Thoughts and Theories on the Air India Disaster
UPDATE. July 14, 2025
The preliminary crash report on Air India flight 171 is out, and what it says is startling.
It contends that at the moment of takeoff, both fuel control switches “transitioned” from the RUN to CUTOFF positions, essentially killing both engines.
One of the pilots is then heard asking the other why he did this. The second pilot responds that he did not. (Which pilot was speaking to whom is not specified.)
A few seconds later the switches transitioned back into the RUN position. The engines began spooling up again, but there was nowhere near enough time for them to produce adequate power.
In other words, it seems that one of the pilots switched off the fuel intentionally.
Or did he? Notice the investigators’ use of the word “transitioned.” The report is not fully clear as to whether the data recorder was tracking the flow of fuel independent of the switch position, or if it shows the switches themselves were, in fact, physically moved by hand. If they weren’t, it remains possible that an electronic glitch in the plane’s digital engine control system could be the culprit.
But if the switches were moved to CUTOFF manually, the billion-dollar question is why? Were they moved by accident, or nefariously? Was it an act of absurd absent-mindedness, or one of willful mass murder, a la EgyptAir, Germanwings, and (almost certainly) MH370.
As preposterous as the idea of a pilot mistakenly shutting off the engines sounds, I’ll note that it’s happened before. Consensus, however, is trending toward premise number two.
At least to me, though, the dynamics of the crash don’t really fit the suicide theory. You’re saying that the pilot’s plan was to cut both engines and let the plane glide into the ground? That seems an awfully conservative scheme. There would be a lot of unknowns in such a scenario, and no guarantee that the ensuing impact would be as disastrous as it turned out to be.
Just to the left of the impact zone was an area with no buildings and fewer obstructions. Perhaps with just a little more altitude they could’ve reached it, resulting in a crash that was partly survivable. Calculating the exact impact point ahead of time would’ve been nearly impossible.
Further, the second pilot denied shutting off the engines when queried by the first one. Why?
Rumors are circulating that the captain was going through a divorce and had been treated for depression. Whether or not this turns out to be true, just keep in mind that although depression sometimes turns people suicidal, only in the rarest cases does it also turn them murderous.
One thing that should pique interest is the report’s inclusion of a Boeing service bulletin issued a few years ago. The concern was fuel control switches failing to properly lock. The bulletin pertained to 737s, but the same basic switch is used on the 787 as well. Is it possible the fuel switches on Air India 171 were defective, and were moved from RUN to CUTOFF by nudge or vibration?
UPDATE. July 9, 2025
The focus now is on the fuel control switches. These are a pair of manually positioned cockpit controls that effectively turn the engines on or off.
These switches lever-lock into place cannot move automatically. Whichever position investigators say they were in, a pilot had to put them there.
If an engine fails during flight, one of the steps in securing that engine is to shut off its fuel control. In the case of Air India, what’s being whispered is that an engine may have failed on takeoff, and one of the pilots then grabbed the wrong switch, thus rendering both engines inoperative only a few hundred feet over the ground.
The biggest problem with this idea is that engine failures aren’t handled this way. All pilots are trained to wait until a safe altitude — usually at least a thousand feet — before commencing any steps to secure or troubleshoot a malfunctioned engine. You don’t start moving critical controls until there’s adequate time and altitude. Additionally, any time a fuel control switch is lifted out of the RUN position, both pilots must verbally confirm the correct one is being manipulated.
Is it possible a pilot could’ve have reacted in a panic and done exactly the wrong and catastrophic thing? Yes, and it’ll be quite troubling if it turns out one fuel control switch was found in RUN, with the other in CUTOFF.
And what if they were both found in CUTOFF? Well, putting aside the a possibility that the plane was crashed on purpose, this would make some sense. One of the steps addressing a dual engine failure — rare as such things are — is to cycle both fuel control switches. That is, move them from RUN to CUTOFF, then back again.
If something in the plane’s electronics went haywire and caused the engines to lose power simultaneously, we can imagine a desperate crew, with only seconds to spare, attempting the re-start process, with both fuel controls repositioned to CUTOFF and not enough time to get them on again.
Nothing in the preliminary report will be definitive, but if the switches were found in different positions from one another, this hints at pilot error. If both were found in CUTOFF, this hints at a dual failure and a last-ditch re-start effort.
UPDATE. June 24, 2025
It’s looking more and more likely that flight 171’s ram air turbine (RAT) was indeed deployed just prior to the crash. If so, this more or less confirms a loss of power in both engines.
Why it lost power is another issue altogether, and a potentially calamitous one for Boeing.
See below for more.
June 15, 2025
AIR INDIA flight 171 plunged into a neighborhood seconds after takeoff from Ahmedabad airport on Thursday afternoon, killing more than 270 people. The exact death toll is uncertain, as the search for bodies continues. One passenger survived.
My rule this soon after an air disaster is to avoid conjecture. Crash investigations can run for months before causes are nailed down, and first-glance theories, however convincing they seem in the moment, often turn out to be wrong.
That being said, evidence suggests the Boeing 787 suffered either a loss of thrust in both engines, or an inadvertent retraction of the plane’s flaps and slats before reaching sufficient speed.
The plane climbed to only about 400 feet above the ground (the 650 feet being reported by the media is the altitude above sea level), leaving the pilots no time to troubleshoot or turn back. All they could do was glide straight ahead. The flight path was stable, but the jet slammed into buildings at over 150 miles-per-hour, exploding into a fireball.
The loss of thrust theory is evidenced a few different ways. Most notably, one of the pilots, in a mayday call to air traffic control, reported power loss to air traffic control. In addition, some of the video footage appears to show deployment of the 787’s ram air turbine (RAT), a device that extends from the fuselage automatically, triggered by the loss of both engines, to provide flight control power. The “bang” heard by the surviving passenger could have been the RAT dropping into place, and the buzzing noise in one of the videos could be the sound of the device doing its thing (it’s essentially a propeller driven by the oncoming air). The footage is grainy and unclear, however.
