There seems to be a lot of misconception about use of autothrottle.
The simple fact is, if you want to apply power manually when the autothrottle is in use, you simply disconnect it.
That is why Boeing provide a disconnect button on the side of each throttle which is pressed by the thumb as one applies manual power.
it is absolutely instinctive.
I don't see the problem.

So how come we know they existed?

Because there were 25 hours of data on the FDR, and investigators spotted the RadAlt glitches. The crews might not have noticed, never mind reported it.

PLEASE read the thread! The answer is there.

theamrad: The elevator would be maintaining ALT, V/S, a pitch attitude or the G/S and it should not be beyond the capacity of a competent crew to co-ordinate manual power adjustments with an A/P flown approach to preclude any 'pitch hunting' caused by over large variations in thrust. Though if A/P A/T approaches are the norm and manual power inputs the exception I could perhaps understand how unfamiliarity with required power inputs might lead to some rough handling, even absent turbulence or gusts. But then are we not in danger of producing pilots geared to flying the automation rather than the airplane with all the downside potential that creates when the automation for whatever reason fails to perform as expected?

I think that most semi-sensible, also the nonsense, theories have now been beaten to death several times over and it is time to wait for more info being released......the poor mods must need a rest from policing this thread :)

7777777777777777777777

Rananim,

May I politely suggest that you switch the brain on before pressing the keys on the keyboard.

The points I were making were facetious. Of course you don't need the items in my list, that was my point. I put a ridiculous one at the bottom (flashing the dome light if the aircraft sensed peril - bit of a half Hitchhikers Guide semi reference), but no, someone on Pprune thinks I'm being serious. Maybe a flashing peril-sensitive dome light was not ridiculous enough.

For your enlightenment, the NG is crude and simple. You will discover this when you get on to some more modern machinery, possibly the 777 or an Airbus. Incapable during non normal - you will also discover that on other aircraft when flying single engine you can do full flap landings, autolands, and fully automatic go-arounds. Nasty to fly = unstable in pitch (NG only, EG was great). It's only really obvious when you keep swapping types.

Just in case you still don't get it, I was saying that you can put as many beeps and warnings into all those automated systems as you like, but on this occasion it was the fault of the crew for not simply disconnecting and flying the plane.

Well that doesn't sound like let's look into all the factors that led a crew of 3 who accumulated many thousand hours, including an F4 pilot (does the F4 have autothrottles?) to end up short of 18R?

That sounds more like but on this occasion it was the fault of the crew for not simply disconnecting and flying the plane.

Let's call the board and save the taxpayer some bucks.

Please advise if my brain unexpectedly went into RETARD (FYI, I did get the subtlety) .

ant1,

Indeed - I then went on to mull over what we really would like to know. Exactly what happened in there to cause 2 or possibly 3 people to fail to realise that the a/c was not maintaining speed? We will have to wait for the CVR for that. It may well be that the "primary cause" of the accident is rooted in the behaviour in the flight deck during the approach - though there will be many secondary factors thrown in.

What ever the scale of automation or not, it´s up to the crews to check and manage the system. The whole idea of automatic flight is to give the pilots less of a work load and more time to concentrate on scaning the instruments. Now, if that goes to pot and the scaning goes out of the window...well, your guess is, as good as mine to where the blame might fall? Now, not undestanding how the automation works is another kettle of fish!!!

jackharr:

I see. Well, As far as I know the engines are not mounted in such way that they provide thrust that does not go through the center of mass. That means the thrust from the engines will produce torque which will try to rotate the aircraft. So what Safta probably experienced in the sim was immense back-trim in conjunction with the torque produced by the engines. Yoke fully forward was not enough to get the nose down. Maybe he (she?) could have recovered with not that much power... Anyway, I've learned that applying full power is the textbook way to recover from a stall. Hence, full power it is...

But this doesn't apply to you question as they've already lost control. A stall situation is the very definition of loosing control, isn't it?

RadAlt Problems Noted?
Quoting me in prior post:

"If a sensor, such as the radio altimeter, is giving you erroneous information, you have the option to disengage its circuit breaker."

