Note: This story was corrected on August 10th at 10:23 am, thanks to the help of a sharp-eyed reader.
Making an ATR-72 Spin
I wasn’t in Brazil on Friday afternoon, but I saw the post on Twitter or X (or whatever you call it) showing a Brazil ATR-72, Voepass Airlines flight 2283, rotating in a spin as it plunged to the ground near Sao Paulo from its 17,000-foot cruising altitude. All 61 people aboard perished in the ensuing crash and fire. A timeline from FlightRadar 24 indicates that the fall only lasted about a minute, so the aircraft was clearly out of control. Industry research shows Loss of Control in Flight (LOCI) continues to be responsible for more fatalities worldwide than any other kind of aircraft accident.
The big question is why the crew lost control of this airplane. The ADS-B data from FlightRadar 24 does offer a couple of possible clues. The ATR’s speed declined during the descent rather than increased, which means the aircraft’s wing was probably stalled. The ATR’s airfoil had exceeded its critical angle of attack and lacked sufficient lift to remain airborne. Add to this the rotation observed, and the only answer is a spin.
Can a Large Airplane Spin?
The simple answer is yes. If you induce rotation to almost any aircraft while the wing is stalled, it can spin, even an aircraft as large as the ATR-72. By the way, the largest of the ATR models, the 600, weighs nearly 51,000 pounds.
Of course, investigators will ask why the ATR’s wing was stalled. It could have been related to a failed engine or ice on the wings or tailplane.
The ATR has had several accidents related to ice over its 40-year history. The local Sao Paulo weather today indicated icing aloft was a definite possibility. If the ATR’s deicing system could not remove the ice the aircraft encountered at 17,000 feet, the wing or tail could have stalled. Worse yet would have been if the wing iced up and later shed only a portion of that icing load, perhaps on only one wing. A stall encountered under this condition could have initiated the rotation.
Then there is the possibility of an engine failure. Pilots who possess a multi-engine rating will remember the VMC demonstration during training. This maneuver demonstrates to the pilot applicant that things can go to Hell in a handbasket pretty quickly in a twin-engined airplane when one of those engines stops producing power. It’s natural for a twin to begin slowing following an engine failure. The remaining engine is then normally brought to 100 percent power. But as the aircraft slows, the flying pilot must add rudder pressure toward the operating engine to keep the aircraft under control. But … there’s only so slow that aircraft to fly with the good engine producing full power. Vmc, then, is the airspeed at which the aircraft – even with the rudder jammed to the floor – will no longer maintain wings-level flight. Without enough rudder, the airplane can begin to roll, unless the pilot pulls the power back on the good engine. Once the aircraft begins to roll, the flying pilot has only a few seconds to try and wrestle back control. And for those who watched the Vmc demo video thinking it didn’t look all that tough, remember this pilot knew what was coming. Voepass 2283’s crew would not have known what was ahead.
In most twins, because Vmc airspeed is not all that many knots greater than the aircraft’s stall speed, a spin is almost inevitable. Most twin-engine aircraft are NOT recoverable from a spin, no matter how the aircraft entered it.
The Crew
Investigators will also be looking closely at the two pilots responsible for flight 2283 on August 9, 2024 and anything that might point to a reason these two aviators lost control of their aircraft. What was this crew’s experience with the ATR-72 and how recent was that experience? What story might their initial and recurrent training records show. Also, what kind of duty day did these two pilot have? Was this their first trip of the day or perhaps their fourth or fifth leg. Were these two pilots familiar with each other or were they new and still learning the other’s habits. Did the weather forecast for this hour and a half flight turn out as expected, or might the pilots have encountered a surprise or two.
Any of these crew issues could have resulted in their not recognizing an impending stall and the loss of control that followed.
Again, I’ve outlined a few of the more obvious possibilities that this aircraft got away from the two pilots. There’s sure to be more to the story, so stay tuned.
Rob Mark, publisher
BTW – If you enjoyed this story and found it useful, please share it with a friend. And remember, a subscription to Jetwhine costs you nothing. Just add your e-mail in the box beneath this story.
PS – Thanks to my pal Max Trescott from Aviation News Talk for his help with this sad story.
Rob, is the statement “Most twin engine aircraft are recoverable from a spin, no matter how the aircraft entered it.” correct? I seem to remember from my multi-engine training that my instructors said that most twins are very difficult to recover from a spin. Just wondering. Thanks for the story.
