Possible Causes of the Boeing 777 Crash Landing at Heathrow?

[Garth]

This is the reason why there are airworthiness directives to aircraft owners. All aircraft have problems that only show up after many hours of flight time. Its the reason why a database is kept and notices are sent out when a problem is seen among many airplanes of the same type. If this were a problem on all 777s, an AD would have been issued and sent out.

Not if this were the first time the problem revealed itself.

So to say that there is an 'unresolved computer glitch' which could cause a catstrophe at any time is nothing more than uninformed nonsense. As of now, there is no issue with all 777s, so don't expect them to start dropping out of the sky.

One nearly did.

A 777 is one of the best airplanes out there.

A touching example of faith placed in modern software, I hope it doesn't prove to be a delusion.

Garth

 

[grant9076]

I posted this additional information in the Air Crash thread in the Mechanical and Aerospace Engineering forum.

http://aviation-safety.net/database/record.php?id=20080117-0

http://www.aaib.dft.gov.uk/latest_news/accident__heathrow_17_january_2008___initial_report.cfmInterestingly, there was a problem with the 777 in Australia.
http://www.airlinesafety.com/faq/777DataFailure.htm

I wonder if this failure had any commonality with the BA038 problem.

Until the flight recorders are analyzed, it's too early to tell.

Although I have not flown a B-777, this shows the classic indications of a (partially or fully) blocked pitot probe which can happen in any aircraft. If that is the case, the airspeed indications will increase with altitude and will eventually show excessively high airspeeds. The automation will respond by increasing the pitch attitude in an effort to slow down. This will aggravate the erroneous high speed readings while the aircraft is slowing down and approaching a stall and will activate the stall warning system simultaneously with the overspeed indications. If the aircraft is descending, the reverse will happen and the airspeed will show excessively slow indications.

If this is the cause (I am not saying that it is), then it is totally unrelated. However, I am shocked and disappointed that the crew elected to re-engage the automation after experiencing a known flight instrument malfunction. This is an extremely foolish and dangerous choice in any aircraft that is suffering any flight instrument malfunction for any reason.

My estimation of a 3000 ft/min impact is a bit more pessimistic, I would not expect to walk away from that.

I share your pessimism. The wing tanks and/or center tank are containing fuel and/or fuel vapor. Neither one responds very well to tank ruptures.

 

[Astronuc]

Although I have not flown a B-777, this shows the classic indications of a (partially or fully) blocked pitot probe which can happen in any aircraft.

Could the pitots iced up? What about redundancy of pitot's or an independent system? GPS or doppler radar.

Problem with the autothrottle or throttle control?

Any news on the blackbox?

 

[Andre]

Could the pitots iced up?

Unlikely, but in this case it would not make sense. As climbing with a blocked pitot tube is dangerous for over speed indication, the opposite is true for descending, the decreasing airspeed indication prompts the system to give more and more power, leading to comfortable high airspeeds for flying purposes. The high sink rates confirm that this was not the case.

What about redundancy of pitot's or an independent system? GPS or doppler radar.

Not really that gives you ground speed which is off about four variables with indicated or pressure airspeed required to control systems, the most important: altitude or density pressure and wind.

 

[grant9076]

I get the feeling that Astronuc is referring to the incident in Australia, while Andre is referring to the accident in London.

Could the pitots iced up?

I do not know the whole story but pitot icing will cause the symptoms that I just mentioned.

What about redundancy of pitot's or an independent system?

Yes. Reduced Vertical Separation Minima (RVSM) rules do require the pilots to have separate and independent pitot-static systems. However, if conditions are conducive to icing up of one system, then they are also conducive to icing up of the other. The standby flight instruments are independent of these 2 but can be iced up as well.

Problem with the autothrottle or throttle control?

When I use the term "automation", I am referring to both the autopilot and the autothrottle. The autothrottle will follow the erroneous readings just as blindly as the autopilot will, which is why it should not be used.

GPS or doppler radar.