Smilin' Ed:
"I don't think anyone ever noticed the RadAlt problems on this or any of the previous eight flights. If they did, I don't think they reported it."
--------

Have you and I been reading the same thread? The radalt going to -8 feet, as reported, will give GPWS "Too Low, Gear", which was reported somewhere on the approach. It also caused the AT retard. It may have cycled between correct and erroneous, causing multiple alerts and warnings; things such as "Too low, flaps." It would also display "-8" on the EADI, probably in a yellow meatball.

Had the crew pulled the Radalt circuit breaker after the first erroneous indication, the AT would not have pulled the throttles back again in the stall recovery.

GB

FORESIGHT is quite correct, it is VERY easy to disconnect the autothrottles and move them manually. EVEN if you forgot the disconnect button (which is easier than picking your nose), you could simply overpower the clutch mechanism and move them just fine.

RAINBOE...if that lear couldn't stop its climb using manual throttles, the pilot should lose his license.

Some pilots just need a huge amount of training and the discipline to keep things working right.

:ugh::rolleyes::rolleyes::rolleyes::}

737 AMM 22-11-00-146 (DFCS Functional Description)

Engage and Disengage Functions
24. The signal from the SMYD that shows a stall condition and activates the stick shaker must not be active.

So the answer to your question is yes; if the pilots had not yet pushed the control yoke forward, the autopilot is supposed to disengage when a stick shaker is activated by the SMYD.

100% Percent Please,
Sorry,I didnt realise you were being ironic.This thread moves quickly and I jumped the gun.Pilot error.
Smilin Ed,

No AoA indicator on 737 flightdeck.Two AoA sensors feed 2 SMYD computers and they use TE Flap position and TAI bias to determine what is known as the STALL ID(identification) AoA.If actual angle of attack>Stall ID AoA,the SMYD's activate the stick-shakers.The SMYD refines the STALL ID AoA by incorporating other parameters such as high engine thrust bias taking N1 input from the DEU's etc..

- have you ANY idea how long it would take to locate? Have you any idea how long one pilot (the Captain in this case) would be 'completely out of the loop' (ok - I know...), twisted around looking at a panel of 40 or so breakers? Have you noticed they probably did not notice it anyway?

It is an option for the 737NG to be displayed on the PFD. It's controlled (on or off) by your CDS OPC software.

AvEng - this is useful stuff! If only we had the AMM......................thanks for the info.

Rainboe: On the contrary. I'm fine with automation as long as it doesn't violate the KISS principle by making things more complex than they need to be for a given operation and as long as the specifics of any configuration doesn't inadvertently lead to the opening up of more holes in the cheese than the automation is intended to eliminate.

Your second para indicates that perhaps you may be guilty of the very thing you have roundly been castigating others over - namely not reading what people post.

As to the example you relate. If the Learjet was cleared to an altitude or FL above you then I would blame ATC for bad separation, if cleared to an altitude below you then I would respectfully suggest that we not promote the idea of A/T as a crutch for those who are incapable of managing their thrust manually.

AOA indications available to the pilots would be very useful in some situations like wrong weights or in this case pilots not increasing power on the glide slope and getting slow not being aware of it until stick shaker. Forgetting speed brakes out at night in a tight descending turn in a 727 is the only time in over 20,000 hrs I could have used it. Paying attention to flying the airplane at all times would make it not required.

We are entering an automated flight era that probably requires it along with idiot lights to bring attention to it like cars with oil pressure low lights in the 70's.

Quoting Rananim:

:ugh: I haven't been able to find this info yet. Can you reference the section? I've looked at the differences between the Smiths Autothrottle Computer (which is what the accident aircraft had installed by line number) and the integrated Collins autothrottle card (line numbers 1270, 1272, 1278 and on have the autothrottle computer integrated into the FCC - FCC A is used for autothrottle).

Could you possibly be seeing the difference in rad alt connection to the FCC's? The Honeywell FCC's are not dually fed (directly) rad alt. The Collins FCC's are. There is a Honeywell FCC software change that prevents one radio altimeter failure from causing an FCC to default to the 8 degree bank angle limit during an LNAV takeoff. This was standard ops with the Collins FCC's from the get go.

BOAC - I thought they did notice it was giving a false reading? This was mentioned in the first briefing wasn't it?

3001A: it means "one" does not "like" automation FGS!! Please read between the lines just a little.

