Voepass flight 2283
Voepass Linhas Aéreas flight 2283 was a scheduled Brazilian flight from Cascavel, Paraná, to Guarulhos, São Paulo. This is a routine, two-hour flight and the aircraft (and presumably crew) had flown it regularly, including the Guarulhos to Cascavel leg earlier that morning. Voepass Linhas Aéreas is a small airline that has offered domestic flights around Brazil since 1995 (initially as Passaredo Transportes Aéreos). Until the 9th of August 2024, they had never experienced a fatal crash. Their current fleet consists of 15 ATRs, including the accident aircraft.
ATR is a Franco-Italian manufacturer (FR: Avions de transport régional; IT: Aerei da Trasporto Regionale) owned by Airbus and Leonardo and based in Toulouse. The flight that day was an ATR 72-500 acquired by Voepass from an Indonesian carrier in 2022. The 14-year-old twin-engine turboprop (serial number 908) was registered in Brazil as PS-VPB.
There were four crew that day: two on the flight deck and two in the cabin. The number of passengers was initially reported as 58 and then as 57. On the 10th of August, Voepass confirmed 58 passengers on board, citing technical issues as to why the 58th passenger was not on the manifest.
That day in Cascavel, there were another ten passengers wishing to board the flight who had been waiting at the wrong gate. When they realised and rushed to the correct gate, they were told that they were not able to board. One man told the press that he’d been adamant that he needed to be on the flight.
this passenger who missed his connection with the plane that crashed in Vinhedo. He even fought with an employee for not being able to board pic.twitter.com/o0zHBvCk4X
— Clover Lavender (@AyoolaMatthee) August 9, 2024
I even pressured them a little. ‘Mister, put me on this plane, I have to go.’ Then he said, ‘There’s no way, all I can do is rebook your ticket.’
The flight to Guarulhos departed at 14:58 UTC (11:58 local). The ATR climbed to its cruising altitude of FL170 (17,000 feet). As you can see on Flightradar24’s playback of the flight, the flight proceeded normally until it was ready for the descent to Guarulhos.
Brazil’s air traffic control (FAB Brazil) stated that they last heard from the flight at 16:20Z (13:20 local) and there was no cause for concern. There was no emergency call or even a mention of severe weather.
Based on the ADS-B data, the aircraft briefly lost altitude. This was quickly corrected and the ATR climbed to FL172 (17,200 feet). Then, the aircraft entered a steep descent. The aircraft went into a flat spin. Flightradar24 recorded a descent rate of 24,000 feet per minute (250 km/hour, 160 mph) just before it crashed into the ground.
Radar contact was lost at 16:22Z
There were only 89 seconds from that first brief descent to the crash but long enough for people on the ground to notice the aircraft spinning to the ground.
Here are two of the clearest videos taken of the aircraft coming down, showing the flat spin. I do not have any information as to who took the video. Readers receiving this by mail will need to click through to the post to watch video. Please note that this footage shows disturbing images of the aircraft crash.
The aircraft impacted the grounds of a house in a residential area, 40 nautical miles northwest of Guarulhos Airport, and burst into flames. The owner of the property said that he was at home with his wife and the maid but everyone was fine. He mentioned that the final descent of the ATR was very loud. “We thought it was a helicopter due to the noise.
Firefighters arrived on the scene to find the aircraft still in flames. They have confirmed that the bodies of the passengers and crew were still in their seats. There is no information, yet, as to where the cabin crew were in the aircraft.
Imagens mostram o local da queda pic.twitter.com/xm2dgcI3zw
— AEROIN (@aero_in) August 9, 2024
The METARs (METeorological Aerodrome Report) for Guarulhos Airport, from the hour before and after the crash, do not show any problem weather.
SBGR 091700Z 33009KT 9000 NSC 17/13 Q1019=
SBGR 091600Z 36006KT 8000 NSC 17/13 Q1019=
The above shows good visibility (9,000 to 8,000 feet) and no significant clouds (NSC), with stable pressure, no precipitation and mild winds.