Holy Smoke Mahesh – you caught a really big typo. Thank you so much for mentioning this. I corrected the text. That will teach me to try and send something out late at night.
Rob thanks for your quick analysis. When I saw that video I was horrified. The first ATR-42s had a funky de-icing system on the horizontal stab that was quickly replaced as I recall. Maybe you’re right to suggest an uneven ice shedding led to the spin, but whatever it was, that was a long, long fall. Awful.
I can’t imagine what it must have been like in the cabin during that fall Mal. Thanks for reading the story.
It sounded to me that the left engine/propeller had failed into a reverse thrust condition. The sound from the engine was as if the propeller blades were stalled.
Most twin-engine aircraft with engine failure ARE recoverable from a spin, especially from altitudes in the mid-teens (if you’re not over the Andes), by pulling the power to idle on the good engine and executing a proper spin recovery. In general, the main problem is that most pilots are not taught handling the aircraft at the extreme boundaries of the flight envelope. So if they get into trouble they don’t have the experience to recognize what must be done to reestablish smooth airflow over the wings and flight control surfaces. This airplane was likely recoverable unless structural failure occurred with the stall/spin.
We must have had very different flight training Jerry. When I penned this story I said most twins are not recoverable from a spin, exactly because of what you mentioned. No one anywhere – that I know of at least – would even attempt to train a pilot how to recover from a spin in a twin engine aircraft because the manufacturers do not flight test these aircraft under those circumstances. I was taught that if you spin a twin engine airplane, you just became a test pilot. Please do share your source about recovering twin from a spin. I think we could all benefit from that.
There is this YouTube video from a C90 King Air recovering from a spin. They got into a spin after dropping sky divers who were holding on to the fuselage, and then all jumped at once, prob with engine idle on that side. (Actually, the pilot got almost into a second spin, after avoiding hitting on jumper.) This was from 2021 from South Afrika. I read the airframe was bent. https://youtu.be/4H5JQ-ESc3M?si=INBM_O-iqwVJgv-P
Ah yes. Thanks, Stephan. Desperate times call for desperate measures no doubt. These guys were lucky.
From the video it is clear that this aircraft wasn’t just in a spin, it was in a flat spin. That is the worst kind of spin from which to recover the aircraft. It’s very possible that an ATR-72 simply cannot be recovered from a flat spin.
Re the below, of course, is accurate, but one might also note that as air density decreases, in many aircraft VMC drops below stall speed.
“In most twins, because Vmc airspeed is not all that many knots greater than the aircraft’s stall speed, a spin is almost inevitable. Most twin-engine aircraft are NOT recoverable from a spin, no matter how the aircraft entered it.”
That is a factor in many Baron training crashes. I don’t know if that’s the case for the ATR, but it could be a factor.
Matt
Matt
I knew Vmc could decline as the density of the air decreased Matt, but I was not aware it could drop to a speed less than stall speed. Thanks for mentioning this.
I’m 73, flew Army fixed wing in Vietnam and was in a flat spin during training in Alabama. We had 4 persons onboard the Beechcraft B55 Baron. I was in the back left seat. The instructor (warrant officer) was in the left front and a lieutenant student was in the right seat. Practicing left engine out (throttle retarded to simulate) stalls, at about 8,500 ft. The left wing stalled first, wing dropped, student applied full right engine power and full up elevator. The airplane entered a left turning flat spin. With the two pilots pushing and pulling and doing all sorts of things, the airplane started to dip nose down and we finally broke out of the flat spin at about 2,500. The terrain was approximately 1,000ft MSL. Not sure of exactly what got us out of the spin. The Warrant Officer instructor was discharged from the Army after this incident.
My buddy, sitting beside me in the back seat, said, “We just bought the farm” as we’re falling towards the ground.
Beechcraft and the U S ARMY, interviewed us individually. They asked me how did I know I was in a flat spin? Let me tell you, if you’ve been in one, you know! The representative from Beechcraft said that the Army Baron couldn’t get into a flat spin. He was proven wrong.
About two weeks later, on another training mission, another Beech Baron flat spun into the ground, killing all three on board.