737AvEng,
Section 22-31-00 pages 28,29,30 dated 10/10/06
3 schematics entitled AT SYSTEM-DIGITAL INPUT INTERFACE
One shows RA#2 having a direct(not via FCC b)input interface with AT CMPTR.
Effectivity is:
GEF 002, 003, 006, 007, 009, 011, 012, 028, 030, 301-305, 309-399, 713, 714, 717, 719, 721-726, 728, 729, 733, 735, 737, 805, 810, 811, 817, 821, 826-829, 833, 834, 837, 839, 840, 843,848, 859-861; GEF 812, 825 POST SB 737-34-1758; GEF 815 POST SB 737-34-1941

Other two schematics show input to FCC B.

Hope you can make more sense of it than I can.

Yes,I see AoA indicator is an option.I have not seen one installed myself.

Thanks 737 AvEng.

So what we could have is speed rolling back and AP trimming nose up until Stick shaker when AP disconnects and If no one has their hands on the controls the aircraft pitches up sharply.

I have seen things in the simulator that in real life you would not believe it could happen - especially with experienced pilots. In fact if someone wrote a book about the quite frightening lapses of flying "skill" seen in the simulator it would probably plunge the airline industry into despair as passengers head for other means of transport after reading the book. "Thinks - not a bad idea - might make a buck or two with the book!"

Two examples: Experienced captain fails to notice his ADI had jammed during a reversal turn in a holding pattern. This in a 737-200 sim IMC. He continues to apply wheel pressure in an attempt to "force" the ADI to come good, meanwhile ignoring or not noticing comparitor lights. The aircrft is rolled beyond the vertical and the nose drops until with a startled Asian oath he looks at the correctly functioning standby ADI and calls "Standby ADI failure". He then proceeds to pull the fast erection button to level flight. Needless to say we crashed. All this time the F/O with his 250 hours said exactly nothing while watching open mouthed at his own perfectly functional ADI.

Second. On take off in low vis, the captain failed to rotate at VR and eventually went off the end of the runway at around 190 knots still on the ground. Why? The F/O had been quietly instructed to call 80 knots as per SOP but to "forget" to call V1 and VR.

The object of the exercise was to ensure the captain rotated on his own ASI readings and not to act purely on a VR call from the PNF. Because invariably a VR call coincides with the other pilot also reaching VR, there is an oft seen tendency to rely on the PNF call of "rotate" as the initiation action. If someone omits to call "Rotate" it should make no difference. But here it did and that is not the first time I have observed this blind reliance on a support call before a specific action is taken.

In both examples given, the captains had over 10,000 hours and both F/O's were new with barely 250 hours CPL. And that is only two of hundreds of similar style examples I have seen in many years of simulator instruction on the 737. I can quite see how the Turkish Airlines accident could evolve with a minor problem with an autothrottle, causing jaws to drop and stunned mullet disbelieving looks and finally lack of corrective action in time to stop a crash.

That is the old (Smiths Autothrottle) install which has had both radio altimeters wired to it since inception on the first NG.

On the diagrams where you see with the #2 Rad Alt going to the right FCC, the next item to the right should not be the authrottle computer, it should say FCC-A underneath it as that's the Collins setup.

The SB's noted are for installation of the second FMC, so that's the change for GEF 812, 825 and 815.

What we've know is that the Smiths Autothrottle computer has both radio alitimeters wired to it. What we don't know is how the software inside uses the right radio altimeter input. I am still looking for absolute confirmation that my opinion is that the A/T computer looks at the left until it sends out NCD (Fail) and then switches to the right is correct.

What will be interesting to see is how fast the speed bled down from Alpha to stick shaker. I need to research alpha a bit more as this is the only info I have on it:

The autopilot (A/P), flight director (F/D), and autothrottle (A/T) alpha floor limit is the minimum speed available for airspeed control that will override manual speed selection or FMC commands. The limit is approximately 1.3 times the stall speed.

It would be interesting to know that autothrottle alpha protection is only based on one radio altimeter. When stating A/P above, it almost leads you to believe that the aircraft would abandon the G/S and pitch down to gain speed.

But the speed trim is more interesting:

The speed trim system gives automatic stabilizer trim for positive speed stability during low-speed high-thrust conditions. The speed trim is only operational when the autopilot is not engaged.
The stall detection circuit monitors the flap position and the angle of airflow. Near stall, the speed trim function trims the stabilizer to a nose down condition to allow for trim above the stickshaker AOA and idle thrust. The trim continues until the stabilizer gets to its limits or the aft column cutout position is exceeded.
If the roll angle from the ADIRU is more than 40 degrees, it opens an electronic switch and stops the speed trim signals.

