Air India 787 Crashes in Western India

Attachments

billsgs126 said:
This has not been my experience. I don't for a second doubt that if you have kept your 44-year old Commodore 64 in a climate controlled surrounding and the game cartridge hasn't been eaten by rodents or in a flood you can insert it and play whatever games are on it.

But, and maybe I'm wrong, I don't see any reason that a coffeemaker, a dishwasher, a clothes dryer, or a refrigerator needs a computer chip. Or why it would be connected to any supporting infrastructure. And, it seems to me that they were manufactured to the lowest-possible standards to ensure the lowest possible manufacturing cost.

Bottom line, they don't work. Not one item or two. But well over a dozen in my short lifetime. That isn't a statistical anomaly, it is a structural defect.
Click to expand...
I don't want to put words into the mouth of the poster you're responding to here, but I believe he specifically said software doesn't degrade with age. The same sequence of ones and zeros will do the same thing every time the hardware executes it. This is why the user interfaces, features, look and feel, etc. rarely change on systems where a software failure is catastrophic. The developer invested all of their QC on pushing out the most reliable software, tested against every conceivable failure. If they make a change, all of those QC tests have to be run again. If you want to change a database update from 3.5" floppies to flash, you have to repeat all the QC for essentially all software involved in that system.

This doesn't translate in anyway to consumer products. Samsung doesn't give a lick about their software failing on a dryer. They have to develop to a lowest common denominator of users' abilities, meaning if grandma wants to get a notification that her dryer is done, she needs to be capable of getting the dryer on her wifi and adding the app to her flip phone. Also, none of these devices have a "CPU" (computer chip). They have just augmented their microcontrollers, which they've been using since the dawn of digital controls, to support screens, buttons, a wifi adapter, etc. None of this "computer chip" tech on appliances degrades their reliability. What degrades their reliability is keeping manufacturing costs down in the wake of massive inflation in inputs over the past decades. Grandma isn't going to buy a dryer if it costs $5000. So if you use cheaper everything (including labor), you get less reliable products. This is why a dryer (whether a "smart" dryer or not) costs $500 and a 787 costs $300 million. If you bought all the parts to make your 787 off alibaba and put it together in your backyard, I'm certain it would be a fraction of $300 million, but also very likely to fail in ways you hadn't anticipated.

Personally I enjoy the tech in everything and find it to be relatively reliable. I like being able to tell Alexa to turn on my ceiling fan or make a pot of coffee tomorrow at 6am. I also have a skill set to troubleshoot the problems that occur when they do. Pilots of the transport category aircraft certainly are given the tools to troubleshoot their automations as the manufacturer attempted to understand and document all failure modes of the systems. I've never piloted professionally or have any experience in that world, but I'm guessing procedures and troubleshooting checklists are updated more often than any system or actual software as new failure modes get discovered or refined. Pilots in GA bugsmashers probably ought to know how to troubleshoot their automations, but more times than not they're here complaining about all these youngsters and their whiz bang auto-pilots while drinking a cup of coffee from their 1981 Mr. Coffee and shaking a fist at the clouds.
 
morleyz said:
Also, none of these devices have a "CPU" (computer chip). They have just augmented their microcontrollers, which they've been using since the dawn of digital controls, to support screens, buttons, a wifi adapter, etc. None of this "computer chip" tech on appliances degrades their reliability.
Click to expand...
As a point of clarification, many microcontrollers do have standard microprocessors as their main component. And many appliances, yes including fridges and washing machines, run linux as their operating system. These mass produces parts are sometimes quite overpowered for the job they are being asked to do, but because of mass production it's far cheaper to buy a system on a board, or system on a chip, than to design something vastly simpler as a bespoke solution.
 
172andyou said:
As a point of clarification, many microcontrollers do have standard microprocessors as their main component. And many appliances, yes including fridges and washing machines, run linux as their operating system. These mass produces parts are sometimes quite overpowered for the job they are being asked to do, but because of mass production it's far cheaper to buy a system on a board, or system on a chip, than to design something vastly simpler as a bespoke solution.
Click to expand...

and some of the current microcontrollers have more processing capability then old CPUs.
 
