Air India Crash

[Penfold101]

But would you notice that they were flipped in less than a second? How long would it take for you to even look? Oh and you're taking off, a very busy time to boot.

I suspect (with no evidence or experience) that the EICAS probably threw up a warning message (or many) saying words to that effect rather than him looking down to see the position of the switches themselves. I don’t think that would be the immediate thought that would run through your head if your engines suddenly shutdown - many other things (bird strike etc) but not unless you physically saw them being moved. Sully is a great example of the time it takes for humans to take in a given situation and process it before choosing a response - we don’t just react immediately unless we know what’s coming.

 

[Feek]

suspect (with no evidence or experience) that the EICAS probably threw up a warning message (or many) saying words to that effect

Yup, there will be many.

Excuse the fact it's the ol' 744 sim again but the 787 FCOM doesn't suggest anything different.

/edit - That direct link doesn't seem to work very well. Let's try this:

It's the sound, or lack of it as the engines spool down that scares me.

 

[MOOGLEYS]

I have replaced a few of the same type of switches as these on military aircraft. They do fail the spring can break which is what holds the switch in place usually from being kicked as the aircrew climb in and out. But both broke on the same panel and not noticed before flight. Seems a bit far fetched but not impossible I suppose.

 

[esmozz]

But would you notice that they were flipped in less than a second? How long would it take for you to even look? Oh and you're taking off, a very busy time to boot.

You're right, and if only talking about the switches in isolation you may well not notice them in a busy flight deck after takeoff but the point I was making was that the first one was flipped back on then 4 seconds later the second one. It's just seems an unusually long time between.

This looks more like either a catastrophic mistake or, since I'm not familiar with the switches which were subject to an maintenance advisory, maybe there is something about the switches which has caused this.

 

[esmozz]

I have replaced a few of the same type of switches as these on military aircraft. They do fail the spring can break which is what holds the switch in place usually from being kicked as the aircrew climb in and out. But both broke on the same panel and not noticed before flight. Seems a bit far fetched but not impossible I suppose.

I once took over an aircraft in which the rudder trim return to center spring had broken. A gentle push in either direction and the rotary switch remained in that position, pushing the rudder all the way out to the stops. I'm really glad I found that one out on the ground.

 

[Jokester]

It could just be an artifact of the scan time of the controller or what ever it's connected to.

 

[ubersonic]

Having both switches simultaneously switching to the cut off position because of vibration or anything like that is incredibly unlikely.

In addition, it's also worth noting that even with the locking gate on the switches non-functional, any vibration significant enough to move both switches would also be significant enough to move pretty much every other toggle switch in the cabin too. Granted that may be difficult to check post crash but I would wager such an event would have caused many other recordable/observable effects.

 

[200sols]

You're right, and if only talking about the switches in isolation you may well not notice them in a busy flight deck after takeoff but the point I was making was that the first one was flipped back on then 4 seconds later the second one. It's just seems an unusually long time between.

This looks more like either a catastrophic mistake or, since I'm not familiar with the switches which were subject to an maintenance advisory, maybe there is something about the switches which has caused this.

The switches were subject to that in 2018 but the throttle console was swapped out twice after those advisories the last being 2023 so those swicthes shouldn't be an issue.

 

[mattyfez]

If you watch some of the analysis by actual pilots it's highly likely to be human error or done on purpose... they go into great detail and talk about the switches and the advisory on the switches and it's... tenuous at best...

 

[esmozz]

The switches were subject to that in 2018 but the throttle console was swapped out twice after those advisories the last being 2023 so those swicthes shouldn't be an issue.

That being the case, then I'm familiar with those exact switches as they are used on other Boeing aircraft and it's very difficult to see how they could move inadvertently.

 

[esmozz]

If you watch some of the analysis by actual pilots it's highly likely to be human error or done on purpose... they go into great detail and talk about the switches and the advisory on the switches and it's... tenuous at best...

The GD accident investigation team is more than up to the job.

 

[Penfold101]

The switches were subject to that in 2018 but the throttle console was swapped out twice after those advisories the last being 2023 so those swicthes shouldn't be an issue.

You’re assuming the at the spares/replacements they used were subject to the SAIB. They could have been sat on the shelf for years before being used, and the inspection was not mandatory.