No long after that, the training procedure was taken out of the Army manual. A chart was added showing how altitude and VMC were affected and crossed at a certain altitude. That chart may still be found in multi-engine manuals today. For the Beech Baron I was training in, it was approximately 8,500ft.
This all happened in 1970. As far as I know, my crew, having survived, may have been responsible for the VMC graph. Our friends weren’t so lucky. I attended their memorial service. It’s been over 50 years, but I remember it all like yesterday.
I may be leaving out some details, it’s been over 50 years. I graduated from flight school in January 1971 and went to Vietnam in June 1971.
As an anecdote… I think the Baron was a fine aircraft. Stall strips were added to the leading edge of the wings to help reduce the chance of one wing stalling before the other.
My personal opinion is that any aircraft with a near center CG, can flat spin given the right conditions.
BTW, I retired some years ago, from Delta Air Lines.
Y’all have a good day! :-)
Philip Lewis…
Thank you for your service and getting untold thousands home safety in your second career.
And, thank you for sharing your amazing “near death” experience – it certainly gave me a real world insight into the hazards of spinning twins.
As usual, we have been inundated by opinionated click-bait experts who “know” the cause of this event. Thanks, Rob, for providing a relatively thorough examination of the possibilities, free of any claims of having “solved” it.
It occurs that one more possibility is a level-off from a descent under autopilot, accompanied by a distraction at the critical moment leading to a failure to advance the throttles. Do ATR’s have auto throttle? I’ve seen it happen more than once in a variety of airliners-always caught prior to the stall.
No mention of the ATR crash outside Chicago on Halloween night many years ago.
I’m always baffled by these accidents and situations ! The first thing I was told in my initial training was: in any and all cases a pilot should first remember to “fly the plane”! I see to many cases where accidents could have been prevented by good hands on flying piloting including twin turbo prop aircraft. Yes there has been cases of spin in accidents with twin turbo props ,I believe one was shown on one of my favourite documentary shows “ Mayday”. The initiating cause of that accident (also in South America) proved to be the result improper manual de-icing instructions if I remmber. I gained my licence in Canada back in the 70s where full on spin and stall training was mandated by the DOT not theatrical. I know hindsight is 20/30 but it seams at least in some cases we need to take a stronger look at “hands on” pilot training no matter er what the aircraft size!
The FAA needs to put full stalls and Spin training and recovery back into Private Pilot ACS’s.
You can’t do this in a simulator or your iPad. If you got your license after 1983 there was no requirement to ever do a full aerodynamic stall in an airplane.
Spin recovery is the same in all airplanes. Power to idle on all engines. Neutralize flight controls and allow the airplane to turn into a “lawn dart” as the airspeed increases pull out of the resultant dive..
This was a flat spin. it takes power for it to go flat (right engine continued to make power). To say the airplane rolled is not correct. It initially yawed that cause the spin. This is the same deal as the Air France Airbus over the Atlantic years ago.
Improper training!
Rolled versus yawed is an interesting point Patrick. I would imagine the flight data recorder will eventually be capable of setting the record straight. A former F-18 pilot and a guy who has probably forgotten more aerodynamics than I know made another interesting comment on this story. “With aircraft having large fuselage mounted or wing-mounted engines, there is a very real risk of them being inertially locked in a spin (meaning it is likely that no amount of control input or differential thrust can decouple auto-rotation).
I didn’t mention it last week because there was no evidence that the two accidents were similar, except for the aircraft type.
Excellent story Phil. I’m glad you’re still here to share it. Another aspect of surviving spins in a twin-engine aircraft, in addition to what has already been mentioned here, is how the aircraft is loaded. With people in he back seats of a light twin, the CG might be so far aft that no amount of forward pressure on the control wheel will be enough to make the wing fly again.
I didn’t mention much about the flat spin investigation, but, it was determined, at the time, that we two passengers, being in the back seats, did contribute to the stall spin, moving the CG towards the aft. On previous flights, the Baron behaved normally, so CG should have been within limits. We probably did a W&B in the classroom, but I honestly don’t remember.
Onto my buddies death, the Baron cratered into ground, almost intact… no slide at all.
Besides terminating that training in the SOP, we were subsequently required to wear aviation type helmets, after that crash. As you can imagine, an absolutely unnecessary (fill in your own words) response to help in surviving a flat spin into the ground.