Engine failures are rare. A loss of both engines is exceptionally rare. A bird strike, a la Captain Sullenberger, would be one possible culprit, but so far nothing points to this. Other possibilities include a malfunction of the 787’s electronic engine controls, fuel contamination, or ingestion of runway debris. There’s also the chance, however far-fetched, that a pilot shut the engines down, either out of carelessness, or, in a worst-worst-worst case situation, deliberately.
Getting back to those grainy videos, it’s hard to see much detail, but the wings look strikingly “clean,” which is to say the flaps appear retracted. Flaps, which extend from the trailing edge of the wing, together with slats, which extend from the front, provide critical lift at low speeds. Jetliners almost always take off with these devices extended (the particular setting varies with weight and runway length). The pilots then retract them incrementally as speed increases. Perhaps in this case they were brought up inadvertently — or by way of some bizarre malfunction — immediately after takeoff, resulting in a loss of lift without enough altitude for recovery.
An Air India 787.
Then we have the landing gear. It was not retracted after liftoff, as normally would be the case. Curiously, this hints at either of the two scenarios just discussed. In the first one, we imagine the pilots, distracted by engine warnings and a sudden loss of power at the worst possible time, simply neglecting to raise it. In the other, a pilot mistakenly retracts the flaps rather than landing gear. The probability of such a mistake is absurdly low, but it’s bigger than zero.
On the other hand, the 787’s aerodynamics are uniquely sculpted, and when the flaps are extended the wings take on a camber, a smoothly downward curve, front-to-back, that makes the flaps less conspicuous than they are on other airplanes. Looking at videos of 787s taking off under normal circumstances, loads of which can be viewed online, those wings, too, look very clean. And photos from the crash site show a wing with flaps and slats that appear to be extended, at least partly.
The stronger evidence points towards power loss. My hunch is that something went wrong with the jet’s digital-electronic engine management system.
Whatever it was, the data and voice recorders will tell us shortly. That would be the how. Figuring out the why might take more time. For Boeing’s sake, let’s hope it’s not a design flaw buried in the plane’s high-tech architecture.
This is the worst crash since the downing of Malaysia Airlines flight MH17 in 2014, more than a decade ago. As of June 20th the body count is 274 people, including dozens killed on the ground.
The media keep reminding us how terrible the last twelve months have been for aviation, with at least four high-profile accidents. It hasn’t been a good run, but as I underscored in a recent post, historical perspective and context are important. Even including the past year’s spate, we see far fewer plane crashes than we used to. Heck, in 1985 there was a serious accident every 13 days, on average (including the bombing of an Air India 747 that killed 329 people). Multiple deadly disasters were once the norm, year after year. This is no longer the case, despite the number of commercial flights more than doubling since the 1980s.
Related Stories:
MH370: A STORY IN PIECES
PILOTS AND MENTAL HEALTH
WHAT’S GOING ON?
Photo courtesy of Unsplash.
Leave a Comment
Maximum 1500 characters. Watch your spelling and grammar. Poorly written posts will be deleted!
74 Responses to “Thoughts and Theories on the Air India Disaster”
You are viewing newest comments first. Click to reverse order
Well well well — a reminder of just how fallible crash investigations can be. Talk about confirmation bias.
Anyone who has watched Magnar’s videos over the years knows his soft voice & studious, undramatic manner. It can make his presentations almost dry on occasion, even mildly hypnotic. But that predisposes me to take this video on the Germanwings disaster very seriously. As he points out, it has been cited repeatedly in connection with the Air India crash.
https://www.youtube.com/watch?v=y_MkjcUUTrk&ab_channel=FlywithMagnar
This theory deserves a look https://www.youtube.com/watch?v=fl1kUudq7Tw
The Indian Air Investigation Bureau says that the fuel switches were in the “cutoff” position when found and produced a photo of the blackened console. If a pilot had cut them off, how did the engines re-spool unless the switches were manually reset to run again? I think nobody touched the switches as the voice recorder states and that the theory of a catastrophic power failure is the most likely scenario with the RAT restarting the engines.
The fuel switches were found in the RUN position.
Also, the RAT cannot start the engines.
anonymous: Cuz the name “Dreamliner” is 100% BS.
Why does the photo of the Air India 787 have the registration number “N100BS” seemingly crudely edited onto it? No Air India plane is likely to be registered in the US, and that particular tail number belongs to a Piper PA-23 that’s been around since 1966.
Actually, it’s not uncommon for pre-delivery aircraft to carry temporary N-numbers, even when in the colors of foreign carriers. I suspect that’s what we see in the picture. Additionally, some foreign carriers (Avianca for example), routinely register some of their planes in the United States.
I can’t explain why the number is associated with the Piper. Maybe the Piper recently switched to that tail number from a different one. I believe the in 787 in photo is several years old.
Craig Ivinson, the investigators report one pilot asking the other why he had moved the switches to cutoff. This is a smoking gun: one pilot (presumably the first officer) SAW the other move the switches. (If he hadn’t actually seen it, his reaction would presumably have been different). Even if faulty soldering, or whatever, caused both engines to shut down simultaneously, that fault could not have moved the switches – each held by a spring in a detent — to the cutoff position.
Assuming (& we must) that the investigators are not maliciously inventing this exchange, we are left with two possibilities:
— a case of colossal absent-mindedness;
— deliberate suicide/murder.
Another possibility is that the guy who moved the switches asked the other why HE had done it, so as to put the investigators off the scent. This might be hard to imagine but should be considered.