These provide groundspeed info which is not useful to the automation. However, there is the technique of using minimum ground speed which is used by some experienced pilots. It involves figuring what the final approach true airspeed should be and then subtracting the headwind (or adding the tailwind) component that is reported by tower to give you the mimimum groundspeed that you should be maintaining on final approach. I find this technique to be very useful for 2 reasons:

1. In addition to known pitch and power settings, it is useful for safely landing the aircraft in the event of a known pitot-static malfunction.

2. It gives an extra layer of protection against windshear. This is particularly true for types of windshear that are not detectable by the onboard predictive windshear warning systems.

 

[Cyrus]

Not if this were the first time the problem revealed itself.One nearly did.
A touching example of faith placed in modern software, I hope it doesn't prove to be a delusion.

Garth

The 777 has over 3.6 million flight hours based on a boeing spokesman. Your argument is pathetic at best. Each 777 has different electronics packages and engine combinations. To hint that its a problem with the 777 in general shows a total lack of understanding about aviation.

Its more likely that a particular engine/avionics combination is having issues.

 

[Cyrus]

Still underestimating the inertia. What for instance might have been the order of magnitude of the vertical velocity on impact of this 777?

Given: initial altitude 700, distant to cover 2 miles minus undershoot say 1,75 nautical miles, Say that the speed had been reduced to 120 knots then time to impact would be 52.5 seconds, covering that 700 feet to lose with an average sink rate of 800 feet a minute.

Grant says:



hence the 800 ft/min order of magnetude is four times the energy of the normal no flare impact. My estimation of a 3000 ft/min impact is a bit more pessimistic, I would not expect to walk away from that.

Again, 3000 FPM is only a 40 mph crash in the vertical direction. No big deal for such a big structure to absorb.

 

[Garth]

The 777 has over 3.6 million flight hours based on a boeing spokesman. Your argument is pathetic at best. Each 777 has different electronics packages and engine combinations. To hint that its a problem with the 777 in general shows a total lack of understanding about aviation.

I wasn't saying that it was a problem with 777s in general, I think that was Art.

Actually I have held a PPL since 1966, so I might know a little about aviation.

Its more likely that a particular engine/avionics combination is having issues.

Agreed that the problem may be with "particular engine/avionics combination" but that configuration should be treated with suspicion until the cause of this incident is discovered.

Note: As far as an pitot icing problem is concerned the weather conditions near ground level in London on the day in question were well above freezing, that does not mean of course that the aircraft might have suffered icing conditions at altitude en route from Beijing.

Garth

 

[Cyrus]

I wasn't saying that it was a problem with 777s in general, I think that was Art.

Actually I have held a PPL since 1966, so I might know a little about aviation.Agreed that the problem may be with "particular engine/avionics combination" but that configuration should be treated with suspicion until the cause of this incident is discovered.

Note: As far as an pitot icing problem is concerned the weather conditions near ground level in London on the day in question were well above freezing, that does not mean of course that the aircraft might have suffered icing conditions at altitude en route from Beijing.

Garth

I think the pilots would have noticed if their altimeters still read 33,000 feet on final due to pitot blockage that occurred at altitude. Flying in lower air at above freezing temperature will melt pitot ice most of the time. Its the reason why your told to fly lower if you suspect icing.

As a pilot, I find such statements as:

A touching example of faith placed in modern software, I hope it doesn't prove to be a delusion.

As nonsense, and not helpful to informing the public. Modern software has made flying much safer. Aircraft avoidance, GPS maps with ground speed, multiple redundancies, in air weather and terrain maps, full glass panel displays with systems information, you name it -all thanks to modern software.

 

[grant9076]

Note: As far as an pitot icing problem is concerned the weather conditions near ground level in London on the day in question were well above freezing, that does not mean of course that the aircraft might have suffered icing conditions at altitude en route from Beijing.

I was referring to possible causes of the incident in Australia and explaining why I think that it has no correlation to the accident in London.

Note: It is static port icing that causes the altimeter readings to "freeze" (no pun intended) while pitot icing causes the airspeed indicators to act like mis-calibrated altimeters. Again, I am saying this for information only and I do not believe that it has any relation to the Heathrow accident.

 

[Cyrus]

Grant, how fast do blocked Pitot and static ports take to melt when flying in above freezing conditions, if you don't have pitot heat available (in something like a 172)?

