|
https://avherald.com/h?article=4d1de8cc&opt=1
|
#
?
Jan 17, 2020 17:13
|
|
|
|
| # ? Apr 20, 2024 13:40 |
|
|
I would blow Dane Cook posted:Anyone ever slept in the crew rest compartment on an airliner? They’re not terrible on the 747s, especially the ones that were built as freighters versus converted, none of our 767s have them though, which sucks.
|
|
#
?
Jan 17, 2020 17:25
|
|
|
quote:The brave pilots had to open the side window to clean z windshield to land the plane.( un official report) Wow.
|
|
#
?
Jan 17, 2020 17:34
|
|
|
e.pilot posted:They’re not terrible on the 747s, especially the ones that were built as freighters versus converted, none of our 767s have them though, which sucks. Lots of times when I would sleep back there the slight nose high deck angle would trick my inner ear into thinking I was falling over backwards just as I started to fall asleep. I actually slept better in the seats in the back.
|
|
#
?
Jan 17, 2020 17:56
|
|
|
I'm probably missing something obvious, but why would they need to clean the windshield? Couldn't they do an instrument landing?
|
|
#
?
Jan 17, 2020 18:11
|
|
|
Craptacular posted:I'm probably missing something obvious, but why would they need to clean the windshield? Couldn't they do an instrument landing? Can't see the runway at minimums if the windshield is covered
|
|
#
?
Jan 17, 2020 18:13
|
|
|
hobbesmaster posted:Can't see the runway at minimums if the windshield is covered Not seeing the runway hasn't stopped Jerry. 
|
|
#
?
Jan 17, 2020 18:18
|
|
|
I thought it was possible to do instrument landings now with no visibility at all (?).
|
|
#
?
Jan 17, 2020 18:19
|
|
|
Craptacular posted:I thought it was possible to do instrument landings now with no visibility at all (?). I don't think ils cat iiic exists in the real world?
|
|
#
?
Jan 17, 2020 18:21
|
|
|
F-14s can do it on a carrier
|
|
#
?
Jan 17, 2020 18:24
|
|
|
Sagebrush posted:F-14s can do it on a carrier Carriers actually have a bespoke radar system for that. Had, anyway, that spectrum (CBRS) is now being freed up for private LTE networks.
|
|
#
?
Jan 17, 2020 18:29
|
|
|
Dumb question, and while sorta MCAS related but... AoA sensors, one would better than two right? Three would be good, otherwise you get to the split-brain problem. As in, which sensor do you trust or do you take a vote scenario.
|
|
#
?
Jan 17, 2020 19:00
|
|
C-Bangin' it 
|
Colostomy Bag posted:Dumb question, and while sorta MCAS related but... yeah think about it like this, if we save a few bucks by skipping on an aoa sensor the ceo's bonus will be a bit larger
|
|
#
?
Jan 17, 2020 19:06
|
|
|
I think a lot of single redundant systems will compare results to what the computer thinks should be reasonable, like if you maybe have your two AoA sensors disagreeing and your inertial navigation says you going in a straight line, your VSI says your altitude isn't changing, your ASI is reading reasonable speed, etc, it might decide the one sensor saying you're stalling is faulty and stop paying attention to it. I pretty much work on Pipers and King Airs though, so take a few grains of salt
|
|
#
?
Jan 17, 2020 19:17
|
|
|
Colostomy Bag posted:AoA sensors, one would better than two right? Three would be good, otherwise you get to the split-brain problem. As in, which sensor do you trust or do you take a vote scenario. I'm pretty sure there was a crash caused by two AoA sensors going bad simultaneously, so 3 is good for redundancy but don't solve with a vote.
|
|
#
?
Jan 17, 2020 19:22
|
|
|
EvenWorseOpinions posted:I think a lot of single redundant systems will compare results to what the computer thinks should be reasonable, like if you maybe have your two AoA sensors disagreeing and your inertial navigation says you going in a straight line, your VSI says your altitude isn't changing, your ASI is reading reasonable speed, etc, it might decide the one sensor saying you're stalling is faulty and stop paying attention to it. Also remember to turn on the (optional) red light that says the sensors disagree.
|
|
#
?
Jan 17, 2020 19:23
|
|
|
Cat Mattress posted:I'm pretty sure there was a crash caused by two AoA sensors going bad simultaneously, so 3 is good for redundancy but don't solve with a vote. There are 2 AoA sensors, one for the captain's instruments, one for the F/O's instruments. The problem is that if the captain's failed while the F/O's was ok MCAS would go nuts and a "AoA disagree" warning would show up ( but only if you had the instrument panel AoA indicators turned on )
|
|
#
?