Given Air India chose not to do it on their serviceable aircraft, there is no reason to believe the spares would have been checked either. I’ve dealt with inspections like these in the military world, and they will instruct whether any items held by the supply system must be checked too.

I agree that a very unlikely chain of events must occur for something else to have operated them apart from the crew, but the board felt the SAIB (and the airline’s actions) were significant enough to warrant mention in the preliminary report so it shouldn’t be completely discounted.

 

[ubersonic]

The fuel cutoff switches are identical to ones on newer versions of the 737-800 and 737 Max aircraft and I've operated them many times. They are gated and in order to operate them you need to pull the switch up, move over the gate before releasing it. You need to do this in order to either move the switch up (on) or down (off).

Question if you please about the switch function.

Aside from the gate are they like normal toggle switches where they stay in position unless physically moved? (with the gate being just a safeguard against catching them inadvertently). Or are they just floppy switches with no resistance that would drop from run to cut by themselves without the gate holding them up?

I would presume the former (as the latter is an extremely stupid design) so if that's the case are they notable weaker in resistance than the other toggle switches in the cabin?

My reason for asking is that a lot of noise has been made around the idea that the switches could be vibrated/jolted into the cut position if the gate failed, but the way I see it any jolt/vibration capable of doing that to both of them with the gate broken would also have done it to half the switches in the cabin and probably caused the pilots to react on the CVR.

 

[mattyfez]

Question if you please about the switch function.

Aside from the gate are they like normal toggle switches where they stay in position unless physically moved? (with the gate being just a safeguard against catching them inadvertently). Or are they just floppy switches with no resistance that would drop from run to cut by themselves without the gate holding them up?

I would presume the former (as the latter is an extremely stupid design) so if that's the case are they notable weaker in resistance than the other toggle switches in the cabin?

My reason for asking is that a lot of noise has been made around the idea that the switches could be vibrated/jolted into the cut position if the gate failed, but the way I see it any jolt/vibration capable of doing that to both of them with the gate broken would also have done it to half the switches in the cabin and probably caused the pilots to react on the CVR.

From 5.00 mins in this video...

 

[esmozz]

It's occurred to me that the 4 second delay between is probably explained by the electronic checklist system on the more advanced 787. On the 737 we still use paper checklists for everything, including memory items.

With the 787 I believe all checklists are electronic and presented on an information display with the requirement to confirm each checklist items complete with a button press. This could explain the delay.

Although it is probably not relevant, in the event of a dual engine failure in a B737, the memory items call for the movement of the engine start levers (fuel cutoff levers) to cutoff after the engine has failed and switch them back to on after the EGT (exhaust gas temperature) has been seen to decrease.

I.e. In the event of a dual engine failure you first move the levers to cutoff then back to on again. As happened in this accident, although I think if the timeline in the report is correct then this doesn't explain what happened here.

Question if you please about the switch function.

Aside from the gate are they like normal toggle switches where they stay in position unless physically moved? (with the gate being just a safeguard against catching them inadvertently). Or are they just floppy switches with no resistance that would drop from run to cut by themselves without the gate holding them up?

I would presume the former (as the latter is an extremely stupid design) so if that's the case are they notable weaker in resistance than the other toggle switches in the cabin?

My reason for asking is that a lot of noise has been made around the idea that the switches could be vibrated/jolted into the cut position if the gate failed, but the way I see it any jolt/vibration capable of doing that to both of them with the gate broken would also have done it to half the switches in the cabin and probably caused the pilots to react on the CVR.

I honestly don't know as I've only ever seem them work perfectly. They have a spring which pulls the lever towards the body of the switch, you pull up against this spring and move the lever up and over a gate along the action of the switch. The spring then pulls the lever back towards the body of the switch on the other side of the gate and holds it in place. The gate is a metal protrusion along the action of the switch - it's not something that can fail, its either there or it isn't. If the spring fails - I don't know, possibly?

 

[squerble]

It's occurred to me that the 4 second delay between is probably explained by the electronic checklist system on the more advanced 787. On the 737 we still use paper checklists for everything, including memory items.