Bob Noel said:
and some of the current microcontrollers have more processing capability then old CPUs.
Click to expand...
For sure. Not even close in most cases. Again, because of mass production. If you can by a mass produced part to add wifi to your fridge for $10 you don't care if it also has enough juice to run Doom. You just care that your component cost would be an order of magnitude higher if you did something with only the capabilities you actually needed.
 
172andyou said:
Again, because of mass production.
Click to expand...

Not really? Yes, but also no. Semi-conductor manufacturing is a very strange beast. A manufacturer will take a circular Wafer of a given diameter and make as many integrated circuits (or Dies) as will fit. A smaller die size will result in more Dies-per-Wafer, and the manufacturing cost is per Wafer. Furthermore, there will be defects on the wafer. A single Defect might completely ruin whatever die it hits, so, again, having a smaller die will result in a higher percentage of successful dies form a Wafer (or yield).

Over the past several decades, the manufacturing process has decreased the physical size of the logical design, thus a chip that was designed twenty years ago to be 200mm^2 (an expected size for a CPU) might now be smaller than 100. This results in greater computing power from ever smaller Integrated Circuits and a subsequent increased in computing power in general appliances. It is also a large part of the justification for the prices of high-end GPUs or AI accelerators. Those dies are hundreds of square millimeters from the most advanced nodes. A single wafer might cost more than $10,000, and if you only get three or four good dies from that wafer... well.
 
172andyou said:
As a point of clarification, many microcontrollers do have standard microprocessors as their main component. And many appliances, yes including fridges and washing machines, run linux as their operating system. These mass produces parts are sometimes quite overpowered for the job they are being asked to do, but because of mass production it's far cheaper to buy a system on a board, or system on a chip, than to design something vastly simpler as a bespoke solution.
Click to expand...
I'm not trying to get into semantics on this, but yes I would imagine all microcontrollers have a microprocessor. My point was more that they're generally not running an Intel Atom or some type of small, full blown computer under the hood, but the IoT craze just enhanced the existing microcontrollers that were already running the fridges and washers for decades.

If my fridge ever shows a Windows boot screen though, I'm pulling the plug.
 
Surprised nobody has blamed cosmic radiation flipping the wrong bits yet; smaller architecture means more vulnerability.

(I am making a joke. Hundreds of people died and were injured. I think it's funny how far the discussion has veered away from "what could have caused this tragedy")
 
2-Bit Speed said:
Not really? Yes, but also no. Semi-conductor manufacturing is a very strange beast.
Click to expand...
I was talking about the total package, not just the chip. But, yes, chip manufacturing is complicated. The feature sizes on the kinds of chips used in consumer appliances hasn't been cutting edge for a long time though, so I would expect that yields are pretty good. It's not like a fridge requires TSMCs latest 3nm process.

morleyz said:
I'm not trying to get into semantics on this, but yes I would imagine all microcontrollers have a microprocessor. My point was more that they're generally not running an Intel Atom or some type of small, full blown computer under the hood, but the IoT craze just enhanced the existing microcontrollers that were already running the fridges and washers for decades.

If my fridge ever shows a Windows boot screen though, I'm pulling the plug.
Click to expand...
Wasn't Windows the OS on some LG fridges almost a decade ago already?
 
morleyz said:
If my fridge ever shows a Windows boot screen though, I'm pulling the plug.
Click to expand...

I have some different thoughts on what I'd do ...

(it would make
look like a soothing massage).
 
Hengelo said:
(H)ow far the discussion has veered away from "what could have caused this tragedy")
Click to expand...
I don't think many people would disagree that insufficient lift caused the tragedy. Or that insufficient power at low altitude was a root cause of that insufficient lift and the pilots' reaction in trying to keep the nose up. What everyone is waiting for are metrics and perhaps an explanation of what specific processing error (human or machine) occurred for the accident aircraft to have so little lift as to be so very low after using the entire runway.

Discussion of automated systems seems germane.
 
flyingcheesehead said:
Which part are you asserting isn't true? :dunno:

Are you saying that they didn't use the whole runway (they did), or that they weren't low after takeoff?
Click to expand...
I've read comments that assert video examination of similarly loaded 787 takeoffs indicate the accident airplane left the ground at generally the same point on the runway.
 