I am here again with my thoughts about everything I have managed to read. So will start by saying I do not believe it was pilot errors
There was a maintenance bulletin issued by Boeing for a soldered connection on a component /pc card that could cause a break then make connation due to faulty soldering?? a weird one this, not a dry joint but a connection that may break and reconnected in tenths of a second due to vibration. it was deemed to be a non imperative action to inspect and repair WHY? This has been forgotten.
This very aircraft had electrical problems on its pervious flight They where reported by the crew No idea if any inspection was carried out. This has been forgotten.
On the fateful day of the crash one passenger survived who reported that he heard a loud ‘bang’ at some point on the aircraft during its taking off procedure. This seems to have been forgotten.
The tragedy that unfolded after this reported ‘bang’ was horrendous we have no idea at present fully happened on the flight deck.
I think too much emphasis is being given to the fuel cut of switches by the media and everyone else, there is a vast amount of other electrical and electronic equipment associated with these two switches. But there has been no evidence presented to say everything else was working okay and no malfunction of other components caused the RAT to deploy because both fuel cut off switches had been turned OFF and then back ON?.
Well, it remains to be seen if the fuel cutoff switches were deliberately actuated or somehow managed to be moved to the cutoff position without either pilot having touched them. I have never heard of fuel cutoff switches moving on their own, but also don’t claim to have the details on other accidents in which this sort of thing happened.
I’m not familiar with the 787’s controls, but do know my way around a 767’s cockpit. In the latter, the fuel cutoff switches cannot be “accidentally” actuated in either direction. If the 787’s fuel cutoff switches are of the same type/design—which I believe they likely are, then the fact that the EAFRs logged fuel cutoff, followed by a return to run (in a way that strongly suggests manual actuation), suggests to me this was a deliberate act. The pilot who tampered with the switches clearly was intending to put the plane into an unrecoverable situation…which he did. That the FADECs had initiated a restart once the fuel cutoffs were returned to the “run” position tends to eliminate mechanical or electronic failure as the cause of loss of thrust.
Patrick, you raise some interesting points. But your theories about possible causes are only very remotely plausible. If there were a serious problem with the Boeing 787-8, why isn’t every one grounded (at least in the U.S.)?
As William of Ockham mused in the 14th century (and I paraphrase), the simplest explanation is usually the correct one. The simplest explanation, however uncomfortable we are with it, is murder-suicide.
Going forward, I think the only appropriate measure is to install cameras inside cockpits. I understand the pilots’ union doesn’t like this approach, but it would provide answers that we may not clearly get otherwise. Plus, it could be a deterrent to murder-suicide, since a pilot contemplating such an awful action would know that his actions could be detected.
Very sad incident, but let’s hopefully learn from this and install those cameras.
I haven’t seen anyone link the evidence, landing gear not raised, fuel cut-off switch moved. You’ve come the closest. Initially it was thought that the flaps were raised by accident instead of the landing gear; Why not the cut-off by accident instead of the landing gear? Yes it would be a really stupid move, but it accounts for all the available facts, even the suggestion that the pilot was distracted by personal circumstances, which could make him more likely lean into mistaken muscle memory. Yes murder/suicide seems to be one plausable explanation, but not pulling the landing gear first is not explained in that theory. It may actually fit best that it was unintentional cut-off with intent of lifting gear.
Here’s a question: were these guys alone in the cockpit? I’ve flown Delhi-Paris — a similar stretch — & it’s almost 11 hours. So, was this flight being conducted by a two-member crew? If not, wouldn’t all pilots be present in the cockpit for takeoff?
Here’s a question: were these guys alone in the cockpit? I flew Delhi-Paris once — a similar stretch — & it was almost 11 hours. So, was this flight being conducted by a two-member crew? If not, wouldn’t all pilots be present in the cockpit for takeoff?
Gimlet Winglet — Good points all. Yes, it’s foolish to underestimate the potential magnitude of random absentminded acts.
On the verbal exchange, we’ve seen how the paraphrase has put the cat among the pigeons in the commentariat. I’ve done a lot of transcribing of speech & translation of transcriptions, & both are minefields. But then so is paraphrase.
So if I were doing the PM & worried about future discrediting, I wouldn’t have mentioned the conversation at all.
As I understand it they’ve assembled AI 787 pilots who’ve flown a lot with both the pilots involved & can help sort the voices. And others as well who will know what this noise is as opposed to that noise. Then there’s the matter of understanding & interpreting the words uttered.
I’d wait for this to be achieved with confidence before publishing anything.
Has any read the document appearing under the following link ?
https://ad.easa.europa.eu/blob/NM-18-33.pdf/SIB_NM-18-33_1
It pertains to the FAA switch mechanism recomendation.not executed by AI.
JFD
Greybeard —
If it helps, on the 787 all significant control inputs are sent over an ethernet bus. The cockpit fuel cutoff switches are not wired directly to the FADECs. Instead they are wired to remote data collectors which detect state transitions and send datagrams to the systems that have subscribed to that type of event.
As for why rocking the switch through off-on might work in an uncommanded shutdown scenario, should the switch be physically defective (eg. a weakening solder connection), chunk-chunking the switch could readily restablish electrical contact. Meanwhile, you’ve likely sent out two datagrams to interested subsystems that the cockpit switches went to CUTOFF then RUN, which might reset those subsystems out of whatever goofy state they’re in. Rocking the switch thru CUTOFF-RUN is part of the trained response to uncommanded engine shutdown.
Personally, I do not think the switches failed in any way. The Boeing 7 series planes have been in service for over half a century, they’ve had decades to learn from all the real world things that can go wrong and observe the creative ways cockpit humans can screw things up. But I would prefer the switches were moved to someplace other than the center console, that location just puts them too much in realm of target of action slip (brain fart).
rod — “But there were two switches involved here”
Moving both fuel switches to cutoff is something every pilot does every other leg, part of parking the plane. That’s hundreds of times a year. So yes, it’s in the hopper of routine muscle memory things that you can brain fart, or if you prefer, “action slip” swap in for what you were supposed to do. If I were such a pilot I would try hard to see if I could do it quickly with one hand while doing another bit of routine cockpit cleanup with the other; you get 8 hours’ rest from landing to next pushback, every second saved is another second of shuteye.