Jan 17, 2020 19:27
|
|
|
MCAS sure is something. I've been listening through back episodes of an aviation podcast (Airline Pilot Guy) and am up to the ones in late 2018. The Lion air plane has crashed and they've been talking about it a bit, some blame directed at Boeing but also a lot directed at the pilots for not disconnecting the electric stabilizer trim and recovering the airplane. Very curious to see if that changes as time goes on and Ethiopian 302 goes down followed by the MAX being grounded.
|
|
#
?
Jan 17, 2020 19:46
|
|
|
hobbesmaster posted:There are 2 AoA sensors, one for the captain's instruments, one for the F/O's instruments. The problem is that if the captain's failed while the F/O's was ok MCAS would go nuts and a "AoA disagree" warning would show up ( but only if you had the instrument panel AoA indicators turned on ) Ah poo poo, had a dumbass attack with this one. Was thinking of one instead of two pilots and getting the proper reading. Automotive pedal type values and readings type thing.
|
|
#
?
Jan 17, 2020 20:03
|
|
|
The real problem is that MCAS is a garbage system designed by idiots who were cutting corners to solve a problem that shouldn't have needed solving but for the fact airlines are whiny and cheap, which itself comes back to the fact that the vast majority of consumers have no idea what air travel actually costs to do properly and would kill their own mother to save $10 on a fare. On a related note, one of our ULCCs had another incident in Mexico where a flight got cancelled and they handled it in an ultra-low-cost manner and now everyone is pissed. Where do these people think the ultra-low-cost comes from? PT6A fucked around with this message at 20:14 on Jan 17, 2020 |
|
#
?
Jan 17, 2020 20:11
|
|
|
Cat Mattress posted:I'm pretty sure there was a crash caused by two AoA sensors going bad simultaneously, so 3 is good for redundancy but don't solve with a vote. Yeah, if you have three sensors and one is known to be failed, then when your two remaining ones disagree you're stuck. Obviously you need four sensors so that you can put one out of service and still vote with the remaining three. This is the GPC (general purpose computer) status readout in the space shuttle:  The shuttle had four identical computer systems running the same flight control software, constantly comparing against each other and voting. If there was a failure or consistent disagreement in one of the systems the commander would be alerted through this panel and he could manually take that system out of the loop, while leaving three to continue voting. But what if all four computers failed simultaneously? They're all identical and running the same software, so it's possible that an unknown hardware or software error could take them all out at once. In that case the commander can switch in #5, the backup flight computer, which is immune to the errors in the first four computers because it uses completely different hardware and software built by a different contractor, and which only contains the logic required to get the shuttle in and out of orbit. Anything less is basically criminal
|
|
#
?
Jan 17, 2020 20:13
|
|
|
Maybe manufacturers should just build planes that someone can fly with their hands and feet and poo poo without involving a "should I dive at the ground??? Y/N" computer also what PT6A said
|
|
#
?
Jan 17, 2020 20:16
|
|
|
Mao Zedong Thot posted:Maybe manufacturers should just build planes that someone can fly with their hands and feet and poo poo without involving a "should I dive at the ground??? Y/N" computer Let's not go too far. MCAS was definitely a problem, but computerization is essential in being able to run a huge and complex system like a modern transport aircraft with a flight crew of two, and have a decent chance of not dying if one pilot is incapacitated. It is, in essence, a flight control in its own right at this point, and a faulty system like MCAS is no different from a poorly engineered mechanical system that has an obvious and severe failure mode.
|
|
#
?
