Announcement

Collapse
No announcement yet.

Breaking news: Ethiopian Airlines flight has crashed on way to Nairobi

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • 3WE
    replied
    Originally posted by BoeingBobby View Post
    I'll take the one with the Engineer!
    But it's so loud and noisy and smoky...and I can't see where you can watch the game on TV.

    Leave a comment:


  • TeeVee
    replied
    yet another problem with the MAX: https://www.bbc.com/news/business-48503610

    Leave a comment:


  • BoeingBobby
    replied
    Originally posted by Gabriel View Post
    Ok so maybe it is time to define "complex" and how to measure complexity.

    Yes, planes will ALWAYS be of balancing drag with thrust, balancing weight with lift, and being stable and controllable. And in a purposely designed airplane you can achieve that with just 2 analogical controls (one for thrust and one for rudder) and zero instruments. Such a plane would take-off, climb, cruise, descend and land at the same speed (because throttle controls altitude) and will have a hard time taking off and landing in anything but the slightest crosswind. Not very efficient or even practical.

    So complexity comes in how you provide the shortest take-of and landings with the highest cruise speed, the highest cruise speed with the lowest fuel consumption, the highest payload with the minimum empty weight, the greatest dispatch reliability, and why not, the safest safety (all that while keeping the development and production cost as low as possible and being able to fit the plane in a standard gate).

    Aviation has become much more efficient and safer, and technology and complexity are at the core of these improvements. Will the airplanes have more wings, more engines, more fuel tanks, more control surfaces? No, but thy will have more sensors, redundant actuators, computers, and lines of code. And COUNT is just a small part of the complexity. The other part (that counts more than count) is connections.

    As an example, take a route system where there is one central hub that serves 5 major destinations and each of them serves other 5 minor destinations. You have 31 cities and 60 routes (counting both legs).

    Let's double the cities that are served from each point. Central hub with 10 major destinations each of them serving 10 minor destinations. Now you have 111 cities and 220 routes. You doubled the cities from each point and the result was more than TRIPLE the total number of cities and routes.

    As you see, adding count adds complexity, but that's nothing compared to adding connections. Let's take the original 30 cities and let's start flying point-to-point. Now you have 870 routes. That's almost 15 times!!!

    This is the same that happens in a plane when you start to interconnect components and systems.

    Let's take an example: Compare this:

    [ATTACH=CONFIG]25147[/ATTACH]

    with this:

    [ATTACH=CONFIG]25148[/ATTACH]

    Which one do you think offers a simpler user interface, is more reliable (less probability of failures), is more robust (works better when failures do occur), is safer?
    And which one of them has more components, more energy and information pathways (interconnections), more lines of code, more data storage, has more potential points of failure, has more potential modes of failure, is harder to predict the full implication of failures (especially combined failures), and it is harder to tell "what is it doing now" when it is not reacted as expected (either because failures in the system or in the expectations/understanding)?

    ATL says that the future will be isolating failing systems and working with what's left. And he is probably right in the short term, for the evolution of the current systems in the next generations of the current airplanes.

    But I think that in the longer term (but not as long as having only 1 pilot in the cockpit, let alone zero, both things that almost sure will become true eventually), the system will do that by itself and keep enough robustness that the pilot will not even realize, the plane will not lose any function, characteristic, or displayed parameter since all the parameters will be measured from different sources AND calculated based on other parameters with enough redundancy that it will almost never happen that you lose so much redundancy that the autoflight, control law, flight characteristics or information presented will be affected.

    I'll take the one with the Engineer!

    Leave a comment:


  • 3WE
    replied
    Originally posted by Gabriel View Post
    Ok so maybe it is time to define "complex" and how to measure complexity...[a big lecture]
    Again, the vast majority of your talk is about the past and the present, NOT THE FUTURE.

    You again focus on computing being the simplifier.

    I just don't see you engaging to any great degree in the question if the plane itself is going to be lots more complex in the future.

    You show a picture of two cockpits, but don't acknowledge that one of those planes is DESIGNED simpler in having two less power levers and two less N1, N2, EPR, EGT, Oil pressure, Fuel Flow, EDIT- THE ALL IMPORTANT TURBINE INLET TEMPERATURE, Intake heat? Crossfeed and who knows what other gauges...Yeah, the computer interface helps, but that's not the whole story.

    I did catch "more sensors" (and "connections") and a term in agriculture we call "interactions" which is more important than connections...I don't argue that they may sneak in some sensors here and there (and take out the AOA sensor, except for stall warnings), and it will take computer power to keep it all straight, but maybe I forced it out of you that the actual design of the aircraft may not get all that more complex...

