July 09, 2025, 18:20:00 GMT
permalink Post: 11918562
One thing that I remember from when I was a simulator TRI/TRE on a Boeing was that as an instructor you get very used to operating critical
switches rapidly without following any procedure, in order to set the sim up for a single engine landing etc. When I was then line flying next I had to guard against doing the same thing in the real aircraft.
switches rapidly without following any procedure, in order to set the sim up for a single engine landing etc. When I was then line flying next I had to guard against doing the same thing in the real aircraft.
However, it can also bite us. The Delta dual engine shutdown during takeoff from LA (referenced way back when in the 1st accident thread) was caused by muscle memory - the pilot reached down to set the EEC switches (located near the fuel On-Off switches) but muscle memory caused him to do something else - set both fuel switches to OFF. Fortunately, he quickly recognized his error, placing the switches back to RUN and the engines recovered in time to prevent a water landing (barely).
It is conceivable that a pilot - reaching down to the center console to adjust something unrelated - could have muscle memory cause him to turn the fuel off to both engines. While all new engines are tested for "Quick Windmill Relight" - i.e. the fuel switch is set to CUTOFF with the engine at high power - and the engine must recover and produce thrust withing a specified time (memory says 60 or 90 seconds) - it takes a finite amount of time for the engines to recover (spool down after a power cut at high power is incredibly fast - plus moving the switch to CUTOFF causes a FADEC reset, which means it won't do anything for ~ 1 second). Doing that at a couple hundred feet and the chance that an engine will recover and start producing thrust before ground impact is pretty much zero
July 09, 2025, 19:23:00 GMT
permalink Post: 11918592
Muscle memory is a strange and (usually) wonderous thing. It allows us as humans to perform amazing things without actually thinking about what we are doing. Professional Athletes have perfected this to a high art, but the rest of us do things using muscle memory on a regular basis. Back when I was still racing, I happened to look down at my hands on the steering wheel in fast, bumpy corner, and I was simply amazed at the large, rapid steering inputs that I was making to compensate for the bumps - with absolutely zero conscious thought. Muscle memory at its best.
However, it can also bite us. The Delta dual engine shutdown during takeoff from LA (referenced way back when in the 1st accident thread) was caused by muscle memory - the pilot reached down to set the EEC switches (located near the fuel On-Off switches) but muscle memory caused him to do something else - set both fuel switches to OFF. Fortunately, he quickly recognized his error, placing the switches back to RUN and the engines recovered in time to prevent a water landing (barely).
It is conceivable that a pilot - reaching down to the center console to adjust something unrelated - could have muscle memory cause him to turn the fuel off to both engines. While all new engines are tested for "Quick Windmill Relight" - i.e. the fuel switch is set to CUTOFF with the engine at high power - and the engine must recover and produce thrust withing a specified time (memory says 60 or 90 seconds) - it takes a finite amount of time for the engines to recover (spool down after a power cut at high power is incredibly fast - plus moving the switch to CUTOFF causes a FADEC reset, which means it won't do anything for ~ 1 second). Doing that at a couple hundred feet and the chance that an engine will recover and start producing thrust before ground impact is pretty much zero
However, it can also bite us. The Delta dual engine shutdown during takeoff from LA (referenced way back when in the 1st accident thread) was caused by muscle memory - the pilot reached down to set the EEC switches (located near the fuel On-Off switches) but muscle memory caused him to do something else - set both fuel switches to OFF. Fortunately, he quickly recognized his error, placing the switches back to RUN and the engines recovered in time to prevent a water landing (barely).
It is conceivable that a pilot - reaching down to the center console to adjust something unrelated - could have muscle memory cause him to turn the fuel off to both engines. While all new engines are tested for "Quick Windmill Relight" - i.e. the fuel switch is set to CUTOFF with the engine at high power - and the engine must recover and produce thrust withing a specified time (memory says 60 or 90 seconds) - it takes a finite amount of time for the engines to recover (spool down after a power cut at high power is incredibly fast - plus moving the switch to CUTOFF causes a FADEC reset, which means it won't do anything for ~ 1 second). Doing that at a couple hundred feet and the chance that an engine will recover and start producing thrust before ground impact is pretty much zero
It's ironic that cognitive science arguably started with 'The Cambridge Cockpit'; an attempt to make sense of, and mitigate, pilots doing this sort of thing when tired, stressed and so on. This kick started an ergonomics revolution which appears to have come full circle. Now we have cognitive science offering Bayesian accounts of neural function that might explain how innocent but unfortunate priming of 'muscle memory' when practicing for emergencies could, almost predictably, lead to this sort of complex, protection overriding, error.
As non consciously executing a complex, well practiced, but unintended, action is a fairly common experience in less critical situations, I'm surprised that there isn't already a more effective ergonomic fix than the safety switches fitted.
Last edited by Subsy; 9th July 2025 at 21:58 .
July 12, 2025, 02:55:00 GMT
permalink Post: 11920123
I assume the 10 second pause is based on how long it took the two 767 event pilots to realize their error and return the switches to RUN.
July 13, 2025, 22:36:00 GMT
permalink Post: 11921747
While this is thread is still deeply into hamster wheel status, there are some valuable inputs being made. I'll try to continue with that...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
July 13, 2025, 22:58:00 GMT
permalink Post: 11921757
While this is thread is still deeply into hamster wheel status, there are some valuable inputs being made. I'll try to continue with that...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
The EAFR data revealed that the thrust levers remained
forward (takeoff thrust) until the impact.