 

[Garth]

I think the pilots would have noticed if their altimeters still read 33,000 feet on final due to pitot blockage that occurred at altitude. Flying in lower air at above freezing temperature will melt pitot ice most of the time. Its the reason why your told to fly lower if you suspect icing.

I would have thought pitot tube icing would have affected the airspeed indicator...

Not knowing the particular configuration of this 777 I am not sure, but would not the pitot tube have electric de-icing in any case?

The PFD would indicate radio altitude, which of course would not be affected by icing.

As a pilot, I find such statements as:

A touching example of faith placed in modern software, I hope it doesn't prove to be a delusion.

Aircraft avoidance, GPS maps with ground speed, multiple redundancies, in air weather and terrain maps, full glass panel displays with systems information, you name it -all thanks to modern software.

And as a pilot I would prefer not to place total faith in electronics, especially the software.

I was at a lecture at Queen Mary College, (University of London), about the Space Shuttle in the 1980's, pre-Challenger disaster, and asked why a recent launch had had to be postponed because there had been zero wind velocity at the Cape. The answer was that the computers on board had been designed with the rest of the craft around 1975 and they had 1970's memory shortage. As there was seldom zero wind velocity at the Cape they were programmed for a wind speed of 10 - 70 knots. (If I remember correctly).

I asked, "Why not replace the computers with modern ones with more memory?"
The answer came back, "Because we trust the software!"

The software in modern computers is so complex that, although it can be de-bugged and tested to the n^th degree, you can never be sure that you have got rid of all the bugs.

My intuition is that the investigation will discover that this incident was caused by one of those undiscovered bugs, but then what do I know?

Garth

 

[Cyrus]

I would have thought pitot tube icing would have affected the airspeed indicator...

Yes, sorry. I switched the two by accident.

Not knowing the particular configuration of this 777 I am not sure, but would not the pitot tube have electric de-icing in any case?

Im sure it would on an A/C that large.

And as a pilot I would prefer not to place total faith in electronics, especially the software.

As this is the first accident on a 777 in 12 years of service, its a very very good system and A/C.

I was at a lecture at Queen Mary College, (University of London), about the Space Shuttle in the 1980's, pre-Challenger disaster, and asked why a recent launch had had to be postponed because there had been zero wind velocity at the Cape. The answer was that the computers on board had been designed with the rest of the craft around 1975 and they had 1970's memory shortage. As there was seldom zero wind velocity at the Cape they were programmed for a wind speed of 10 - 70 knots. (If I remember correctly).

I asked, "Why not replace the computers with modern ones with more memory?"
The answer came back, "Because we trust the software!"

The software in modern computers is so complex that, although it can be de-bugged and tested to the n^th degree, you can never be sure that you have got rid of all the bugs.

My intuition is that the investigation will discover that this incident was caused by one of those undiscovered bugs, but then what do I know?

Garth

I don't know how you are comparing 1975 software with modern fly-by-wire systems. Thats over 33 years of improvements. I don't doubt it was a bug, as I said so myself. But blanket statements such as 'faith in electronics' makes no sense. A large portion of airliners are fly by wire for many years. It works, over and over, again and again. One time in 12 years something goes wrong and people get all worked up. Amazing.

I would venture that the space shuttle would have a fly by wire system as, if not more, complicated than airliners today.

 

[grant9076]

Grant, how fast do blocked Pitot and static ports take to melt when flying in above freezing conditions, if you don't have pitot heat available (in something like a 172)?

This is an excellent question and I honestly do not know the answer.

However, I think that the 2 biggest factors which affect this are the type/severity of the icing, and the total air temperature TAT. In turn, the TAT is a function of the static air temperature and the true airspeed.

On a side note: Although lower altitudes are generally warmer, this is not always the case, especially with a strong temperature inversion like the one I encountered 3 days ago on the descent into Dallas Fort Worth. The TAT was showing 11 degrees celsius at 12,000 ft (although we were doing 300 Knots indicated airspeed). By the time we rolled out on the ILS course at around 3,500 ft, the TAT was showing 3 degrees celsius.