Jan 17, 2020 20:28
|
|
|
sanchez posted:MCAS sure is something. I've been listening through back episodes of an aviation podcast (Airline Pilot Guy) and am up to the ones in late 2018. The Lion air plane has crashed and they've been talking about it a bit, some blame directed at Boeing but also a lot directed at the pilots for not disconnecting the electric stabilizer trim and recovering the airplane. Very curious to see if that changes as time goes on and Ethiopian 302 goes down followed by the MAX being grounded. Ultimately it's the swiss cheese model:  This is a standard model for understanding process risk. Engineering controls attempt to prevent the risks by engineering them out of existence. Administrative controls prohibit actions or decisions that increase the risk. Behavioral controls teach the user how to operate the system safely and correctly. When an error gets through all of those, it becomes an incident, and then you have physical barriers (failsafes, shields, etc) that try to stop the damage. If those too fail, you have a catastrophe. In this case no single factor was entirely responsible for crashes, and changing any one would be enough to prevent them. - The plane was designed with less than optimal handling characteristics in the first place. (engineering) - There was not enough physical redundancy in the AoA system. (engineering) - There was no standard warning system to indicate a failure in the AoA system to the pilots. (administrative/enginerering) - The MCAS system was not designed with much thought put into its ergonomics or the consequences of its failure. (engineering/behavioral) - The MCAS system and its effects on the airplane's flight characteristics was poorly documented. (administrative) - The Boeing lobbyists managed to get simulator conversion training waived when it really should have been required. (administrative/behavioral) - The training that was given, in the case of Lion Air at least, was probably insufficient in the first place. (behavioral) Change any one of those factors and things probably would have been different. The people blaming the pilots for not disconnecting the trim are focusing on the "mitigate" step in the image above, where the error has already happened and it's now up to the pilot to correct it in the moment. It's true that that could have saved the planes; there were indeed dozens of cases of probably MCAS-related "runaway trim" with the 737 MAX before the two crashes, where the pilots disconnected it in time and flew the rest of the flight uneventfully. But focusing on mitigation ignores all the other factors that had to line up to let it get that far. Sagebrush fucked around with this message at 20:42 on Jan 17, 2020 |
|
#
?
Jan 17, 2020 20:35
|
|
|
PT6A posted:Let's not go too far. MCAS was definitely a problem, but computerization is essential in being able to run a huge and complex system like a modern transport aircraft with a flight crew of two, and have a decent chance of not dying if one pilot is incapacitated. It is, in essence, a flight control in its own right at this point, and a faulty system like MCAS is no different from a poorly engineered mechanical system that has an obvious and severe failure mode. Yeah, for sure, I was being flip. Not implying we need to ditch computerization, but if your design involves literally a "point it at the ground" computer you have hosed up.
|
|
#
?
Jan 17, 2020 20:56
|
|
|
I want to be able to fly a 737 with an Xbox controller.
|
|
#
?
Jan 17, 2020 21:01
|
|
|
Charles posted:I want to be able to fly a 737 with an Xbox controller. join the navy, become a submariner.
|
|
#
?
Jan 17, 2020 21:08
|
|
|
Mao Zedong Thot posted:Yeah, for sure, I was being flip. Not implying we need to ditch computerization, but if your design involves literally a "point it at the ground" computer you have hosed up. You're at the same point in your flight training as me; you should know that an airplane's fundamental layout is designed to make "point it at the ground" (gently) the stable default state  Anti-stall computers are a great invention. I would agree that the specific implementation of MCAS is bad, of course.
|
|
#
?
Jan 17, 2020 21:08
|
|
|
PT6A posted:On a related note, one of our ULCCs had another incident in Mexico where a flight got cancelled and they handled it in an ultra-low-cost manner and now everyone is pissed. Where do these people think the ultra-low-cost comes from?
|
|
#
?
Jan 17, 2020 21:19
|
|
|
SeaborneClink posted:join the navy, become a submariner. "Stop pissing, Yuri. Give me an Elite Series 2 and a map and I'll fly the Alps in a plane with no windows."
|
|
#
?
Jan 17, 2020 21:40
|
|
|
slidebite posted:Just read that too. I understand people wanting to save money but it I wouldn't fly one of those ULCC unless I had to, especially on a flight out of country or more than a couple hours. ULCCs open up the possibility of flying anywhere to a huge number of Americans.
|
|
#
?
Jan 17, 2020 22:00
|
|
|
Sagebrush posted:Yeah, if you have three sensors and one is known to be failed, then when your two remaining ones disagree you're stuck. Obviously you need four sensors so that you can put one out of service and still vote with the remaining three. If you have 2 that disagree you have to turn off the system that uses it for decision making. For example on an airbus if these instruments were disagreeing it would kick over to an alternate law mode and automation would be off. Then you’d only crash if one of the pilots command full nose up or full nose down, which surely would never... oh
|
|
#
?
Jan 17, 2020 22:09
|
|
|
Meanwhile, over yon'quote:COMAC engineers miscalculated the forces that would be placed on the plane’s twin engines in flight - known in the industry as loads - and sent inaccurate data to the engine manufacturer, CFM International, four people familiar with the matter told Reuters
|
|
#
?