Also, there are other dual-control “routine” actions done and trained for: moving thrust levers, activating thrust reversers, switching off stabilizer electric input (those switches are guarded).
rod — “reason a full CVR transcript can be published in the final report but Not in the preliminary report?”
I can think of some. First, you do not put anything into the prelim report (PR) that you cannot defend as an under oath technical expert in court. For two reasons: your professional competence and integrity are at stake, and if even one minor detail in the PR has to get walked back after more study gets done, in the real courts and court of public opinon that calls into question everything else in that report. So don’t go out on a limb. I do not think they had sufficent time to analyze the CVR to that level of confidence. If they had, they would have timestamped their paraphrases of cockpit conversation.
It’s ok (legally) to be wrong if you can show that by applying professional best practices you reached that conclusion at that time. It’s not ok in court of public opinion to be walking back anything from the PR.
Second, the CVR raw content was not necessary for the purposes of the preliminary report, which is stating the known facts and determining if any urgent fleet action is necessary.
Third, both India and Britain law prohibit recording and replaying plane-ATC convos, and I expect (but do not know) the same law covers CVR convos. So, while the final report likely will reach conclusions and recommendations that require publication of CVR transcript and/or raw audio to support them, the PR did not, so sensibly they did not request permission to include them.
Question for you, Patrick: The report states that the Fuel Control Switches transitioned from RUN to CUTOFF position, with a gap of one second. Since the Enhanced Airborne Flight Recorder monitors the Fuel Control Valves, which physically regulate fuel flow to the engines, and not the Fuel Control Switches next to the pilots, could that have been an error in the report? Unless they corroborated the movement of the FCSs from sounds of the CVR, natch.
What many people don’t realize: The preliminary report is meant to be “factual only”. So a fact is: “the fuel cutoff switch was recorded as switching OFF”, or the fuel cutoff switch was recorded in the ON position during takeoff and then transitioned to OFF just after liftoff. That second sentence is shortened as “…transitioned to OFF”. And as Patrick notes: A glitch/fire/whatever in the wiring might cause them to record (and act) as OFF while the switches remain in the ON position. This wasn’t the case here: One pilot noticed them in the OFF position and moved them back to ON. And I agree with Patrick’s addition that we don’t know that someone deliberately moved them, it is also possible they didn’t fully lock in the ON position and were bumped to OFF.
GW — Good points.
— God knows we all have hair-raising tales of cerebral flatulence to tell about ourselves & others. That most don’t end disastrously is a matter of luck. But there were two switches involved here, & if they were functioning properly it really is hard to imagine someone executing the deliberate double motion required, & doing it twice, let alone reaching down with two hands & doing them together.
— Can you suggest any reason at all why a full CVR transcript can be published in the final report but Not in the preliminary report?
Meantime, check out Christine Negroni’s piece on this. She asks, for example, whether it should be merely assumed that RAT deployment followed, & was a result of, switch “transition”, & not vice-versa.
https://christinenegroni.com/lapses-in-air-india-crash-report-invite-criticism-and-worse/#comment-48117
Clearly much more detail about what happened, especially what was said, during the flight itself could have been included in the preliminary report, but I disagree with those who characterize this part of the report as “vague.” A better word would be “constrained.”
The word “transitioned” seems mealy-mouthed, but its use is justifiable at this point because it acknowledges the possibility that fuel cutoff might have occurred without pilot intervention. Remote as that possibility may seem, it cannot be ruled out definitively at this stage, or perhaps ever. I have yet to see any analysis of exactly how the CDR records a switch input. Does it record this event through an independent physical circuit, or does it “infer” it electronically? There is room for doubt here, or at least for argument.
Full analysis of the voice/sound recorder, with complete time stamping, will not be a strictly objective process. There will be disagreements, some fueled by uncertainty and others, alas, by politics. So it goes. I predict people will be arguing about this crash long after the final report is released, some sincerely, others “interested.”
A question from a non-pilot who is a programmer and is thus thinking about this as a debugging exercise: will (should) the recorder data show the movement of the switches in BOTH directions? That is, there are two possibilities:
1) The switches WERE turned off
2) The switches were NOT turned off: a malfunction cut the fuel flow
Since we’re told the switches were turned back on and at least one engine relit, then if it’s #2, the switches would have to have been manually turned OFF and back ON after the engine failure. Seems like that might be discernible from the data.
Meanwhile, if a malf caused the failure, why would turning the switches back on even help? Or is this command-driven operation, where flipping the switch sends a “start fuel” or “stop fuel” command, and the supposition is that a malf could have sent a bogus “stop fuel”? Point is, if it was a solder joint failure (that somehow activated both at once??), flipping the switches wouldn’t help, I’d think.
Patrick (and other pilots), I have a question about what the one pilot said to the other (granted that we do not have an exact quote nor know who was the speaker (for some stupid reason)): “why did you cut off?”
Would you phrase a question like that just purely based on warnings suddenly appearing? Or on observing that the switches are in cutoff? I don’t think that’s a natural way to question what’s happened WITHOUT having seen some motion (even if just out of the side of one’s eye). A more natural question would relate to what one observes, without asking what YOU did.
Rod — “The other denies that he did it — doubly strange”. Not really, in the brain fart scenario the pilot who moved the switches instead of some other cockpit control (eg. gear level) would not yet have realized his error. In the suicide scenario, the pilot who moved the switches would be stalling, knowing that once the engines have spooled down below idle speed (about 5 seconds) there’s no chance of getting thrust back before impact.