However, the SIGMETs tell a different story. A SIGMET (SIGnificant METeorological Information) is used to inform aircraft in flight about severe weather on their route. They cover a larger area and are forecasts, as opposed to METARs which offer real-time observations. SIGMETs are broadcast on the Automatic Terminal Information Service (ATIS) and included in the weather information for aircraft in flight (VOLMET) to warn flight crews of significant weather which may threaten the safety of the aircraft. Brasilia and Curitiba FIRs border each other north to south and the aircraft would have flown from Brasilia FIR to the Curitiba FIR
The Curitiba region had nine SIGMETs that day, showing that the severe weather was more persistent in that area.
SBCW SIGMET 9 VALID 091530/091930 SBCW – SBCW CURITIBA FIR SEV ICE FCST WI
S2024 W05806 – S2204 W05801 – S2221 W05549 – S2359 W05525 – S2349 W05440
– S2418 W05416 – S2422 W05417 – S2439 W05316 – S2848 W04527 – S2645
W04345 – S2356 W04006 – S2140 W04452 – S2248 W04546 – S2314 W04552 –
S2338 W04639 – S2307 W04734 – S2220 W04955 – S2020 W05127 – S1809 W05326
– S1823 W05638 – S2045 W05321 – S2209 W05437 – S2024 W05806 FL120/210
STNR NC=
This says: Valid 9th September from 15:30 to 19:30: severe icing in the Curitiba FIR from 12,000 to 21,000 feet. No movement, no change.
The Brasilia region’s SIGMET was the second of that day.
SBBS SIGMET 2 VALID 091530/091930 SBBS – SBBS
BRASILIA FIR SEV ICE FCST WI S1809 W05326 – S2020 W05127 – S2220 W04955 –
S2307 W04734 – S2338 W04639 – S2314 W04552 – S2248 W04546 – S2140
W04452 – S1804 W05226 – S1809 W05326 FL120/210 STNR NC=
This gives the same warning of severe icing for the Brasilia region.
The flight would have flown from the Cascavel FIR through the Curitiba FIR to São Paulo. The Brasilia FIR covered central Brazil and was north of the flight as it passed through Curitiba.
In addition, radar reports for the São Paulo region show signs of thunderstorms and icing.
The crash will be investigated by CENIPA (CENtro de Investigação e Prevenção de Acidentes Aeronáuticos), the Brazilian Aeronautical Accidents Investigation and Prevention Centre. This is a unit in the Brazilian Air Force which investigates all aviation accidents and incidents in Brazil. The French BEA (representing the state of manufacture of the aircraft), and the Canadian TSB (representing the state of the manufacture of the engines) have confirmed they will also be taking part. ATR has reported that their specialists are ready to support the investigation and the airline.
The Flight Data Recorder (FDR) and the Cockpit Voice Recorder (CVR) were both recovered from the crash site and sent to CENIPA, who has already confirmed that they were able to extract the data from the two flight recorders.
Photo courtesy of the CENIPA Communications Office
The head of CENIPA spoke about what comes next (machine translation from the original Brazilian)
Regarding the CVR, a detailed study is being carried out of the dialogues and sounds established in the cabin and with airspace control. Furthermore, through the CVR, possible audible alarms must be identified, the research of which may require, if necessary, the use of sound spectral analysis software. Through FDR, at the beginning of the process, after extracting and obtaining the information recorded in the black boxes, the binary electronic data is converted into an engineering unit.
On Monday 12th of August, the investigation progressed to the data analysis phase, looking at flight-related activities, operational environment and human factors as well as the equipment, systems and infrastructure.