Jan 17, 2020 22:42
|
|
|
Sagebrush posted:Ultimately it's the swiss cheese model: 10/10 post 
|
|
#
?
Jan 17, 2020 22:46
|
|
|
Switzerland posted:Meanwhile, over yon' Just throw some coins in em
|
|
#
?
Jan 17, 2020 23:14
|
|
|
The only reason this doesn’t happen in north America is that locusts were inadvertently eradicated there.
|
|
#
?
Jan 18, 2020 01:12
|
|
|
Came back from a break a little while ago and one sector was completely down the shitter with ufo reports. "A string of like 16 lights" getting reported by 15 aircraft. I walk back out to find out where the starlink satellites are. It's them. Thanks Elon.
|
|
#
?
Jan 18, 2020 03:09
|
|
|
fknlo posted:Came back from a break a little while ago and one sector was completely down the shitter with ufo reports. "A string of like 16 lights" getting reported by 15 aircraft. I walk back out to find out where the starlink satellites are. It's them. Thanks Elon. Those things are pretty hard to miss once its pointed out.
|
|
#
?
Jan 18, 2020 03:11
|
|
|
KYOON GRIFFEY JR posted:ULCCs open up the possibility of flying anywhere to a huge number of Americans. I agree in principle, but it's always a tradeoff. It probably wouldn't be a problem if you flew to some place in Canada or the US which is reasonably safe and you can easily manage a disruption. Less so when it's Cancun, and there's a bunch of shady poo poo and you don't speak the language. I mean, fly them if you want, but be ready and able to eat the consequences when something goes wrong. It's a great option for visiting family or something like that, less so for visiting a foreign country you obviously don't feel at all comfortable in. There's only so much you can possibly cut in flight, the rest of the savings come from being perfectly willing to leave you exactly as far up poo poo creek as legally allowed when things go awry. PT6A fucked around with this message at 03:44 on Jan 18, 2020 |
|
#
?
Jan 18, 2020 03:42
|
|
|
|
| # ? Apr 20, 2024 13:40 |
|
|
Sagebrush posted:Ultimately it's the swiss cheese model: This is a great post, I'd like to address MCAS based on engineering principles alone as a safety system. Redundancy should be baked into any life safety system anywhere, in my field how we address this is having either 2/3, or more commonly 3/4 redundancy. Basically, let's say you have a minimum of 3 sensors, if 1/3 sensors trip a "bad" value (we're assuming one sensor is hosed from the start, in this case stuck at 0⁰ AoA, so 2/3 is all that's required to trip), a warning is displayed, giving the operator a chance to cut the bad value out of the loop and engaging 1/2 redundancy, this means if one of the remaining sensors registers a "bad" value (in this case an AoA approaching stall) the automatic safety feature actuates, again it sounds a warning concurrent with the actuation, giving the operator the ability to cut out the remaining bad value, along with the lone operational sensor, returning the safety feature (trim) to a default value (whatever it normally is during level flight) disengaging any automatic override and returning control to the operator, this shoul be controlling trim with the electric actuators as the normal method, with manual control as a backup. Inherent in this system is a minimum of two independent computing systems, i.e. each sensor has one set of mechanical linkages and two electronic paths, this allows the operator to select away from a faulted electronic signal path. Automatic safety features are critical to any modern system that is vital to maintaining human life, old-school full manual controls are too prone to operator error, and too slow, it's been born out time and time gain that automatic safety's beat out human responses 9 times out of 10, but this requires a properly designed system, and a safety system that's not covering for an inherent design flaw, as Sagebrush said. Anyone in a field of engineering critical to human life can tell you how the "swiss cheese model" (what Sagebrush posted) or "pyramid model" (wide base of redundancy leads to stability) are essential to preventing failure. Operator training and control is the first layer of blocks in the pyramid, and the last piece of swiss cheese in the diagram, which is why it's unacceptable for Boeing to not provide training on their new poorly engineered system. SeaborneClink posted:join the navy, become a submariner. In case anyone wondering he's not kidding, XBOX controllers are not used to drive submarines, but are used for a variety of critical functions onboard, because they're more reliable and cheaper than bespoke controllers and talk güd with windows computers. The joysticks used to drive a modern submarine are like $50,000.
|
|
#
?
Jan 18, 2020 07:59
|
|