    But, alas, that's a bit of extrapolating on my part.

    Leave a comment:


  • 3WE
    replied
    Originally posted by Gabriel View Post
    Nonsense.
    Nonsense: Your opinion.

    Cameras are not in the cockpit: Fact.

    Leave a comment:


  • elaw
    replied
    There is one particular group of pilots I can understand might not like the cameras... the type this NTSB report discusses: https://www.ntsb.gov/investigations/...ts/AAR1503.pdf (Gulfstream overrun on takeoff at KBED in 2014).

    From the second paragraph of page vii: "A review of data from the airplane’s quick access recorder revealed that the pilots had neglected to perform complete flight control checks before 98% of their previous 175 takeoffs in the airplane, indicating that this oversight was habitual and not an anomaly."

    Leave a comment:


  • Gabriel
    replied
    Originally posted by elaw View Post
    On the topic of cameras, I would think the main concern by pilots would be the recording of conversations criticizing their employers or coworkers, and the resultant backlash if the employer found out. But a CVR is 10x better at that sort of thing, and planes already have those!
    And the airlines don't listen the CVoiceRs. And won't watch the CVideoR.

    Leave a comment:


  • elaw
    replied
    On the topic of cameras, I would think the main concern by pilots would be the recording of conversations criticizing their employers or coworkers, and the resultant backlash if the employer found out. But a CVR is 10x better at that sort of thing, and planes already have those!

    Leave a comment:


  • Gabriel
    replied
    Originally posted by 3WE View Post
    If I were a pilot, I would strongly oppose cameras. (Maybe ATL or Bobby can list some key bullet points why this is a bad idea).

    If I were an investigator, I would find it useful in some instances.

    I consider it one of those tough arguments: good points on both sides...maybe things will change, but so far pilots have won...maybe things won't change.

    Of note is that TOTALLY VOLUNTARY (and driver-paid-for) use of cameras is growing (rapidly?) in the car and truck-driving world and cameras are apparently somewhat standard in the train-driving world.
    Nonsense. Many workers (including many "drivers" of public means of transportation) are getting video recorded all the time they are working, and the don't have the strict protection standards that a cockpit video recorded would be subject to (on par to CVRs). I don't see how a video camera passively recording the cockpit would be a distraction for the pilots, and I don't see how the pilots are entitled any superior privacy while at work, especially when their work is critical for the public safety.

    Leave a comment:


  • Gabriel
    replied
    Originally posted by 3WE View Post
    My comment was "maybe not"...maybe THE PLANES THEMSELVES won't be ALL THAT MUCH MORE COMPLICATED THAN TODAY. (and blah blah blah)
    Ok so maybe it is time to define "complex" and how to measure complexity.

    Yes, planes will ALWAYS be of balancing drag with thrust, balancing weight with lift, and being stable and controllable. And in a purposely designed airplane you can achieve that with just 2 analogical controls (one for thrust and one for rudder) and zero instruments. Such a plane would take-off, climb, cruise, descend and land at the same speed (because throttle controls altitude) and will have a hard time taking off and landing in anything but the slightest crosswind. Not very efficient or even practical.

    So complexity comes in how you provide the shortest take-of and landings with the highest cruise speed, the highest cruise speed with the lowest fuel consumption, the highest payload with the minimum empty weight, the greatest dispatch reliability, and why not, the safest safety (all that while keeping the development and production cost as low as possible and being able to fit the plane in a standard gate).

    Aviation has become much more efficient and safer, and technology and complexity are at the core of these improvements. Will the airplanes have more wings, more engines, more fuel tanks, more control surfaces? No, but thy will have more sensors, redundant actuators, computers, and lines of code. And COUNT is just a small part of the complexity. The other part (that counts more than count) is connections.

    As an example, take a route system where there is one central hub that serves 5 major destinations and each of them serves other 5 minor destinations. You have 31 cities and 60 routes (counting both legs).

    Let's double the cities that are served from each point. Central hub with 10 major destinations each of them serving 10 minor destinations. Now you have 111 cities and 220 routes. You doubled the cities from each point and the result was more than TRIPLE the total number of cities and routes.

    As you see, adding count adds complexity, but that's nothing compared to adding connections. Let's take the original 30 cities and let's start flying point-to-point. Now you have 870 routes. That's almost 15 times!!!

    This is the same that happens in a plane when you start to interconnect components and systems.