A further aspect of this case is that the ATR-72-500 was involved in a tail strike incident earlier this year. As documented on The Aviation Herald, the aircraft was mid-flight when the crew received a low hydraulics warning. They continued to the destination safely but suffered a tailstrike on landing at Salvador. The damage was described as minor at the time; however the ATR was under maintenance for three months, which doesn’t sound very minor. The aircraft was returned to service on the 9th of July 2024, one month before the accident. Brazilian media have reported that there have been multiple issues with the aircraft since then (cabin pressure lost, air conditioning system failing); however none of this has been confirmed by CENIPA.
In fact, CENIPA has reportedly released a statement to the media trying to combat other media coverage claiming inside information about the cockpit voice recorder. Again, the below is machine translated:
⚠️ UPDATE – AIRCRAFT CRASH IN VINHEDO ⚠️
Dear Journalist,
The Brazilian Air Force (FAB), through Center for Investigation and Prevention of Aeronautical Accidents (CENIPA), assures that no media outlet has had access to the audios, transcripts, or data from the flight recorders, commonly known as black boxes (Cockpit Voice Recorder and Flight Data Recorder), of the aircraft registered PS-VPB, involved in the aviation accident in Vinhedo (SP), last Friday (09/08).
CENIPA also emphasizes that it strictly follows the protocols and specific procedures established by Law No. 7,565/1986 (Brazilian Aeronautical Code – CBA), Decree No. 9,540/2018, and Annex 13 of the 1944 International Civil Aviation Convention.
Finally, the FAB reiterates its commitment to transparency and seriousness in conducting the investigations, as well as respecting the grief of the families of the victims involved in the accident.
According to Aeroin.net, Voepass have sent a related statement to their crew:
We do not yet have additional information about the accident, and as such, we ask everyone not to disclose or share any information regarding the operation with persons not involved with VoePass operations. VoePass is already taking all necessary measures to support and assistance to passengers, crew and their families. We ask everyone to stay focused on the continuity of operations, raising their alert level to the expected operational routine. If you do not feel able to perform a flight, inform the execution of a stopover for your replacement. We will be sharing new information soon.
It is clear that whatever caused the loss of control of the ATR, it happened without warning during a critical phase of flight: the transition to descent.
However, this leaves us with more questions than answers. The turboprop might suffer in severe weather but icing alone should not cause such a sudden and dramatic loss of control. What does the brief climb to FL172 mean, if anything? And finally, does the previous maintenance point towards undetected structural damage?
Looking at the videos, are the tailplanes still attached? I can’t tell, though it’s clear the fin is still there.
Ya, hard to tell, but there are a couple of frames in the first video where you can clearly make it out. (At about 7 seconds in the first video, just before it goes behind the treeline. You can advance frame by frame in the video using the comma and period keys on you keyboard.)
Well, that disposes of that idea. Is loss of cabin pressure at 17,000 feet likely to disable the flight crew?
See Rudy’s comment below but generally speaking, assuming two fit pilots without other issues, I’d expect them to have enough time to react at 17,000 feet.
Something happens… Maybe a pitot ices up, it can’t sense airspeed and autopilot turns off. (Or maybe an engine shuts off.) It starts to lose altitude, the pilot notices/panics and pulls back on the yoke to gain altitude back. Autopilot had the throttle reduced, and pilot is used to autothrottle (or something), or just does not notice. This slows the plane to near stall (and maybe it has differential thrust). Plane enters a flat spin stall. Pilot throttles up (reports of loud engine(s) as it descends), but continues to pull back on the yoke for altitude gain, and does not counteract differential thrust (or the spin) with rudder. Flat spin all the way down.
The plane looks viable. Even if it lost power, and gained ice, it should have been able to glide, if not in the spin. They couldn’t get it nose down, or did not try. Tragic. It will be interesting to see the report on this one. I often think of the Air France flight out of Rio, and how a pilot can be deceived about a situation, although in this case they seems to have visibility.
Note the repeated references to the noise level from the people on the ground. I am thinking it was power on all the way down.
ETA: Sorry, I reread and saw you dealt with that.