One item that no one has mentioned here is the fact that the ATR is a T tail aircraft. When this type of aircraft stalls, the wings block the airflow over the horizontal stabilizers so that they become ineffective. Stall recovery is very unlikely once this happens. I believe the first time this was discovered was during the crash investigation of the prototype BAC-111 in England many years ago. That aircraft also entered a flat spin and struck the ground in one piece.
Cheers
Great article and commentary! Thoroughly enjoyed reading all of it! Glad the FlightAware weekly newsletter featured it!
It appears this accident was due to weather condition, specifically icing. The conditions were very similar to the Chicago ATR crash years ago. As a result of that accident flight testing at Edwards AFB back in the nineties with the ATR and the EMB-120 revealed SLD (supper cooled large droplets) causes icing to form beyond the leading edge of the boots. This causes separation and loss of controls. I was directly involved in these icing test conducted behind a specially modified KC135. What was discovered was these conditions are difficult to detect and if the auto pilot is engaged it will cause airspeed deterioration resulting in a loss of control and in some cases un-commanded rolls. I suspect the crew was not properly trained for these conditions which are most susceptible to turbo prop aircraft which operate in the regime of plus or minus 5 degrees C of the freezing level. This will probably be determined to be the cause of such a loss of control. It probably was not a VMC loss of control but a stall due to SLD. I’m sure the Brizilian government will be looking at these test conducted in the 90s since they were part of out FAA test team.
What about the possibility of deep stall?
A common problem in some “T” tail gliders.
I don’t think the number of engines has anything to do with spin characteristics per se. With training aircraft (usually single engine), spins are induced by stalling the aircraft while simultaneously applying full rudder to induce yaw. The throttle would normally be at idle for this exercise. These aircraft usually enter with a nose down attitude. Standard recovery is to apply full opposite rudder and full nose down control stick. The aircraft will literally snap out of the spin and so one must immediately centralize the rudder and then gain airspeed to fly out of the stalled condition. With a twin, it may be that stalling with asymmetric thrust might provide the yaw to induce a spin. I would defer to test pilots here.
If the aircraft tends to enter a flat (as opposed to a nose down) spin, recovery becomes more problematic because there is no lateral airflow – hence the elevators and rudder become useless. You can’t lower the nose and you can’t stop the rotation. When I flew fighters, the only recourse was to deploy the drag chute to force the nose down. The only thing I can think of with a transport aircraft would be to drop the gear and flaps in the hope of causing the nose to drop. When you see the video and how fast the aircraft was falling (about 20,000 ft/minute), you realize how little time the crew would have to think through this.
I think this accident reinforces the need to program flight simulators to replicate real aircraft performance outside the “normal” flight envelope and to expose crews to this kind of abnormal condition during their recurrent training.
Billy T
This is a superb thread. Thank you to all for your keen insights. It is humbling.
Andy Hoffer
Former PPL, 20 yrs experience, C172
Very good and complete analysis from Robert, thanks.
But still curious to know about the data in the CVR and FDR, if made available some day.
Why no mention of the Speed & Altitude graph? The aircraft was level at 17000’ for an hour and the ground speed went below 100 kt four times followed by 300 kt plus GS. Also at least a dozen additional variations of over 100 kt. Was the aircraft circling in strong winds or was this a straight line flight?
Can a large airliner flat spin into the ground? Yup, I think so. While mishandling a stick=shaker recovery at FL250 in a simulator 757 rating ride, the ATP candidate entered a deep stall spin. As the DPE, I watched in horror trying to decide whether to freeze the sim and declare the ride unsatisfactory or let it hit the ground. Yes, I know . . . actual flight data is employed by the manufacturer Flight Test people to ensure simulator accuracy. Or, a computer-derived outcome. If I had been a test pilot, and my boss had said,”OK, today we do flat spin recoveries”, I would have run screaming out onto the runway. Anyhow, the candidate was pink-slip’d and returned to training.
PS . . . I froze the simulator at 100 feet AGL before the windshield visual went red with cracks and the crashing audio sounded.