This brings up another side note. We turned on the engine anti-ice at 12,000 ft because engine icing can occur at total air temperatures as warm as 10 degrees celsius (depending on the engine) if there is any type of visible moisture (including clouds or visibility less than a mile) present.

A third note: Every 777 pilot that I know describes it as possibly the best designed aircraft that he/she has ever flown.

 

[Garth]

I don't know how you are comparing 1975 software with modern fly-by-wire systems. Thats over 33 years of improvements. I don't doubt it was a bug, as I said so myself. But blanket statements such as 'faith in electronics' makes no sense. A large portion of airliners are fly by wire for many years. It works, over and over, again and again. One time in 12 years something goes wrong and people get all worked up. Amazing.

I would venture that the space shuttle would have a fly by wire system as, if not more, complicated than airliners today.

I was simply saying that NASA were sure of the simple 1970's software because the lines of code could be de-bugged 'by hand'. Modern software is so complex involving millions of lines of code that you cannot be sure that every bug has been dealt with.

I am sure modern fly-by-wire systems have been robustly tested and reliable, however we will see whether this case is an example of an unknown bug being missed until now.

Garth

 

[Garth]

Another possibility is that the emergency was caused by another electronic device, such as a mobile phone (Hi love I'm almost home!), interfering with the normal operation of the flight control computers.

Garth

 

[Cyrus]

Could you explain more on that please?

 

[grant9076]

I just recently saw the latest press release from the UK Air Accidents Investigation Branch. The report indicates that there was adequate fuel in the tanks, and that the autothrottles and the electronic engine controls functioned perfectly as advertised.

They are now investigating all possible causes including the entire fuel flow pathway from the tanks to the engine fuel nozzles.

http://www.aaib.dft.gov.uk/latest_n..._17_january_2008___initial_report_update.cfm"

 

[chroot]

The MythBusters did an interesting show on cell phones aboard aircraft. Quite surprisingly, basic nav radios, the kind you'd find in Cessnas, are DRAMATICALLY affected by cell phone signals. On the other hand, they were unable to get a properly shielded jet aircraft's avionics to respond in any way at all to cell phone signals. They weren't legally allowed to use a cell phone while in a flying aircraft, though, so they had to do their tests while powered up and parked on the tarmac.

Either way, I'd put the odds of a cell phone causing a crash at about one in a billion.

- Warren

 

[Cyrus]

The point I was getting at is that a cell phone would affect navigation. That does not play a roll in the pitot probe, nor the physical connection between the autopilot and engines. I don't see how that could possibly be an issue.

BUT, thanks for that warren. I didnt know it was an affect on the nav that much in a cessna.


EDIT: You know, open up your sport's pilot catalog. They sell cell phone adapators to hook up to your headset. Hmmmmmmmmmmmmmmmmm...

http://www.sportys.com/acb/showdetl.cfm?&did=19&product_id=7145

http://www.sportys.com/terryc/images/4149al.jpg

 

[Art]

How do you reconcile these two statements?

So to say that there is an 'unresolved computer glitch' which could cause a catstrophe at any time is nothing more than uninformed nonsense

I don't doubt it was a bug, as I said so myself.

 

[grant9076]

How do you reconcile these two statements?

As I mentioned in my previous post, the evidence clearly shows that there is no electronic malfunction. The investigation is now directed at fuel blockage/contamination.

I think that the point that you are trying to make with this post is now irrelevant.

 

[Cyrus]

How do you reconcile these two statements?

I was very clear in pointing out that each 777 is as unique as the next in terms of the combination of engine and avionics packages and a problem in one particular A/C is not necessarily common to all aircraft of that type. I also mentioned Airworthiness directives being issued if it were common to many A/C in the past 12 years.

I also said, it was either a computer issue, or possibly fuel starvation.

I blame those British Rolls engines. They should have put Good American G&E engines.

 

[Cyrus]

As I mentioned in my previous post, the evidence clearly shows that there is no electronic malfunction. The investigation is now directed at fuel blockage/contamination.

I think that the point that you are trying to make with this post is now irrelevant.