For likely legal reasons the preliminary report does not have a full CVR transcript nor, when paraphrasing, identify which pilot said what. We don’t know whether all what was said in that flight is included. For instance, the SOP “positive rate–gear up”, which normally would have occured right around when the fuel was cut off.
The point of the preliminary report is to state the known facts of the accident and to determine whether any urgent fleetwide hardware action is needed. We’re likely going to have to wait a year for a full accident report to get the complete CVR transcript or audio, plus proper time synching of audio events to data recorder events.
>> “one pilot asks other why did he cutoff, other pilot responds he didn’t”
This is where the plot thickens more than somewhat.
1) Note he doesn’t (reportedly) express puzzlement at a sudden power loss, as in “What the hell is going on?!”. Rather he directly asks the other WHY he moved the switches. Which suggests he saw it happen.
2) The other denies that he did it — doubly strange. If pilot A saw pilot B do it, pilot B may reply “Because God told me to crash this plane” or some other plausible response to the question “Why?” But instead he says he didn’t do it.
3) Both switches were found, in the wreckage, in the RUN position.
Mighty odd.
When you are taking off the pilot actually at the Yoke ‘doing’ the take off has two hands on the Yoke. This aircraft has a Yoke not a side stick like on an Airbus.
I have read that it was the co-pilot who was carrying out the take off. The Captain is monitoring everything else I find it very difficult to understand how the co-pilot could take his left hand off the yoke and instigate a fuel pump OFF then ON without the captain instantly taking over control of the take off on his Yoke and immediately saying two hand on the Yoke as soon as the co=pilot moved his arm.
At the very least the fuel cut off switches have a pull up twist then down to instigate a fuel cut off. The guards at each side of the switches is a bit of a simple idea to stop accident knocking of the switches and maybe a removable quick ‘hand on pull up release cover’ over the two switches clipped to and on each guard would give a more visual and extra safety margin before this operation needs to be carried out. I appreciate that the extra time 1/2 seconds to remove any switch cover could? may ? compromise the emergency operation? But that’s for the industry to decide. The other item of concern is the fact the end of the runway is not clear of obstructions ie buildings or anything else ALL runways should have a CLEAR area at each end in case of emergencies if the plane over shoots the runway end
If this runway was a clear open space at the end this tragic situation may well have been a different outcome.
I think we can say at this point that one of the pilots deliberately cut off the fuel right after takeoff and then denied it. It seems inescapable that this was a suicide/murder. Given the political and religious tensions in the area I think it is possible that the pilot who cut the fuel had religious or political motives. Either that or the pilot experience some kind of mental breakdown.
Please note, the report does not say either of the pilots moved the fuel cutoff switches, only that the switches “transitioned” to CUTOFF, then 10 and 14 seconds later transitioned to RUN.
The data recorder almost certainly tracks the electrical output from the cutoff switches, not their physical position. I am told the switches make a very distinctive sound when moved over the detent, perhaps those events will appear in the audio record in a future report.
Data sample rate for cockpit switch positions is likely once per second. So both switches moving within “1 second” could be 0 or 2 seconds. But it would be difficult to move both switches in 0-2 seconds with just one hand. Perhaps a line training pilot frequently setting up the sim might be very practiced at that. I will note that when switches were moved back to RUN, 4 (3-5) seconds elapsed between the two.
I am told that under ideal conditions (higher speed, engine didn’t spool down too much), in-flight restart takes at least 30 seconds, and 90 seconds is more realistic. The FADEC in each engine does this automatically as long as the fuel switches are in the RUN position.
To head off speculation, current fuel cutoff switches are very robust, and illicit cockbit debris falling on them is very unlikely to displace them. One could complain that there’s no reason to put them in the middle of the thrust console; some other manufacturers put them at the edges or ceiling of the cockpit and have a guard cover over them.
The preliminary accident report has been released (link at bottom). I highly recommend people go directly to the report rather than relying on mediot interpretations. It is only 15 pages, much of which one can skip (half the report is on where they found various pieces of the plane).
My main comment is, why release it at about 4AM india time? I had given up on getting the report on friday (in north america) saying, well, it’s now past midnight and is saturday in India.
Here’s the timeline, all info directly from the report and solely from the report:
08:07:33 cleared for takeoff
08:08:33 V1 speed attained (153 kts)
08:08:35 Vr speed attained (155 kts)
08:08:39 systems switch to air mode (liftoff)
08:08:42 max speed attained (180 kts)
08:08:42 “immediately thereafter” both fuel switches transition from run to cutoff with time gap of 1 second between switches transitioning. N1 and N2 start decreasing (presumably both engines)
no timestamp: one pilot asks other why did he cutoff, other pilot responds he didn’t. Their exact words are not mentioned.
no timestamp: RAT observed deployed immediately after liftoff (plane was around a wingspan’s height AGL)
08:08:47 (“about”) RAT starts supplying hydraulic power
08:08:52 (“about”) engine 1 fuel switch transitioned from cutoff to run
08:08:56 (“about”) engine 2 fuel switch transitioned from cutoff to run
08:09:05 mayday call
08:09:11 EAFR recording stops
https://aaib.gov.in/What's%20New%20Assets/Preliminary%20Report%20VT-ANB.pdf
BBC is reporting that fuel switches were moved to the cutoff position. In cockpit voice recorder is heard one of the pilots asking the other why he did the cutoff. The fuel switches were moved to the “RUN” position and the engines began to respond, but it was already too late.
The Air Current is saying their sources indicate focus by investigators on the fuel-control switches, located on the centre pedestal just behind the power levers.
I post this only for the purpose of predicting that if this indeed proves to be the case, many will cry “Boeing cover-up!”. The interesting thing about that is it rhymes with general loss of faith across the West in institutions such as media & government authorities. People know they’ve been lied to so much that they refuse to believe anything (except, of course, the things they want to believe).