I have research that tells me that this plane flew into a tornadic vortex. I have even authored a book about these plane accidents titled: Science About How Tornadoes And Vortexes Form And How They Are Causing Planes To Crash (Including MH370, available from Barnes & Noble and Amazon. There was a front coming through and a 101-Knot jet stream sufficient to create a tornado vortex. A vortex can even be seen on the ground on a satellite view. Just think what would happen to a plane if it flew into a tornadic vortex. Contact me if you wish to learn more. I’ve found over 480 such accidents the last few years. I have found on an average of one major accident per year, that is killing over 50 to 250 people in each accident or incident. My research can explain why planes are dropping thousands of feet and injuring people when the plane bottoms out.
Thanks for this comment, Randy. This is an interesting point you raise. I thought the shaker/pusher mechanics made it virtually impossible to stall a transport category aircraft, much less spin it. Can you share any info about the “mishandling” you mentioned? And I too crashed a Dash 8 simulator many years ago and remember all too well that red background and the crashing sound.
Thanks for your input. I do believe you and I spoke when I was still with Flying magazine Ronald.
I did add the graph to the story Todd, but it was probably too small for people to see. So unless the winds were varying rather insanely, you’re correct. On closer look, the groundspeed fell off several times before the spin. They were in trouble for almost a minute before the aircraft entered that spin.
You’re absolutely correct Etienne. I’ll be on the lookout for the reports of the CVR and FDR when they’re made available to the public.
It is humbling isn’t it Andy. I’m still numb thinking that all those people died because the crew was unable to cope with the situation for whatever reason.
I am a product of Army aviation 1956- 1973 fw & helo then Pan Am 747. Mr Lewis made me proud of his discourse
and the other Army pilot with “spin in a Baron” notes. They are trained in “little bitties”, but understand aviation.
Robert . . . the candidate overcame the stick pusher with back pressure and stomped incorrectly on the low wing rudder pedal as the wings were rocking. The way you and I trained spin entry in the old days. The candidate’s instructor was sitting in the rig;ht seat and was mortified and speechless. The spin developed in a heart beat. The checkride was over at that point but I was waiting to see if rudder and power application would stop the rotation. The autothrottle was disabled. As stated above, simulators are not real airplanes. Another DPE was observing the ride. He agreed with the outcome due to incorrect flight control inputs at the beginning of the event. The entire crash sequence was avoided by freezing the sim. This ugly vision would have been a recurrent nightmare for the rest of the candidate’s career. The recheck was successful.
Thanks some story Randy, thanks. I’m just curious. Once the spin was developed, did the candidate make any attempt to recover?
Rob
One thing not mentioned in this thread, generally the procedure to recover from a spin is to firstly stop the turn with the rudder. As to Barons flat spinning, around a year ago a local pilot and instructor got a Baron into a flat spin and unfortunately did not recover and perished.
You must live in Oregon Stan. And I believe if you read through all the comments, a few people did mention the standard spin recovery. One problem is that the certification of most twin-engine airplanes does not include spins. So if you get yourself into one, you’ve become a test pilot. You might get lucky and recover, then again, you might not. On the Baron spin out your way, one concern I heard was that the person in the rear seat might have pushed the envelope near the rear CG limit making the recovery problematic. That hasn’t been proven yet, but we’ll be watching for that report too. Thanks for reading.
It’s sad to hear about flat spins into the ground.
Here it is, some 50 years later and my Baron experience happened again, just last year?
Oh, small world… I, also, flew for Pan Am. Delta bought Pan Am in ‘91 and I went with Delta on the A310.
Two Army guys end up at Pan Am. Small world. :-)
Was. the spin fast enough to put the passengers to sleep? For their sake, I certainly hope so.
That’s an interesting question you pose David. I’m sure some smart Jetwhine reader could calculate the rotation speed of the spinning aircraft from looking at the video, but my guess is that centrifugal force had those people almost frozen in position until they hit. The fall would have been mercifully brief although that doesn’t mean it would have been quiet in the cabin. Just my two cents.
I intend to follow everyone who analyses this accident. it is just the sickest video I’ve ever watched
Sad is the word I used. I can’t imagine what was happening in the cockpit, much less the cabin on this airplane.
I regularly do intentional spins in a single engine aircraft (to keep my recovery skills sharp as I do aerobatics) and no matter how many times you do it, once you go beyond 5 to 6 rotations, it gets really scary, even for me who should be used to it. So, I can just imagine the absolute horror those poor people on board must have gone through. So very sad to see that airplane spinning all the way into terrain.