference in shape though?

Here we go again. The company are going to review their safety procedures. What about a total review of training procedures and experience of crew reviews.

The authorities and companies don't seem to have any idea of dealing with the issue. Just awaiting more stringent and regular medicals and university degrees as a remedy,

Where is ICAO when you need them??

The article isn't the full article about the dangers of trying to nurse a engine home when its delivering less than zero thrust ie not feathered.

It looks like the time period 1995-2000 is correct though.

"The wing stalled....that seems pretty obvious to many here. It was certainly not a deliberate roll to the left, the aircraft simply isn't capable of such a rate of roll.

For those who favour the VMCA theory I strongly disagree. Reducing IAS through VMCA would never produce such a high rate of roll. For certain one wing stalled, and quite likely very shortly after so too did the other one.

A tragedy, and for those fortunate enough to survive, also miracle."

I'd have to disagree with you on this one, Asymmetric roll over happens pretty quickly

https://www.youtube.com/watch?v=YqmomTUVsAw

Look at the graph in the post #104. I think there is some answer. They lost an
engine, but managed to climb to 1300 feet in one minute. That seem to be an excessive climb rate. It seems some quite excessive pitch for being on one engine. It seems the speed for single engine was not taken care off. I am not sure, what is the single engine climb rate in an ATR. But 1300 fpm seem to be high. They went below the safe speed and it goes downhill from there. Of course it is just a theory. Who knows, what else could cause all this. But if it was not a control difficulty problem, than it seems to be overcontrolled.
Sad day.

In the climb phase, especially initial climb, there will inevitably be a lowering of the nose to accelerate. This is what the level off height or third segment is all about. Lower the nose, accelerate, clean up. Even for enroute climb, if you want to maintain the same airspeed, what are you going to do. Cruise altitude and you need to drift down, how are you going to do initiate the drift down. If you are more interested in airspeed(which is the important thing) how else are you going to maintain it or increase it when required.

The ATR does have roll spoilers to augment the ailerons. While not nearly as significant as an MU-2 which has no ailerons, adding significant amount of roll input will raise the associated spoiler reducing lift on that wing. Same with many large jets. I have seen someone in the sim use lots of aileron(with spoiler deflection) in a multiple engine out scenario and it just contributed a continual loss of airspeed(for that type and scenario).

If there was an engine failure with no autofeather in this accident, six-bladed props might make things worse as compared to the older 4 blade engine aircraft.

go123.
That manoeuvre in the video is a spin, not a Vmca loss of control.

The video initially shows a wings level descent which indicates the a/c was under control, but seriously short of power over hostile terrain.

Fuel not getting to the engines would be more likely than two near simultaneous
mechanical failures.

I can understand the crew desperately needed at least one engine back, but may have been too slow to maintain control if an engine suddenly produced thrust.

One can understand why they would stretch the glide as the terrain was as hostile as it gets.

All plausible scenarios, Vmca excursion, Dual flameout, or incorrect securing of a non-failed engine.

I'm tending to think they had lost power on both power plants, because if you look at the speed/graph posted, when the aircraft stopped climbing, it was in a very steady descent, with mostly constant speed.

Seems to me that badly mishandled asymmetric flight would at least provide mostly level flight, or at least a shallower descent. That last minute pull was an automatic reaction to the pilots seeing that they were gonna hit the bridge or power lines. I don't think it was a Vmca roll, it just doesn't seem to have enough yawing and the wing drop looks more as a result of a stall.

Either way, the recorders look in great shape so we should have all the data soon and I'm sure the preliminary report will be quite expedient due to this airline's safety record. This will be a very politically charged event no doubt.

As a boy, I saw the Breguet Atlantique crash at Farnborough in September 1968. It looked very similar.

The altitude reached in the graph suggests that the engine failure may have occurred after the after take-off checks were commenced, one of which (on the turboprop I fly) is to turn off the autofeather (is it the same on the ATR?). After that the prop will windmill, and when instructing in the sim I occasionally see pilots forget or mis-handle the windmilling prop. they still expect the prop to feather.

The graph suggests that the pilot maintained the best speed he could to avoid the stall after the failure. In the final moments he was left with no option but to pull the control column aft to avoid the buildings / bridge and reach the river. The 'g' and speed reduction from this action likely caused a stall and Vmca at the same time.

We do exercises in Vmca in the sim on initial courses. We set up the conditions under which the aircraft Vmca speed is certified; MAOM, aft CG, critical engine failure, take-off flap, gear up, 5,000ft PA. If a gentle descent is maintained for demonstration purposes to allow a slow speed reduction towards Vmca then it can be docile maneuver, but the real world is never like that. If the pilot concentrates on maintaining altitude during the exercise then at Vmca the aircraft will roll very rapidly to the failed engine.

Vmca is generally close to the stall speed. Maintaining aft control column and max rudder deflection at this speed has a name: Full pro-spin action.

As mentioned at #109, I was studying the horizontal stabilisers ("h/s") earlier and thought there was a section missing on the port side.

Bearing in mind the t-tail design which distorts perspective, it's not so much the length, rather than the shape.

BBC News - Taiwan TransAsia plane crashes into river

On the third clip (taken by the car in the right lane) at 18-20s, have a look at the port side. I confirmed with a bit of modelling tonight, that the end of the stabiliser should graduate away bearing in mind perspective and the angles the plane travels through. Instead, it's 'blunter' than it should be; the tapered end part appears to be missing.

So, if part of the h/s had gone, some down force would be lost, the nose would go down, and the AOA of that wing would lower, possibly creating a spin. Does this fit the scenario?