From what I have read the RAT deployed when there is a major electrical failure.
I assume this is a automatic function and not pilot controlled.
The RAT takes a few seconds to come on line with electrical supply.
It does not produce the amount of electrical current required to operate the landing gear hydraulics to raise the wheels.
Due to the height of the aircraft there was no time available for the supply from the RAT to take over the control of the instrumentation in the flight deck or take over the engine functions.(re boot time )
If the aircraft had been much higher in altitude more time is available for this scenario to come on line and keep the aircraft flying and so make a emergency landing with the gear down.
I am just an amateur aviation follower but I am an electrical engineer for the past 50 years and understand switching systems and computer control systems.
To myself it looks like a catastrophic electrical failure not pilot error the big question for everyone is WHY did it happen everyone who perished in the aircraft and on the ground deserves the correct honest answer to this teribble accident.
Well, that’s ONE of the reasons for the RAT to deploy. Loss of engines or hydraulic power also will trigger it.
Parenthetically, what if instead of a statistically very unlikely dual engine failure, they had a single engine failure and brain-farted and shut down the good engine instead of the failed engine? Shutting down the wrong engine in the single engine failure scenario is a known human error pattern that training and cockpit procedures and dual pilot redundancy try very hard to address.
But I do not think this scenario happened that day, for several reasons. First, when a single engine on a jetliner fails there is no urgent need to shut it down. It is not a propeller engine that causes asymmetric drag that can result in an unstable aircraft. You kick in a bit of rudder and climb out on the good engine, get suitable height margin and try to stay on your vertical flight path, then run your engine out checklist. Second, they just didn’t have time to run that checklist. The entire flight was only airborne 13 seconds, part of it they did have engine power and climbed for a few seconds. Third, on the videos there is no sign of the few seconds of yaw which would result from single engine failure.
Three Air India 787 pilots recently tested various failure scenarios in the simulator. What they found was negative information, still useful to rule out various scenarios. In short, there was no combination of single engine out, flaps setting (including flaps 0), landing gear deployment that could put the 787-8 in a state where it could not climb. The only scenario that matched the events of AI-171 was dual engine failure, an event so unlikely that pilots do not even train for dual engine failure at or below 400 feet on takeoff, because there is no recovery. Yet it appears to have happened.
Please note, the article says “that could cause a dual engine flameout”. That’s idiotic reporter embellishment: they were testing for dual engine loss of power, not flameout which is something different and did not happen on that day (no smoke or flame from engines visible). https://www.ndtv.com/india-news/air-india-crash-simulation-pilots-recreated-ai-171s-final-moments-on-simulator-what-they-found-8812035
Have realized something: When they say the recorders are “damaged”, perhaps the correct response is “Not sh*t, Sherlock”. In any kinetically violent accident, they’ll be damaged.
What is doesn’t mean, necessarily, is that some or all of the data will be irretrievable. Maybe it’ll merely require great care in downloading.
After looking at GimletWinglet’s link, my question is: Shouldn’t the NSTB be Actively Involved in this investigation by virtue of the fact that both aircraft & engines were built in the US? Isn’t that common practice? Wouldn’t Boeing & GE be well advised to make that a contractual condition of sale?
Bits of evidence:
1: Film from the rear right of the runway showed the landing gear still fully down – BUT with the trucks tilted from ‘Heel down’ [the normal take-off orientation] to ‘Toe down’ [the position required for landing gear raising and storage. “Gear Up” command on B 787 invokes firstly the truck tilt to ‘Toe Down’; then the folding in of the whole landing gear strut. So a failure occurred at the moment the ‘Fold gear struts in’ command was actuated. What does this imply?
News today that the AAIB has successfully downloaded the data from the forward flight recorder, and is starting to interpret it. “Forward” recorder is important, because that is the one that has 10 minute battery backup, while the one in the tail runs off plane electrical power and is SOL if power fails.
So, perhaps we’ll get some data in the preliminary report around July 16th, as they are either encouraged or “required” to publish (30 days from incident). “required” because the descriptions of reporting rules I’ve read are ambiguous and no enforcement mechanism has been mentioned.
https://www.reuters.com/business/aerospace-defense/efforts-reconstruct-events-that-caused-air-india-crash-says-india-government-2025-06-26/
> its damaged condition
Yes indeed. WTF is what I’ve been saying after a fair number of recent crashes, such as Jeju Air, where it’s “Gosh, sorry guys, but the recorders
— are too badly damaged;
— lost power before the critical moments;
— had been busted for the past two years;
— etc.”
Is this really such an intractable engineering problem?
From various india news sources, they have not yet extracted info from the data recorder they have, and are currently considering whether to send it off to the US where supposedly they can better deal with its damaged condition. It took this long to (presumably) reach that conclusion?
Meanwhile, Air India has without plausible explanation suspended a bunch of its 787 long haul flights until “at least mid-July”. We’re in serious WTF is going on territory. https://news.sky.com/story/air-india-suspends-some-long-haul-flights-after-fatal-crash-and-reduces-the-frequency-of-others-13385918
One of the biggest mysteries to me is how a passenger could have survived. I understand surviving the impact but not the fireball. The video clearly shows the plane bursting into a huge fireball seconds after impact. Where was the time for the man to get out and away from that? Unless he was thrown clear. But that would likely have been fatal.
It seems the vapor lock theory is discredited, at least on the pilot rumor network. That is a good source of information for anyone interested.
They have to find the cause as soon as possible to avoid a repeat of this incident.
> The RAT deployment is confirmed by
Um, no, it is merely indicated quite probable. Confirmed would be info from one of the two (both do voice and data) data recorders saying yep, RAT deployed. It would also indicate whether it happened automatically or was pilot commanded.
> A design fault is unlikely given the track record of the plane.
