-
#250 Reply
Posted by
EEVblog
on 29 Mar, 2014 02:18
-
If any of this is related to the ex-Malaysian Airlines 777 sitting in a hangar in Israel since Nov 2013
Your inability to provide links to back this up is duly noted.
-
-
If he eats that chip, does the Prestige happen when he's on the toilet?
Magic can only happen when you want to be fooled.
-
#252 Reply
Posted by
pickle9000
on 29 Mar, 2014 03:06
-
@TerraHertz
In all likelihood they would have had to find debris well over a week ago to make some use of the acoustic pinger.
-
#253 Reply
Posted by
ron
on 29 Mar, 2014 03:22
-
-
#254 Reply
Posted by
cimmo
on 29 Mar, 2014 03:24
-
I think it might help the discussion if certain speculations are clarified:
Pressurisation: The pressurisation controller allows the cabin to climb (and descend) at a much lower rate than the actual aircraft, typically less than 500 feet per minute. The aircraft itself can easily climb and descend at >3000 fpm. This is done to minimise discomfort to pax ears.
Typical cruise pressure differential is about 8psi (Pounds/Sq in) - equivalent to about an 8000ft cabin. That is why airliners are tubular with hemispherical pressure domes at either end and the underfloor cargo area cannot be independently depressurised. I'll leave it as an exercise to calculate how many thousands of tons of force is on the skin of the fuselage pressure vessel, but that force is why even a tiny bomb can have a big impact. This cyclical force once per flight is also a significant source of metal fatigue and is one limiter on the safe life of the aircraft.
Fly By Wire (FBW): Yes, the B777 does have FBW computer enhancements to the flight controls. But like all Boeings there is always a direct mechanical connection between the control yokes/pedals and the flight control surface hydraulic actuators. In the event of some runaway FBW/autopilot activity (or the hypothetical "remote control"), this direct steel cable control can overpower any autoflight control input.
This Boeing architecture is VERY different to an Airbus. In a typical Airbus, there is NO manual/direct connection - the joystick is used by the crew to inform the flight control computers what they want to do, and if the computer agrees then it moves the control surfaces. Even in degraded modes the pilot is still essentially flying through the computer.
I find it interesting that in all the events where modern airliners have allegedly been "remote controlled", not a single one has involved an Airbus, always a Boeing.
-
#255 Reply
Posted by
cimmo
on 29 Mar, 2014 03:42
-
If any of this is related to the ex-Malaysian Airlines 777 sitting in a hangar in Israel since Nov 2013
Your inability to provide links to back this up is duly noted.
http://www.planespotters.net/Production_List/Boeing/777/index.php?p=2
BUSTED! Israel Caught Housing Flight 370 'Twin' Airplane in Tel-Aviv! Youtube Video as "evidence" ?
Your idea of actual substantive evidence is amusing. So what if a 15 year old ex lease aircraft is sitting on the ground somewhere? It happens all the time. And FWIW, there are over 1100 B777 built, so a few sitting around waiting for the airline capacity demand to cycle up is totally normal.
I would have thought that participants on a site like this could have sense enough to avoid the tinfoil.
-
#256 Reply
Posted by
ron
on 29 Mar, 2014 04:01
-
I would have thought that participants on a site like this could have sense enough to avoid the tinfoil.
Dave asked TerraHertz for a link and I just posted it. Not my video.
-
#257 Reply
Posted by
pickle9000
on 29 Mar, 2014 04:12
-
I wonder if there is an agreement between the searchers to first notify to the families. I don't think that would be unreasonable. I suppose it could also be a free for all.
-
#258 Reply
Posted by
EEVblog
on 29 Mar, 2014 04:31
-
http://www.planespotters.net/Production_List/Boeing/777/index.php?p=2
Whoopdy-doo. Some entry in some database that says the plane is stored in Israel?
Who entered the data?
Who last actually saw it?
Does anyone know why it was bought, or why (presumably) it's being stored? - No.
Why is it so unusual that it's being stored in Israel?
Is is so suspicious that a company actually bought a plane from an airline? - No.
Who bought it? GA Telesis, that's who, a huge company that does exactly that, buys planes and uses them for parts or leases them as fleet planes etc, and does all sorts of stuff. For them to buy and store a plane is completely usual business. They have a billion dollars worth of assets under management.