Sylvia, I have several years working with ATR72 and know their engine noises. In audio from cell phone footage, i heard an engine screaming at high power, contradicted by the swishing of a propeller in Beta /idle. The only way I can reconcile this is the R engine was running at full power in HOTEL mode whilst the other Left engine was in flight idle.
Low Hydraulic pressure can trigger HOTEL mode automatically.
I have been a captain on ATR 42. When I did my training at Toulouse, the first ATR 72 was just ready for flight testing to gain its certificates.
There have been many upgrades since. But auto throttle was not part of the equipment, then. I cannot tell if this had added in more recent years.
There have been reports of incidents ands even accidents involving icing and the ATR
In Ireland and the UK, where the ATR was mainly operated when I was flying it, icing conditions are encountered on a regular basis.
I do not recall having been in a situation where the de-icing system was unable to cope adequately.
I found the ATR a delight to fly, excellent handling and responsive.
Decompression at FL 170 does not necessarily put the crew in immediate danger. Of course, it will necessitate a rapid descent to a level with a more breathable atmosphere but the time available to take action will be substantially longer than, let’s say, FL 350.
Even at FL 370 or FL 390, altitudes at which I encountered rapid decompression in a Citation, we had ample time to assess the situation and start an emergency descent. If I may venture a guess: two supposedly healthy pilots would have probably ten minutes or more before lack of oxygen would affect their ability and judgment. So unless the pilots were very cavalier about a loss of pressurization and did not take action within about 2-5 minutes (a long time in a situation where quick action is required), hypoxia should be ruled out.
Airlines may allow relatively junior pilots to be promoted on the smaller aircraft in the fleet. Sometimes a younger pilot will be promoted to captain on commuter routes and -aircraft types. Not seldom the next step will be to become first officer on the larger types. Yes, they would meet required minimum experience and other criteria, but they may be accompanied by an FO who is just starting his or her career.
Did one of them panic? Extremely hard to say and probably grossly unfair to two people who more than likely were fighting for their lives, and those of the passengers and cabin crew.
I once got into severe icing conditions in a Cessna 310. It was not forecast, and the pilot of another C310 who had just flown through the same area (in opposite direction) told me that he had not encountered any problems. Worse, the heating in our 310 had failed so the temperature inside was way below freezing.
This reduced my reaction time, but when I realised that the aircraft was stalling, I shoved the control column forward, regained control, declared a “Mayday” and asked ATC to guide me clear of mountain tops below.
The 310 had full de-icing equipment and was actually certified for operations in known icing conditions. Anyway, once out of the worst of the icing conditions, I was able to cancel the emergency and continue the flight. But the ATR 72 is a different aircraft. Assuming that icing was a factor, would a rather ruthless action of, quite literally, shoving the controls hard forward, have saved them? I cannot honestly say.
The death of a crew and passengers in an accident, any kind actually, is a sad occurrence and reminds us of the fact that, no matter what technological advances, there are always risks too when we travel.
Maybe we will hear what caused this sad crash, but all that I can add is: May they rest in Peace!
Thank you for the context; I didn’t know you’d flown the ATR as well.
The captain, at least, was very experienced. I didn’t look into the FO. I think the CVR will tell us a lot; even with no discussion, we should be able to hear what immediately preceded the stall. It really is tragic.
Sylvia: the captain was a very young 30-ish looking Danilo Romanos with 8 months as First Officer and 10 months as Captain. His F/O was much older looking Humberto de Campos Alencar e Silva. A transcript of the CVR leaked to Globo TV has the F/O expressing confusion about the engine and demanding more power. No mention of icing in the CVR.
Regular commenter ‘Captain Crunch’ speculates on AVHerald:
The theory is that this caused tailplane icing.
If this is confirmed by the investigation (a preliminary report typically takes 30 days), then I imagine this scenario:
The pilots configure the aircraft for descent, which included reducing power. For the first time since the repairs, the aircraft experiences tailplane icing, because the de-icing equipment malfunctioned. The aircraft pitches up, briefly gaining altitude while losing speed. Then it stalls and spins. The pilots apply power, but failing to understand the cause, follow the wrong recovery procedure.