With the 787 I believe all checklists are electronic and presented on an information display with the requirement to confirm each checklist items complete with a button press. This could explain the delay.

Although it is probably not relevant, in the event of a dual engine failure in a B737, the memory items call for the movement of the engine start levers (fuel cutoff levers) to cutoff after the engine has failed and switch them back to on after the EGT (exhaust gas temperature) has been seen to decrease.

I.e. In the event of a dual engine failure you first move the levers to cutoff then back to on again. As happened in this accident, although I think if the timeline in the report is correct then this doesn't explain what happened here.



I honestly don't know as I've only ever seem them work perfectly. They have a spring which pulls the lever towards the body of the switch, you pull up against this spring and move the lever up and over a gate along the action of the switch. The spring then pulls the lever back towards the body of the switch on the other side of the gate and holds it in place. The gate is a metal protrusion along the action of the switch - it's not something that can fail, its either there or it isn't. If the spring fails - I don't know, possibly?

The metal gate is still there even if the spring fails. It obviously makes it a lot easier that the switch could move back enough to get past the gate but as I said in a previous post, it seems far, far more improbable that both springs had failed and that a smooth takeoff moved them both to cutoff within a second of each other yet managed to stay in the run position during a crash, as opposed to a pilot reaching his hand(s) over and pulling them.

Occam's razor here.

 

[Werewolf]

Question if you please about the switch function.

Aside from the gate are they like normal toggle switches where they stay in position unless physically moved? (with the gate being just a safeguard against catching them inadvertently). Or are they just floppy switches with no resistance that would drop from run to cut by themselves without the gate holding them up?

I would presume the former (as the latter is an extremely stupid design) so if that's the case are they notable weaker in resistance than the other toggle switches in the cabin?

My reason for asking is that a lot of noise has been made around the idea that the switches could be vibrated/jolted into the cut position if the gate failed, but the way I see it any jolt/vibration capable of doing that to both of them with the gate broken would also have done it to half the switches in the cabin and probably caused the pilots to react on the CVR.

If the latches are broken they'll probably move like a normal toggle switch, so not a massive amount of force needed (probably about the same as a bunch of others).

IIRC there have been several incidents that have resulted in switches being knocked/caught by clothing etc and not immediately noticed, hence why they do have the safeguards such as side rails/guards, switch covers and latching mechanisms that require more than just a simple push/pull movement.

Basically if there is an additional safety mechanism assume that at some point someone has caused an accident/nearly caused one because they've either reached for the wrong thing, or because at some point someone/something has accidentally changed the position of a similar switch. or a switch has failed internally in another manner and the additional locking mechanism is in part to prevent an incident due to the original problem.

One of the problems is that switches/controls that have to be readily accessible for routine or emergency use tend to have to be where they are easy to get to, which means they're also more likely to get knocked. IF the latching mechanism was in any way faulty it could mean that the switches got knocked or that another normally minor issue caused a problem. A lot of aircraft incidents are down to multiple small failures, none of which on their own would be a big issue because there are meant to be other checks/safeties that would catch it.

 

[Nuclear Fusion]

If it was a suicide situation it just seems a very unfortunate way to do it, taking so many lives with you. I know as well it's not the first time such a thing has happened either.

 

[mattyfez]

If the latches are broken they'll probably move like a normal toggle switch, so not a massive amount of force needed (probably about the same as a bunch of others).

If the latches were broken they wouldn't stay in the on position, they would drop straight into the cut-off position due to how they work.

They wouldn't have been able to take off, but let's say they did fail, when the switches were returned to the on position, the engines automatically ignited and began spooling up normally, suggesting the switches were fine.

They were in the 'on' position when it crashed - the crash would have caused a hell of a lot of vibration, and yet they remained on.

The problem is it takes a good few seconds for the engines to spool up from near idle speed, to a speed that would generate enough thrust to stay in the air or continue climbing. If they had a few more seconds to allow the engnes to spin up a bit more, they could have potentialy recovered.

 

[ubersonic]

If the latches are broken they'll probably move like a normal toggle switch, so not a massive amount of force needed (probably about the same as a bunch of others).

Thanks, so there's basically zero chance of it happening due to vibration/jolt even if the gates on both switches failed then.