Im guessing the tanks on large aircraft use crossfeeding; but, I would expect each tank to have at least two selector valves incase one becomes blocked. I would expect the tanks to become blocked either because the fuel expanded at a valve and froze or the fuel was contaminated. Could the fuel really have been that contaminated though? Its very unlikely. He should have used his fuel sample cup pre-flight

 

[russ_watters]

As I mentioned in my previous post, the evidence clearly shows that there is no electronic malfunction. The investigation is now directed at fuel blockage/contamination.

I think that the point that you are trying to make with this post is now irrelevant.

Read it again. The article says the autothrottle was working properly both before and after the reduction in speed. It doesn't say it was working properly during the reduction in throttle. I think that means it is still an open question.

 

[Cyrus]

Recorded data indicates that an adequate fuel quantity was on board the aircraft and that the autothrottle and engine control commands were performing as expected prior to, and after, the reduction in thrust.

Not really an open question. I think you misread it. If it was working even after the thrust was reduced, its not the computers fault, or it would have spooled back up.

 

[grant9076]

Read it again. The article says the autothrottle was working properly both before and after the reduction in speed. It doesn't say it was working properly during the reduction in throttle. I think that means it is still an open question.

Actually Russ, I saw an even more detailed report at my airline's website in the section for pilots only. Although the information concerning the accident is not sensitive, it is in a security restricted area and I am thus unable to share the link with anyone here.

However, the information that I gathered showed what happened with the throttles and the fuel control systems during the incident. It showed that the Electronic Engine Controls (EEC's) functioned exactly as advertised. The data records also showed that the fuel control valves opened to the proper settings for the commanded thrust during the incident. However, the records showed that only a fraction of the fuel flow was received (just enough to keep the engines slightly above idle). This is why the investigation is now turned to fuel blockage/contamination.

This is an extremely brief synopsis of the info that I saw but it is why I no longer consider it to be an open question. P.S. The engine's responses were not exactly identical but similar to each other (according to the records). I did not see any indications of cross feeding.

 

[Garth]

Thank you Grant for that inside information.

If it was a fuel blockage the problem would be to explain how both independent fuel systems, from two independent wing tanks and one common central tank, dried up at the same time.

Not knowing the detail of the 777 fuel tank installation I am dependent on others', such as your own, expertise. Would there be a second set of fuel control valves, close to each fuel tank, ready to shut off the fuel in case of emergency? If so might these have been all inadvertently activated by an electronic glitch?

Garth

 

[grant9076]

Not knowing the detail of the 777 fuel tank installation I am dependent on others', such as your own, expertise. Would there be a second set of fuel control valves, close to each fuel tank, ready to shut off the fuel in case of emergency? If so might these have been all inadvertently activated by an electronic glitch?

A very good question. However, I doubt that it would be the case and here is why:

For every boeing aircraft that I have flown, each one has a separate fire detection and protection system for each engine. In addition, each fire detection system has 2 separate and independent loops along with a fault monitoring system which disables the faulty loop. If it detects a fire, it warns the pilots with audiovisual indications and unlocks the fire handle which allows the pilot to pull the fire handle as part of the fire protection. Otherwise, the pilot will have to manually disengage the locking devise for the fire handle in order to pull it. Pulling the fire handle will cutoff fuel, hydraulics, and bleed air in addition to tripping the respective generator off line.

The chances of all 4 separate and independent fire detection loops (2 per system) failing simultaneously is more remote by several orders of magnitude than the fuel in both wing tanks being contaminated. In addition, the chances of the fire handles pulling themselves or the results happening without them being pulled is doubtful.

On a side note, I did have doubts about an Electronic Engine Control (EEC) malfunction from the very beginning and here is why:

Although I have not flown the 777, I know that the B-737-800's EEC's are very similar in operation to those of the B-777.

For the B-737-800:
1. Each engine has its own separate and independent EEC.
2. Each EEC in turn has 2 separate and independent channels with a self monitoring system which disables the faulty channel.
3. If both normal channels fail, the EEC has 2 alternate modes.
4. The EEC's are designed to never share single source data.
5. In the alternate modes, the EEC's will command thrust settings which are always greater than or equal to those commanded by the normal modes for the same throttle position.

I also consider the chances of all of these failing simultaneously to be remote.

 

[Garth]

Yet obviously something happened to both engines/fuel systems simultaneously.

Garth