You may want to familiarize yourself with the uncommanded rudder deflection issues of the 737-200, a plane that first entered commercial service in 1968, and crashed two times in the 1990s due to an emergent design defect with the servo actuator controlling the rudder. A third crash nearly happened in 1999, vigorous pilot intervention and perhaps a bit of luck with the failure mode let them land anyways, and in 1999 they finally recognized and addressed the design defect. That’s 31 years. Yeah, data recorders are much, much better these days (including on the 787-8), but pointing to an 11 year successful operating history doesn’t mean as much as you think.
Good points.
FYI, the 737 rudder thing wasn’t just the -200. I believe UA at Colorado Springs and US at Pittsburgh both were -300s.
I am not a pilot, just interested in aviation. As things stand so far:
1. The RAT deployment is confirmed by audio and visual analysis indicating a failure of both engines, a failure of electrical power or of all three hydraulic systems.
2. No evidence of pilot error since plane was properly configured for take off and flaps were extended as seen in wreckage.
3. No indication of sabotage or deliberate action by pilots to crash plane
4. A design fault is unlikely given the track record of the plane.
5. Improper maintenance seems the only remaining possibility.
Not necessarily. Failure of the electronic thrust control COULD BE tied to a design flaw that has, only now, revealed itself. It also could’ve been caused by something as mundane as a water leak or some such.
> something Electric punked the engines
This airworthiness directive, issued a few days ago, goes into effect tomorrow (weds 6/18/25). Applies to 787-8 -9 -10. Note, I am not asserting this as cause, just saying that this AD is real and the timing is… well, possibly relevant. Please note that this AD had been being worked for a few months, with airlines and boeing feedback. https://www.federalregister.gov/documents/2025/05/14/2025-08346/airworthiness-directives-the-boeing-company-airplanes
Summarizing the AD, a forward galley can leak water into the electronics bay and the waterproofing of that bay can wear out and decompose over time when subjected to frequent water exposure. Directive is to inspect and take remediative action if water damage is found.
Do I think this the cause of the air india crash? No, not at this time, not enough evidence. But I am at a loss to explain sudden uncommanded engine shutdown of healthy engines at ~150 ft AGL, so yeah, I’ll entertain speculation about what could short out in the electronics bay to cause that.
Note that this plane was in service since 2014, was about number 25 off the assembly line. If this is relevant, lots of time for water damage to do its thing.
I realize that we are not dealing with a cargo aircraft in this particular situation, but do you think the possibility exists that cargo that was being carried in the lower holds might have been incorrectly loaded/secured? Thus, just after take-off the cargo could have shifted creating forces great enough so as to drastically change the center of gravity which, ultimately, could not be overcome?
No, the accident doesn’t have any of the hallmarks of a cargo shift or CG issue.
Def not a configuration issue – Boeing has so many items that trip a horn. Flaps, gear not centered, trim not set, cause a loud horn, the Weight on Wheels switch report actual TOW on FMS, Spd brakes extended – not much else configuration wise left.
Yet the biggest culprit if not human error – is the only remaining system – something Electric punked the engines. Yet previous issues here are exceedingly scant for an aircraft in heavy rotation for 10yrs already.
I will forever prefer the good old mechanics of hydraulic and pneumatic systems.
IA and Boeing already know the answer, IA keeps the black boxes, all remains under wraps – not so much as an emergency AD.
The crew fought til the end, may they and all thier passengers RIP.
> I have heard from a couple of sources that a complete electrical failure would cause the engines to shut down on the 787. Is that true?
False. The engines are equipped with a suction feed pump that powers itself off the rotation of the shaft and there’s a valve and bypass piping near the electric pumps at the major fuel tanks, that opens automatically when feeling suction pressure. This has been tested successfully up to 20000 feet flight level, and the suction feed does provide sufficient fuel at takeoff thrust.
The full length of the runway was used, so that is not contributing factor.
One of the reasons I love your blog is because I know you can be counted on to provide thoughtful analysis free of clickbait and sensationalism. It’s why whenever there is a crash or major incident I always come here to read your take on the situation.
Thank you for your calm and reasoned approach to this tragedy. I am a college journalism teacher and appreciate the careful approach you have taken to this crash. The public needs calm voices. Thank you, again .
Since SD Palamar posted the Prospect piece, I’ll post this documentary. Since I watched this years ago, it seems to have been taken down & reposted in shorter fragments (all with the title “Broken Deams”).
It’s alarming stuff & sounds plausible to me. It’s worth pointing out, though, that AlJazeera is a Qatari outfit & that Qatar has had an issue or two with both Boeing & Airbus over the years.
https://www.youtube.com/watch?v=Tqig7dRxAlg&ab_channel=DavinaArch
https://www.youtube.com/watch?v=rvkEpstd9os&ab_channel=AlJazeeraEnglish
Thanks for your insights. I’m a retired single-engine pilot but enjoy keeping up with current air situations. Love to hear and explore situations for solutions. Thanks again.
I have heard from a couple of sources that a complete electrical failure would cause the engines to shut down on the 787. Is that true?
Initial diagrams in the press indicate the plane did not use the entire runway on takeoff. Any confirmation of this?
Patrick, thank you once again for your clear articulation of the knowns and the unknowns involving an air crash. You do a great job of spelling out some complicated things for us non-professionals. #Husker Du is a great band.
Design/Manufacturing Flaws?
One of the Dreamliners That Gave a Boeing Manager Nightmares Just Crashed:
https://prospect.org/economy/2025-06-12-dreamliner-gave-boeing-manager-nightmares-just-crashed-air-india/
Chair malfunction?
A 787 suddenly dove from the sky over Australia and they blamed someone pushing a button on the the pilots seat.
Never mind this looked similar to a dive by an Qantas Flight 72 which suddenly executed two dives on a flight to Australia. It was caused by a software error. Computer failure. A glitch. While landing, all three flight control computers shut down on an Airbus 330 in 2020 (https://www.theregister.com/2021/09/06/a330_computer_failure/).