This is a pretty dumb conspiracy theory, because if you step back and take a look, nothing unusual has happened here at all.
-
#259 Reply
Posted by
EEVblog
on 29 Mar, 2014 04:34
-
This Boeing architecture is VERY different to an Airbus. In a typical Airbus, there is NO manual/direct connection - the joystick is used by the crew to inform the flight control computers what they want to do, and if the computer agrees then it moves the control surfaces. Even in degraded modes the pilot is still essentially flying through the computer.
Yes, but there are several degraded modes, and as I understand it, the lower level mode you go, the less the computer actually tries to do. To the point of the computer making no decisions at all and simply relaying the stick commands directly. But strictly yes, there are no physical control lines from the stick, it's copper wires.
-
#260 Reply
Posted by
tom66
on 29 Mar, 2014 04:45
-
Airbus flight laws
http://www.airbusdriver.net/airbus_fltlaws.htm- Normal law - Plane is almost impossible to stall - can run with one failed flight computer
- Alternate law - Some protections omitted. Multiple failures required to enter alternate law (such as loss of two computers, loss of pitot reading, etc.)
- Direct law - No protections, all controls still available.
There is no true mechanical backup. In the case of complete systems failure, the aircraft will be uncontrollable aside from limited pitch and rudder control. However, such a failure is incredibly rare - it's more likely all four engines will fail for example. There are 7 redundant computers and Direct Law is possible with just one computer.
-
-
. There are 7 redundant computers and Direct Law is possible with just one computer.
I often wonder about the logic used to implement the redundancy in systems like this, and how they guard against faults in the redundancy hardware - I'm sure they know what they're doing but I'd be interested to know the technical details of how it's done.
-
#262 Reply
Posted by
firewalker
on 29 Mar, 2014 10:09
-
What operating system such a system depends on? VxWorks?
Is the computers of the redundancy array clones of each other?
Alexander.
-
#263 Reply
Posted by
SeanB
on 29 Mar, 2014 10:23
-
KISS. Discrete logic is very reliable compared to software, and you can have very reliable units providing you take care to protect it against overvoltage, power transients and such. If that dies you are up the creek and heading to a trip over Angel falls without a parachute. As well you have 3 mechanical actuators and simply rely on the fact that 2 can still control even if the third is not happy or is trying to hit an end stop. There are plenty of single point failure modes in the system, and they try their best to engineer the overall system such that those points are very rarely going to fail.
Airbus tried their best to mitigate bugs by requiring the computers to each have different logic designs internally and to have both different operating systems and different processors. The code in each just has to agree on outputs for inputs and do so within IIRC 200mS of each other, otherwise the redundancy management logic restarts the errant unit and lets it try again. Too many reboots and it is flagged and locked out. You can do all of that without having a processor just using discrete logic. Might be a programmable logic device, but it is NOT running code, just a configuration of internal logic.
-
#264 Reply
Posted by
EEVblog
on 29 Mar, 2014 10:44
-
Airbus tried their best to mitigate bugs by requiring the computers to each have different logic designs internally and to have both different operating systems and different processors. The code in each just has to agree on outputs for inputs and do so within IIRC 200mS of each other, otherwise the redundancy management logic restarts the errant unit and lets it try again. Too many reboots and it is flagged and locked out. You can do all of that without having a processor just using discrete logic. Might be a programmable logic device, but it is NOT running code, just a configuration of internal logic.
I would expect the final hydraulic drive control unit to have dedicated logic, and then dedicated logic for the Direct Law mode. The higher level ones would requires smarter control, so very likely not 7 of the same redundant control computers. As you say, pretty easy to implement in basic logic. i.e. each computer would have a "system ok" flag which the hydraulic control unit could interpret easily the priority inputs and ignore the units not working. The Direct Law would be just that, direct stick to hydraulic control. And there might be a master switch to override enable that in emergencies?
-
#265 Reply
Posted by
SeanB
on 29 Mar, 2014 12:04
-
No master override, as any fault situation where all computers are out implies the electrics are dead, the engines are dead, the batteries are all dead and the RAT is no longer deployed. No power or hydraulic pumps working either, and as the control surfaces are too heavy in flight for a stick movement to operate anyway a direct cable or linkage would not help anyway. That is the time you bend over and kiss your sorry ass goodbye. In that instance you better hope you filled in the IATA insurance form properly so next of kin can get paid out. I only thought about that one time.