Would that position be consistent with a restart attempt?
July 14, 2025, 00:33:00 GMT
permalink Post: 11921803
While this is thread is still deeply into hamster wheel status, there are some valuable inputs being made. I'll try to continue with that...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
Thrust lever position post-accident - it takes very little force to move the thrust levers, a little more than one pound-force at the knob. I'd be very surprised if the post impact position was the same as pre-impact (and lever angle is recorded on the FDR, so the investigators know where they were). In short - the post-accident lever position is not meaningful.
Fuel condition switch position detents - if the Indian investigators had any reason to believe that a failed or inoperative detent was a contributing fact, the authorities would have ordered a fleet-wide inspection (especially since it literally takes only seconds to do the inspection - the paperwork would take many times longer than the actual inspection).
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a long time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes longer for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for at least 30 seconds - probably closer to 60 seconds.
For all the complaining about this preliminary report, it actually goes into more detail than is typical.
BTW, my money is still on the 'muscle memory/action slip' or whatever you want to call it. I can easily imagine a scenario along the line of 'why did you turn off the fuel' - 'I didn't - oh wait - oh
...
The last line, is not a zero possibility, but at present the whole wiring system for the fuel control switches would need to be evaluated for any potential common fault that may be intermittent before I would lean towards the cognitive slip type error. The poor old FO would have both hands on the yoke from V1, irrespective of the company's SOP and so would have been rather unpleasantly surprised by the loss of performance. I think the 10 seconds in this case is remarkably fast observation-orientation-decision-action sequence, and that one engine had already started to recover towards operation is a credit to the design of the engine.
July 14, 2025, 00:50:00 GMT
permalink Post: 11921812
During the design phase, the GE types fought back very hard at the QWR requirement - as I've noted previously, it's a very difficult requirement to meet - and doing the actual flight testing to show compliance does significant damage to the engine (as in a several percent loss in engine efficiency). But the FAA has a very specific requirement that must be met (documented in an "Issue Paper"), and they make the rules.
July 14, 2025, 08:01:00 GMT
permalink Post: 11921930
Engine restart (i.e. "Quick Windmill Relight"): Even if the igniters were firing, at high power they won't actually spark (the electrical resistance or the air at several hundred psi prevents a spark) - so they won't spark until you get down somewhere near idle if you're close to sea level. Once the engine has dropped below the min idle, it takes a
long
time for it accelerate back to even an idle condition. At takeoff power, the compressor components get very hot - do a power cut the air coming in the compressor gets heated by the residual heat in the compressor. This in turn limits how fast you can add fuel in the burner without excessively back pressuring the compressor and causing a compressor stall. So it actually takes
longer
for the engine to accel to idle that it would during a normal (cold engine) start. The 5 seconds to 95% accel requirement referenced earlier is from a stable 'high' idle (we typically call it 'approach idle' since it's automatically selected when landing flaps are selected). Approach idle is ~10% N2 higher than the in-flight minimum idle, so that takes several more seconds. Bottom line, after initiating the Quick Windmill Relight, you're not going to have usable thrust for
at least
30 seconds - probably closer to 60 seconds..
If the restriction to getting the engine relit earlier (well above idle N2) is only the spark gap, I am somewhat surprised that beefier igniters, perhaps with high/low voltage settings (for emergency/normal use), are not used. Compressed air is a reasonable insulator, but it's nowhere the oil, SF6, or vacuum that HV operators use in tight spaces.
Bigger igniters might mean you can spark the fuel at ~70% N2 at which point you're presumably seconds away from having thrust again, and don't do the significant engine damage associated with I assume EGT exceedances from scheduling high fuel to ramp N2 rapidly with already-hot parts.
My final thoughts for this event.
Misinformation, most of us including myself was of the view that Gear Up had been selected because we saw an image believed to be the accident aircraft with the bogeys stowed for retraction.The report image shows that they were not and the Gear Selector was down. A crucial element in the sequence of events.
Misinformation, most of us including myself was of the view that Gear Up had been selected because we saw an image believed to be the accident aircraft with the bogeys stowed for retraction.The report image shows that they were not and the Gear Selector was down. A crucial element in the sequence of events.
I think there is a bit of confusion running in this thread about how the auto restart function works.
Normal start uses a lot of electrical power to drive the two starters. In a situation with only the RAT supplying electrical power there won't be anywhere near enough power to turn even one starter. Restart relies on windmilling only. Igniters don't need a huge amount of power, fuel will be gravity fed to the engine driven pumps. The APU autostart function will use power from the dedicated APU battery only.
Edit to add, it can take an age for the APU to start off the battery. Well over a minute.
Normal start uses a lot of electrical power to drive the two starters. In a situation with only the RAT supplying electrical power there won't be anywhere near enough power to turn even one starter. Restart relies on windmilling only. Igniters don't need a huge amount of power, fuel will be gravity fed to the engine driven pumps. The APU autostart function will use power from the dedicated APU battery only.
Edit to add, it can take an age for the APU to start off the battery. Well over a minute.
You can spin up the engines in three ways: starter motor (electric or pneumatic, depending on type), windmill (but at low speeds, the RPM given by a windmill won't be enough), or the inertia of the already spinning engine. Quick relight I believe is predominantly inertia.