Last edited:
3393RP said:
I've read comments that assert video examination of similarly loaded 787 takeoffs indicate the accident airplane left the runway at generally the same point on the runway.
Click to expand...
Today, I'd imagine that almost any jet and any loading will leave the runway at nearly the same point if they're using reduced-thrust takeoffs. They need to get to V1 at a point where there's enough room left to either abort or to get into the air sans an engine (both of which should be roughly the same amount of length remaining), and then rotate and get into the air with both engines.

I did note from the link that they confirmed that the airplane was properly configured for takeoff.
 
dmspilot said:
What makes you think they used the whole runway?
Click to expand...
I don't mean that they didn't lift off until the end, I mean that they started at the end and it wasn't an intersection takeoff. There's been a lot of confusion because FlightRadar24 initially said that they had done an intersection takeoff, but when they looked at all the data they had, they came back out with a correction saying that flight did use the full length for takeoff.
 
flyingcheesehead said:
I don't mean that they didn't lift off until the end, I mean that they started at the end and it wasn't an intersection takeoff. There's been a lot of confusion because FlightRadar24 initially said that they had done an intersection takeoff, but when they looked at all the data they had, they came back out with a correction saying that flight did use the full length for takeoff.
Click to expand...
I think the person I replied to was under the impression the aircraft lifted off right as they approached the far end of the runway, which is also an internet rumor along with the intersection takeoff rumor.
 
dmspilot said:
I think the person I replied to was under the impression the aircraft lifted off right as they approached the far end of the runway, which is also an internet rumor along with the intersection takeoff rumor.
Click to expand...
I'm the person you replied to. :rofl:
 
Latest guess I’ve seen (not mine as I don’t have one)is fuel starvation caused by vapor lock. Estimating a flow of ~70000 lbs/hr total is one helluva vapor lock.
 
Last edited:
dmspilot said:
Not true. No need to spread Internet rumors.
Click to expand...
They began at one end of the runway, not an intersection. They rotated. They did not really climb.
3393RP said:
I've read comments that assert video examination of similarly loaded 787 takeoffs indicate the accident airplane left the ground at generally the same point on the runway.
Click to expand...
But, those aircraft made positive climbs (higher than 600-700 feet AGL) rather than initiating a negative climb into the terrain.
 
billsgs126 said:
They began at one end of the runway, not an intersection. They rotated. They did not really climb.

But, those aircraft made positive climbs (higher than 600-700 feet AGL) rather than initiating a negative climb into the terrain.
Click to expand...
So in your expert opinion what would a dual-engine flameout in a 787 at 51 ft AGL look like? 100 ft? 200 ft? At what point do you consider the takeoff phase complete and the climb phase begun?

Nauga,
watching more than airplanes dig holes
 
nauga said:
So in your expert opinion what would a dual-engine flameout in a 787 at 51 ft AGL look like? 100 ft? 200 ft? At what point do you consider the takeoff phase complete and the climb phase begun?

Nauga,
watching more than airplanes dig holes
Click to expand...
I'm no expert.

I don't see how they could have rotated and climbed 600-700 feet with the engines at idle. I don't see why they would not have continued to climb (as those other 787s had) if they were at takeoff power. So, what happened? That's what we all want to know.

My OP on this thread was to state that opinions and speculation as to the design and operation of automated systems onboard was germane. And I also disagreed with the idea that we did not understand the cause of the crash. (Insufficient lift, due to insufficient thrust.) What we don't know is what caused the amount of thrust to decrease.

And it certainly could have been both engines flaming out.

I will let others argue as to whether or not the pilot should have lowered the nose as airspeed bled away.
 
Larry in TN said:
Which inhibited EICAS messages would have given them information that could have led to doing something that would have helped?
Click to expand...

I didn’t say an uninhibited EICAS would have given them any info that would have been usable to them and my very next post continued my thoughts in quasi-support of Himalaya’s posts.