So if we didn't have alpha here on the FCC, i.e. it kept trimming nose up to maintain G/S, when the stick shaker activated, the A/P should have disconnected and based on the above, speed trim should have started pitching the nose down. Correct?

Rainboe,

you didn't understand TM:s post did you?

The captain had 10000 hours so it's not a training issue, it's a complete lack of basic airmanship. No trainer can give him that during a conversion training.

Or where you just stirring the pot?

Tee Emm

Great stories and lessons. Please send me notice when your book comes out. :ok:

maybe I can add a couple or two.

The stuff that can be blamed on the PNF is unending

Hello "737AvEng . . .
 

Question for 737AvEng: In case you didn't have the time to read the entire thread, one of the questions hotly debated earlier was: Why would RA1 logic be programmed to send a Retard signal (flare mode?) to the A/T upon indicating -8 feet when auto land was not selected and the approach was flown only on one coupled A/P in GS capture mode? I'm not a 737 jock, but several contributors had mentioned, just for the record, that 1 A/P can in fact do a full coupled CAT-III landing. OK, but in this case, the engaged A/P was not in A/L auto land mode.

Vino . . . I want to hear a programmer's answer, not a pilot's answer. Yes, I'm a 74 pilot, with a question.

...It also means NOBODYS hand was on the TLs during an ILS approach.

Sorry if I seem a bit behind things, but reading 100 pages of technical bable is a bit too much.
I understand the throttle retarded ? How about disconnecting it and pushing it foward ?
A lot has been said about Airbus throttle not moving and how dangerous it is. Well, it happened to a Boeing.
This is just basic airmanship, period.
The fact that no basic action was taken is in my view both cultural and compagny ( training ) culture related.
Maybe it's not all that easy to scream " speed " when a turkish captain is screwing the pooch.
And no, you are not only the product of your compagny culture but also the product of your culture of origin.
Korean airlines has been extensively discussed on this site and in terms far less PC that didn't lead to the termination of the thread.

Cultural factor IS a factor. If I recall correctly, some Newcastle ( Australia ) university professor wrote a paper on the subject some years ago.

Pulling a C/B ??? .............no comment !

100 pages, nearly 2000 posts - and you are no further on than saying that on the balance of probability -

1) no.1 LRRA failed, and ...

2) no-one was properly flying the aircraft, and therefore ...

3) tragically, the incident was almost certainly avoidable ...


JD
:bored:


This thread is boring ... very boring ...

And that is the crux of the matter. The rest is just window dressing!

Late logging parttly responsible?
 

An article in the Danish Magazine "Ingeniøren" (The Engineer) claims that the Dutch secretary general of the AEI (Aircraft Engineers International) Fred Bruggemann has said that the Investigation Board has spoken to several Turkish pilots who have flown the crashed plane in the days before the accident. They admitted that they had not reported the defect altimeter as they should have. He added that the AEI has contacted the plane mechanics in Turkey who confirmed that the log book did not contain anything about a defective altimeter.
The article describes the course of the accident as it is already mentioned here in several posts.
It says also that the plane crashed while the pilots were having a conversation. They noticed the critical situation and gave full power but it was too late.
It further says that the black boxes contained data and audio from the last 25 hours. That is several flights over several days. During the last 8 flights, the altimeter was defect a total of three times in the same way and in the same situation, including the last time, where the plane crashed. The pilots in those cases also noticed the defect altimeter but did not do anything about it.
Fred Bruggemann said that AEI has warned often about late logging.

(link to article in Danish only)
Groft sjusk med fly-logbøger var årsag til flystyrt i Amsterdam*| Ingeniøren

Usually there are 4 arms in a cockpit, specifically in this case there were 6. Unless 2 people were incapacitated.

hands on TL
 

boeing drivers (a.o.) often have complained the airbus TL's are not moving but if your hands are not on them or you are not responding to their abnormal movement then what's the use?

RA's fail all the time, equipment is prone to faillure, that's why we are on board and will be for the foreseeable future. but if we don't respond to faillures or simply take too much time because we were not prepaired ... the GA switches are on the TL ... there is a reason for that.