-
#266 Reply
Posted by
EEVblog
on 29 Mar, 2014 12:46
-
No master override, as any fault situation where all computers are out implies the electrics are dead, the engines are dead, the batteries are all dead and the RAT is no longer deployed.
I was thinking more in terms of one of the flight computers going completely whacky based on sensor data or something, like what happened in QF32.
IIRC the pilot said he could always switch to Direct Law mode if he felt that the computers in the other modes might give problem?
This is on the A380.
-
#267 Reply
Posted by
EEVblog
on 29 Mar, 2014 13:05
-
-
#268 Reply
Posted by
amyk
on 29 Mar, 2014 16:08
-
But like all Boeings there is always a direct mechanical connection between the control yokes/pedals and the flight control surface hydraulic actuators. In the event of some runaway FBW/autopilot activity (or the hypothetical "remote control"), this direct steel cable control can overpower any autoflight control input.
Interesting fact: the
737 is the only remaining Boeing that can be flown even with all failed hydraulics.
-
#269 Reply
Posted by
tom66
on 29 Mar, 2014 16:23
-
Interestingly: even in the instance of total system failure the engines are self powering, which permits directional thrust.
Each engine computer is independently powered by two onboard generators above about 12% thrust. And the thrust levers/controls do not require cockpit power to control - they are routed via control lines straight to the engine computers mounted near each engine.
So even in the absolute worst case if the aircraft is still flying level you can probably land it if all 7 computers fail or if a total loss of all four AC busses (115V/400Hz), APU power, two emergency DC busses (28V DC battery backup), two emergency AC busses and two normal DC busses occurs.
Such a failure is extremely unlikely, but I suppose it's not impossible.
-
#270 Reply
Posted by
SeanB
on 29 Mar, 2014 16:40
-
Any fault taking all those busses out likely is a fire that burns through the main wiring looms, and will take out engine control wiring as well. Then the engines will either remain at previous power or will drop back to idle and stay there until fuel runs out. Depends on the EMU as to what it does with controller wiring fault. Worst case is as the wiring fails one side is commanded to full power while the other drops to ground idle, and then you have a self augering plane or it shreds itself apart in mid air.
-
#271 Reply
Posted by
Kjelt
on 29 Mar, 2014 17:10
-
Changing the search area again?
It now shows that they really don't have any clue where to look/search
![Face Palm :palm:](https://www.eevblog.com/forum/Smileys/default/facepalm.gif)
All is pure speculation.
What I find embarassing is that they still hold press conferences over these 1000km area changes, as if anybody cares where they are looking if they keep finding nothing.
-
-
I often wonder about the logic used to implement the redundancy in systems like this, and how they guard against faults in the redundancy hardware - I'm sure they know what they're doing but I'd be interested to know the technical details of how it's done.
I have done some of that work in software in the past. My not-so-favorite references were "Fault Tolerant Systems" by Koren Krishna and "Software Fault Tolerance" by Laura L. Pullum. Last one was very watered down so I did not like it, but it covers some important subjects like NVP (N-version programming) and Data Diversity pretty well. To keep things on subject I noticed that often describing fault tolerance in software they refer to solutions and techniques that were and still are extensively used by hardware designers but software allowed engineers to take it to another level if they want to of course.
-
#273 Reply
Posted by
pickle9000
on 30 Mar, 2014 01:13
-
How hard is the search?
Just a couple things to keep in mind.
- The Tai Satellite image with about 300 items covers 500 square kilometers, or put another way that 1 item per 1.6 square kilometers(ish). And that is the best lead to date, it does not even mean that is the place. If your searching you have to look where the satellite and data tells you.
- The amount that they suspect the plane flew south relates to the suspected speed and the last ping. Remember the plane could have went down at any time after the last ping but before the next. That is a full hour.
The search is obviously difficult, and no matter what, when or where, someone will say I told you it would be there.
-
-
Has anyone seen any information on the winds aloft data for that region? I mean, I can't imagine they launch balloons in the middle of the ocean and any numerical data that exists probably exists only to feed computations for populated areas.