The inhibited bells & lights would have delayed (momentarily) their reaction to whatever was occurring. Bells/lights alert you that something is wrong, so you check the EICAS for specifically what’s occurring. Therefore, as I said, they probably didn’t even know what was occurring. Also, as Jdm said, EICAS will still show the failures, however, it shows primary and secondary failures (in order) which there were certainly multiple in this incident. I even question a little what would have even been displayed as bus switching, etc would have even delayed what was displayed. I think I’ve only done one dual engine failure in the sim (and it was at altitude), so don’t recall how quickly everything progressed. They were airborne for what, 20 seconds or so (and probably knew they were doomed in maybe half of that time)? Taking into account startle response, a quick glance across the instruments, etc, hardly time to read and recognize the various failures indicated on EICAS. I don’t disagree with this functionality and, as I also said in my post that you didn’t include, was that in this case, it most likely wouldn’t have mattered anyway.
 
billsgs126 said:
I'm no expert.

I don't see how they could have rotated and climbed 600-700 feet with the engines at idle. I don't see why they would not have continued to climb (as those other 787s had) if they were at takeoff power.
Click to expand...
I truly hope you realize that there is a wide range of conditions between idle and takeoff power and the time at which changes in power may have occurred.

Nauga,
unhampered by futility
 
X3 Skier said:
Latest guess I’ve seen (not mine as I don’t have one)is fuel starvation caused by vapor lock. Estimating a flow of ~70000 lbs/hr total is one helluva vapor lock.
Click to expand...
I've seen this kicked around a lot too. I can't wrap my mind around how any engine, let alone an engine that is consuming fuel through a garden hose, would vapor lock both engines at this point in flight. Certainly not an expert in this, but I would have expected vapor lock to occur while the plane was sitting in the sun and would hard start, fail to start or possibly have issues after start, but by the time this plane started, taxied and rolled down the runway it would certainly have pulled all the fuel in the lines through. I know it was hot, but was it really hot enough to vaporize fuel in the lines while they were under pressure and flowing that much fuel? This is a normal daily temp in Phoenix in the summer. Both engines simultaneously? I don't have the expertise to yea or nay it 100%, but it seems like such an astronomical chance for this to happen just this one time.
 
FWIW:

"One of the engines of the Air India plane that crashed last week was new, while the other was not due for servicing until December, the airline's chairman has said."


:dunno:
 
nauga said:
So in your expert opinion what would a dual-engine flameout in a 787 at 51 ft AGL look like? 100 ft? 200 ft? At what point do you consider the takeoff phase complete and the climb phase begun?

Nauga,
watching more than airplanes dig holes
Click to expand...

Well there is a defined first and second segment climb requirement.
 
X3 Skier said:
Latest guess I’ve seen (not mine as I don’t have one)is fuel starvation caused by vapor lock. Estimating a flow of ~70000 lbs/hr total is one helluva vapor lock.
Click to expand...

I'm not familiar with the fuel system in the 787, but there is probably at least 2 fuel pumps in each tank, any one of which can supply both engines. Then there are the engine pumps that should suction feed themselves at lower altitudes. Now I don't know what if any fuel pumps are supplied with electric power when operating only on the rat. So vapor lock may be a possibility.

I do know that above 30,000 ft on 737's the engine driven pumps may not draw enough fuel to keep full power. But any one of the electric pumps can supply both engines.
 
172andyou said:
I was talking about the total package, not just the chip. But, yes, chip manufacturing is complicated. The feature sizes on the kinds of chips used in consumer appliances hasn't been cutting edge for a long time though, so I would expect that yields are pretty good. It's not like a fridge requires TSMCs latest 3nm process.


Wasn't Windows the OS on some LG fridges almost a decade ago already?
Click to expand...
It's in my 2014 Ford Flex; it has a Microsoft boot screen. It's slow and annoying.
 
nauga said:
I truly hope you realize that there is a wide range of conditions between idle and takeoff power and the time at which changes in power may have occurred.

Nauga,
unhampered by futility
Click to expand...
Changes which will be shown in the data everyone is waiting for.
 
Just found out about another Boeing easter egg: TCMA. I have a few space aliens in mind.
 
You must log in or register to reply here.
Back
Top