    Let's take an example: Compare this:

    Click image for larger version

Name:	CSeries_-Flight-Deck.jpg
Views:	1
Size:	247.7 KB
ID:	1039793

    with this:

    Click image for larger version

Name:	707-cockpit.jpg
Views:	1
Size:	57.8 KB
ID:	1039794

    Which one do you think offers a simpler user interface, is more reliable (less probability of failures), is more robust (works better when failures do occur), is safer?
    And which one of them has more components, more energy and information pathways (interconnections), more lines of code, more data storage, has more potential points of failure, has more potential modes of failure, is harder to predict the full implication of failures (especially combined failures), and it is harder to tell "what is it doing now" when it is not reacted as expected (either because failures in the system or in the expectations/understanding)?

    ATL says that the future will be isolating failing systems and working with what's left. And he is probably right in the short term, for the evolution of the current systems in the next generations of the current airplanes.

    But I think that in the longer term (but not as long as having only 1 pilot in the cockpit, let alone zero, both things that almost sure will become true eventually), the system will do that by itself and keep enough robustness that the pilot will not even realize, the plane will not lose any function, characteristic, or displayed parameter since all the parameters will be measured from different sources AND calculated based on other parameters with enough redundancy that it will almost never happen that you lose so much redundancy that the autoflight, control law, flight characteristics or information presented will be affected.

    Leave a comment:


  • 3WE
    replied
    Originally posted by brianw999 View Post
    [Deletion]
    This however did concern me....

    For almost 20 years, many investigators have also called for cockpit video cameras to record what information is actually displayed to pilots. The BEA repeated the proposal after AF447. Pilot unions oppose the idea due to concerns over privacy and fears it could be a distraction.”

    That tells me that pilots don’t want to risk being found out that they fucked up.
    If I were a pilot, I would strongly oppose cameras. (Maybe ATL or Bobby can list some key bullet points why this is a bad idea).

    If I were an investigator, I would find it useful in some instances.

    I consider it one of those tough arguments: good points on both sides...maybe things will change, but so far pilots have won...maybe things won't change.

    Of note is that TOTALLY VOLUNTARY (and driver-paid-for) use of cameras is growing (rapidly?) in the car and truck-driving world and cameras are apparently somewhat standard in the train-driving world.

    Leave a comment:


  • 3WE
    replied
    Originally posted by Gabriel View Post
    Well, you already mentioned FADEC. Let's see what else...Blah Blah Blah.
    I'm not 100% sure what you are saying (sorry, too much stuff)...but maybe you are strongly focused on how computer systems will simplify things and be simplified (at the interface level) themselves.

    I agree- BUT my point is that we have already made huge strides from the 707 to the 787. My context was you and Evan talking about THE FUTURE and the comment that pilots are going to have to know even MORE stuff as the planes get more and more complicated.

    My comment was "maybe not"...maybe THE PLANES THEMSELVES won't be ALL THAT MUCH MORE COMPLICATED THAN TODAY.

    You get overly drawn into nitpicky details, but to a pilot, I think a 787 is simpler than a 707. And a glass cockpit with a computer to help sort through all of the zillion steam gauges...(computers, yes, we agree). BUT HOW ABOUT IT ONLY HAVING 2 ENGINES? That IS simpler than a 707. Maybe fewer independent fuel tanks with less complicated fuel management?

    Could the future have simpler SYSTEMS. (Will Navigation get simpler- seems to me we could eliminate about 10-bazillion cute NAMED fixes for something more logical).

    SYSTEMS simplicity as much as COMPUTERS COVERING FOR COMPLEXITY...(Maybe you comprende- but the vast majority of your talk goes to past details vs. what the future looks like (except to say computers will have a single, simpler interface).)

    I hope that Boings 7-11-7 isn't a low riding 737 fuselage with 32 ft diameter overslung engines that has to have a PullUpCAS for power additions at go-around, and then a separate don't stallCAS for when PCAS looses some load/torque sensor and the TAS calculator thingy gets confused with the AOA indication. I kind of hope that maybe that 32 ft diameter engine goes behind the wing and it's all kind of balanced so the plane remains stable- or maybe we simply tweak the programming of the F-35A Raptor and copy it over to the 7-11-7 so that it simply stays stable-period-end-of-sentence.


    ***END POST***

    ***NEW POST ON ANOTHER ANGLE***

    Along those lines- you mentioned the fancy new airspeed indication that is calibrated. (Yes, I remember IAS, CAS and TAS and the 172 POHFCOM has a calibration table in it!)