Meanwhile, Boeing has admitted this possibility of a glitch on the 787 if the system is not periodically rebooted (https://www.i-programmer.info/news/149-security/8548-reboot-your-dreamliner-every-248-days-to-avoid-integer-overflow.html).
The cause of the accident is still unknown. Still, I wonder if we have reached the limit of our reliance on computers.
Patrick’s comment that the flaps may have been accidentally retracted instead of the landing gear makes sense and would explain a lot of things. Of course you would scratch your head how this could occur, but who knows the mindset of the pilot(s)? Perhaps they pulled three long shifts back to back for some reason (someone called in sick)?
Fatigue, excess stress, burnout ect.
Anyone know the labour laws/regulations in India?
I know I’ve done stupid stuff when I wasn’t concentrating and straight away though WTF did I do that for..
> a plane filled to the brim with fuel
A typical fuel load for the Gujarat-London run is about 50 tons. The 787 can carry 100 tons. So, unless they were ferrying fuel from India to the UK (very unlikely) they did not have a max load.
That said, 50 tons or 100 tons, you go down hitting a building at 100+ knots one should expect the fireball.
> What about chair malfunction?
There’s been a social media chain about this making the rounds, and a Greek mainstream newspaper ran an AI slop article claiming this is verified (their cited FAA docs don’t seem to exist, they got the dates wrong, and they misidentified gender and age and condition of the survivor).
The 787 seats are electric and do not slide on rails. No, this ain’t your driver’s bucket seat suddenly detaching in your crappy old Fiat (been there). Did not happen. Also, to answer Rod’s question, a RAT would not automatically deploy on manually reducing thrust levers excessively, because those engines are still producing electric power. Further, I do not think such an event would be unrecoverable even at that low AGL. Immediately firewall the thrust levers, point the nose a bit down, and off you go. Zero style points but you live to write yourself up.
Having seen the huge fireball in the video (no surprise for a plane filled to the brim with fuel), I am amazed that somebody could survive that at all.
Actually, the plane would not have been “filled to the brim” with fuel. The fuel load would’ve been significant, but, considering the distance, not close to max capacity. Even low amounts of fuel can cause catastrophic explosions when a plane crashes.
Simon — Is that compatible with a deployed RAT?
What about chair malfunction? Is it conceivable that thrust levers were pulled back to idle when a pilot’s chair inadvertently slid back?
That would be consistent with loss of power on both engines and might also explain why no gear up — pilots too busy restoring power. I’d imagine the other pilot would attempt to push levers back forward, but he’d need to fight the weight of the pilot whose chair failed. And then even when levers do go back up, might not have had enough time left before thrust actually develops at only 400′ AGL.
The video of the thing passing overhead has what sounds like a RAT spinning. (If not a RAT, what is it?) As for the sound of engines, I often wonder, listening to nearby airliners in flight, how much of what I’m hearing is merely the sound of the airframe moving through the air.
>”The longer runway roll hints at a possibility of the aircraft not having adequate thrust for take-off,” said a source close to the investigation.<
Make of that what you will-
As GimletWinglet says, while fuel contamination might explain a great deal here, wouldn't other aircraft have had problems with the same fuel.
As for waiting to find out what the recorders show, how many times recently (ex. Jeju Air) has that same expectation been dashed by the recorders failing to yield anything at all?
I’m going to take issue with the expressed “inadvertent flap retraction” theory, mostly due to reading posts elsewhere from many active 787 pilots saying (paraphrasing), “yeah, you can screw up that way and it happens, but you can’t go to flaps 0 without an extra human action from flaps 1 to 0, there’s a physical gate in the cockpit controls to prevent this. Meanwhile the plane flies ok immediately after rotation on flaps 1, just your climb rate sucks and you may have to write some awkward reports post-flight.”
On the fuel contamination, I discount that because no other flights from that airport had issues that day, and fuel contamination almost (?) never takes out both engines simulteously.
So what do I think did happen? I just don’t have an idea. Knocking off some other crapola, they did backtrack to their end of the runway, they rotated at a normal spot (4k feet from end on an 11k ft runway) rotation and lift looked normal until at about 400 feet they just started drifting towards ground. No rudder or ailereon deflection in videos (which removes single engine failure from the probabilies), no puffs of smoke and flame consistent with bird intake.
> It’s reported on the BBC that they were sort of startled by the sound of approaching jet engines prior to the subsequent impact and explosion. The engines clearly were working.
Jet engines, including the “quiet” high bypass ones they use on 787s, still make a lot of noise when close even when at low power or just windmilling.
Since it more or less is established from the grainy footage – and this will, I’m sure, be confirmed, and if at all, by the FDR readout in the weeks to come – that the RAT was deployed, a catastrophic electrical failure leading to engine failure appears quite probable. My own initial thought was that the PM retracted the flaps instead of the gear inadvertently because whilst viewing the CCTV footage, the point at which the gear should’ve been retracted is when the 788 starts losing lift.
If the PIC did indeed say, “thrust not achieved, going down”, it confirms that he was remarkably calm and lucid despite the dire situation looming. That he tried to smoothly stall the bird at the moment of impact is evident from the footage.
What does NOT add up is the statement by blokes in the canteen who survived. It’s reported on the BBC that they were sort of startled by the sound of approaching jet engines prior to the subsequent impact and explosion. The engines clearly were working. Furthermore, the one survivor has only mentioned a bang but not the silence that’d have resulted if the engines spooled down, which would’ve struck him to be strange in what typically is the noisiest phase of flight.
One pilot that was interviewed here in Australia speculates pilot/human error, rather than Boeing fault, because there has not been any grounding of aircraft. At least the black boxes will be easily recoverable.