The flight recorders have been recovered, which means the preliminary report may well confirm or disprove my take on the event.
P.S.: at 17,000 ft (~5000m) in the mountains, air pressure (and thus oxygen content) is halved compared to sea level. Yet most people do just fine without supplemental oxygen, and if there is altitude sickness, the onset of symptoms typically takes hours. It would surprise me if hypoxia was the primary cause, but if there was a low oxygen condition in the cockpit (for which there is no evidence yet), then it might have contributed to the accident by making the pilots think and react more slowly than they would have otherwise.
Now that is very interesting!
Capt Crunch let’s cut to the chase, On the ATR72, with Hydraulic Auxiliary pump in standby, if there is a low-pressure warning, the system automatically activates HOTEL mode. That is when the propeller brake is applied.so the engine can be run declutched. In HOTEL mode on the ground the R engine can be used as an APU, or to pressurize hydraulics with pneumatics. As I am not an ATR pilot, I am uncertain if the propeller feathers in HOTEL mode?
I did however work several years on the Ramp with Air New Zealand subsidiary Mount Cook Airlines who used ATR72 -500 & now -600 series, so I am familiar what Hotel mode sounds like. Listening to cell phone audio of Voepass2283 spinning to earth, I recognized the sound of HOTEL mode with a propeller swishing in Beta/idle.
Based upon what I heard, I believe the Right engine was revving quite hard in HOTEL mode, with the left engine running with propeller in idle.
If that is the case pilots were running their Right engine very hard in a vain effort to pressurize the hydraulics. that seems odd to me, but it is not fair to double guess these pilots fighting to save their plane.
Sylvia, this aircraft PS-VPB suffered a severe tail-strike on March 11, requiring rebuild of hydraulics in the tail empennage. Following repairs the same aircraft was noted to be suffering persistent hydraulic faults as many as 4-5 times per day.
Within 8 hours of the crash the Brazilian air safety regulator sealed copies of the failed safety audit of Voepass (formerly known as Passerado) which was declared bankrupt in 2012.
The issue with hydraulic failure is ATR-72 aircraft use HOTEL mode on the ground, running the R engine with a propeller brake engaged.
When the Hydraulic Auxiliary pump is in standby and there is a low hydraulic pressure warning, then HOTEL mode activates automatically, in earlier -200 & -300 series this was impossible in flight however in the QRH for -500 & -600 series aircraft it is recommended for ditching to allow hyd’ pressure for lowering flap, so whilst not intended for use in flight, it is possible. No prizes for guessing how a braked propeller induces a stall/spin. Do not expect an honest accident investigation from Brazil. The safety regulator gave prior dispensation for Voepass to disconnect 8 input channels to the FDR, including, now get this, the hydraulic pressure in both hyd systems, plus inputs for rudder pedal positions.
Clearly Air safety regulators in Brazil knew in advance that PS-VPB was not airworthy.
Expect the investigators to blame icing to distract attention from the hydraulics issue.
Voepass exited bankruptcy proceedings in 2017.
The French BEA will be a party to this accident investigation, since the ATR-72 is built in Toulouse, and any cover-ups of the sort you describe will be evident to them. The BEA has published their own opinion in the past when they found an investigation lacking.
I have not been able to find sources for your claims. We already know that the “leaked transcript” may be fake—at this point, I wouldn’t be surprised if ChatGPT could produce something believable.
P.S. I am not “Captain Crunch”, I just quoted them. Take the comments of a breakfast cereal with a grain of salt!
Brazil’s TV Globo claim to have a leaked copy of the transcript. In which they claim the first officer was shouting at the Captain for more power. My money is on engines, not, ice.
You said so two weeks ago. The evidence hasn’t gotten any better since then.
I’d take that bet. The Flight Data Recorder will tell.
CENIPA has published a very detailed preliminary report, with all data pointing squarely at ice.
Thanks for this!