    Seems to me that could be part of the chain of failures in the 'MinLav crashes...Why in the hell do we need to adjust airspeed by a couple of knots because the pitot tube is reading it slightly inaccurate? By adding ythe fancy correction algorithm the pilots were told of airspeed disagrees...If the engineers had left that $hit alone, the Ethiopian guys would have seen that the airspeed INDICATION was OK-albeit slightly inaccurate. Instead the guys saw, "we ain't sure about airspeed". Maybe DCAS is NOT to blame for this, but instead engineers and software guys having to be cute and saying, hey, we can do away with the CAS table in the FCOMPOH with this ADDED COMPLEXITY (which adds more failure points).

    As I think about this, I see unnecessary complexity! SCREW THE AOA, give me UNCALIBRATED INDICATED AIRSPEED...That number is perfectly useful to maintain safe speeds and ATC requested speeds and max climb speeds and on and on. Hui Theiu Lo didn't need to know about airspeed correction systems...he needed to know to watch the airspeed and keep your hands on the throttles like we did in 172's and Tomahawks.

    Leave a comment:


  • brianw999
    replied
    Originally posted by BoeingBobby View Post
    I read the last paragraph and agree 100%.

    This however did concern me....

    For almost 20 years, many investigators have also called for cockpit video cameras to record what information is actually displayed to pilots. The BEA repeated the proposal after AF447. Pilot unions oppose the idea due to concerns over privacy and fears it could be a distraction.”

    That tells me that pilots don’t want to risk being found out that they fucked up.


    Leave a comment:


  • Gabriel
    replied
    What flight systems (that's important, hell yeah, the window tint and environment and IFE systems ARE more complex) are on the 787 that are that much more complex than a 707? (Yeah, they botched the battery trying to be cheap and lightweight...let's trash Boeing some more).
    Well, you already mentioned FADEC. Let's see what else...

    In the 707 (and I believe that also in the 737 pre-MAX) the airspeed indicator is a differential-pressure manometer with one end tied to the pitot and the other end to the static port (just with a fancy scale that instead of units of pressure has units of speed but which shows NOT the actual speed unless the conditions are ISA standard sea level), the altimeter is an absolute-pressure manometer tied to the static port (and calibrated in a fancy altitude scale which shows something similar to the real altitude if you set the atmospheric pressure at sea level right blow you in the Kollsman window). Pretty SIMPLE. In the 787 (or the 737 MAX) the airspeed is computed from the difference between the pressures in the static and pitot ports but corrected for OAT (i.e. for Mach number) and angle-of-attack (because the pitot is not always 100% aligned with the airflow). In this way, instead of displaying what is officially called the IAS they display the CAS, except that the CAS is not the old CAS (calibrated airspeed) but the COMPUTED airspeed, which is in fact the old EAS (equivalent airspeed) which is the IAS corrected for instrument error (which depends on AoA and config) which is called CAS (calibrated airspeed) and then corrected for compressibility. And the altitude is still the absolute static pressure as before, but now corrected for total pressure and AoA (depending on the AoA, the static ports can receive a little bit of total pressure). That's nice, right? Except that now you have a lot more of failures modes for each parameter. In that way, when AF lost the speeds the altitude also changed showing a slight but sudden reduction in altitude which some speculate is what triggered the pilot to pull up like crazy (and, by the way, the incorrect too low airspeed also inhibited the stall warnings at time, creating paradoxical behaviors like when the AoA was super high the stall warning was silent and at the one time when the FO took control and lowered the AoA a bit the stall warning sounded because the airspeed increases and the inhibition was removed). And in the MCAS accidents the failure of 1 AoA sensor created disagreeing airspeeds and altitudes. Nothing of that would have happened in the 707 (or pre-MAX 737). But... add a bit more of complexity (like comparing AoA sensors or crosscheck parameters against the combination of other parameters) and that can be fixed.

    Let's see what else... Oh, yes, in the 707 the controls pull cables that go all the way through the fuselage and wings and these cables pull hydraulic servo valves which in turn actuate the controls.
    In the 787, the controls are just a nice-looking joystick (with force-feedback, unlike the passive spring-loaded ones for Airbus) that are connected to some electronics that sense their position, send that position to several computers, the computers use that position as an input, together with a myriad of other parameters including pitch, bank angle, power setting, airspeed, mach number, altitude, temperature, angle of attack, aircraft weight, configuration and load factor, run a super algorithm, conference with each other, and decide what control inputs to command to achieve what they interpret the pilot wants to achieve with those joystick inputs while trying to keep the plane within the envelope at the same time, and send that information electronically to the relevant control units of each control surface which takes that information and actuates the hydraulic valves accordingly.

    I will not mention the changes in flight management, navigation, communication, auto-flight, terrain avoidance, windshear avoidance, instrument panel, and etc. because, among other things, I suspect you would not consider that being "flight" systems.

    My prediction (which can very well be wrong), unlike ATL's, is that there will be more system integration with a unified interface, in a way that differences between redundant systems and lack of consistency between different parameters (like that if we are flying horizontally, the fuselage is horizontal, and the vertical speed is zero, then the AoA is NOT 40 degrees, this is the kind of sanity check that Boeing seems to be adding to the MCAS) will be dealt with internally by the system(s) but the pilots will not need to do anything or even know abut the issue except by an EICAS message of the type "We just killed the AoA#2, you lost 1 live (i.e. level of redundancy), you have 4 lives left". The plane will keep being basically the same plane for the pilots (in the way it flies, in the way the information is presented in the panel, etc) until you reach to perhaps 2 lives left. At that point it becomes a Cessna 172. At zero lives left you are along for the ride (down), but the chances to get to 2 lives (let alone zero) will be extremely negligible.

    Leave a comment:


  • 3WE
    replied
    Originally posted by Gabriel View Post
    Simpler, user-friendly, user interface
    That too was not really my point...(but at least it's related- and yes I saw it).

    But I see you making more posts and verbiage-some incorrect- regarding the off-topic autothrottles). You will note that I didn't jump your case for the interface comment since it was a little bit on target.

    My alternative hypothesis is NOT that in the future, computers will make up for ever increasing aircraft complexity (Comprende? & See footnote 1)

    My hypothesis is why do future airplanes have to be more and more complex?

    I'm saying that perhaps, complexity topped out with the B-52 or 707 or 747.

    What flight systems (that's important, hell yeah, the window tint and environment and IFE systems ARE more complex) are on the 787 that are that much more complex than a 707? (Yeah, they botched the battery trying to be cheap and lightweight...let's trash Boeing some more).

    Yes, there will be incremental improvements in computer stuff and in flight systems...but why will it be MUCH MORE complex? Won't some improvements use a computer and go in the simpler direction?

    Wings push air down, tail planes (and ailerons and spoilers) adjust the direction and computers and human inputs (by varied mechanisms) generate control inputs and computers are giving big doses of stability...Do we need all that much more? (Seriously, I really would like to "deactivate" DCAS unless it's really needed for good stall prevention). (See footnote 2)

    We seem to have locked in on a relatively universal flight data display, and two super duper reliable engines and a lot simpler fuel management than in the days of old...Maybe we'll have the 737 and even more work arounds for 50 more years, but not sure that new airplanes have to be tons more complicated (and require tons more computer simplification). Instead maybe they're slightly more complicated here, slightly less there, and yeah sure, a more powerful computer- that may have interface improvements...but still doing many of the same things we have now (except for AOA indication).

    The scenario above is NOT a "much more complex airplane made more simpler by computer interfaces".



    Footnote 1: Today, computers are already doing lots of things to make pilot's life simple (FADEC)...My hypothesis is that things are kind of going to "stay the same"- and not be on a huge push where you need LOTS MORE computer power. My PC simulates just about every plane there is to a half-ass degree of accuracy- so the systems that drive 787's and CRJ's and ERJS and ARJs and Airbii...are converging to a fair degree.

    Footnote 2: I seem to recall that there was a fantasy that we'd suck a tiny layer of turbulent air in off the wing surface through numerous tiny slots to achieve true laminar flow...are we really going to do that?...Ever??? Seems like we are happier just putting longer narrower wings- and get the nice efficiency boost...but what's more complex about a longer wing, that affects a flight system that a Pilot will have to learn about?

    Footnote 3: This isn't a black and white Evan statement, I imagine there will be some complexity creep, but I also see us sort of topping out and who says aeroengineers won't make systems simpler on from-scratch airplanes? Again, we are not talking about what started as a DC-9-30 with five stretches, new wings and huge ass engines that require a sneaky mini elevator on the engine struts. (Not all that different from a 737-200 that stretches and gets taller and gets DCAS). Maybe the fresh planes will be kept simpler (repeated for the 30th time?)...maybe.

    Footnote 4: Evanie did get my post and comment that "workarounds" go along with bringing older aircraft up to newer requirements (with the implication that new designs may not need the work arounds- along with a jab that indicates he has disdain for fundamentals.

    Footnote 5: Autothrottles are already a widely used complexity- but arguably a well managed complexity. I do not forsee huge growth or huge change in autothrottles...since they are already on most airliners today (some CRJ's excepted).

    Leave a comment:

Working...
X