The difference is that with speed off the A/T might come forward due to Min Speed Reversion if they were not designed to retard. So the feature is not there "in case you forgot to retard", it is implemented because of what I've written in bold.

A/T come off 2s after touch down.

BTW, no single channel to touch down, right?

The only allowed possibility to touchdown with A/T armed, other than dual channel, is speed off, A/P off, right?

• The AT is a complementary system to the AP and may be used with or without the AP.
• The 737 is certificated for autoland using dual APs; the AT, if engaged, will retard at 27ft.
• The 737 is also certificated for single AP approach to a minimum approach altitude (80 ft … cert authorities/737 versions may differ). The AP has to be disengaged for landing, but the AT may remain engaged and will retard at 27ft. Presumably this option enables a single AP MEL auto approach.
• If a pilot elects to fly a manual approach with AT, it may remain engaged and will retard at 27 ft.

Why have so many AT options … ? Because over the years pilots thought that it would be a good idea – it reduced workload, made life easy, wouldn’t it be nice if … etc, etc, and Boeing attempted to satisfy the widest possible market and retain a good MEL – which is also operator driven.
It appears that it is the method of interfacing the inputs and components of the FGS to provide these options, which has left a loophole in operations in the event of malfunction (but not necessarily for a failure).
You can keep some of the people happy all of the time, but … … … and that’s the point where opportunity for mistakes, misinformation, and false knowledge arises.

Speed reversion is an alternative pitch mode providing limited protection from abnormal low speed. It is normally engaged automatically from another mode. It is not available from GS.
GS mode is a AP ‘path on elevator’ mode, where the AT controls the selected speed datum. A speed reversion would require AP ‘speed on elevator’ mode, where a climb or descent is determined by the thrust setting. Reversion from GS would require a high thrust setting to ensure a climb. This logic is already built into the TOGA mode which provides a pitch up and an AP ‘speed on elevator’ climb, commanding the appropriate thrust setting on the AT.
Speed reversion is not an Airbus 'alpha floor' type of function, which is a higher order safety protection and operates independently of the AP/AT.

If someone wanted to end the thread on a high note,this would be it.Clear,concise explanation.

safetypee:

Presumably this option also assumes a correctly functioning RA#1 with no redundancy?

Ok, RB, got it! In your B738 there is no separate autoland "A/L" or "APPR" selector in order for the A/P [and A/T] to do full autoland. Understood.

Downin3Green - couldn't agree with you more!!! Right on. Nowadays, those things that stick out of the center quadrant, often referred to as thrust levers, or throttles, really should be renamed as hand rests!

Pending release of further info from the investigators may I respectfully request that this thread be temporarily locked.

It would:
a. permit many (most) to read all the current posts;
b. stop the circular arguments.

As a former Tower controller, I do wonder what I might have said to a pilot with decreasing ground speed (well below what one would have previously seen or might reasonably expect at that stage of flight).

Faulty RA
 

This is what the MEL states, and why in retrospect it is important to check this even before commencing, or even to execute a missed approach, to get the full picture with respect to downgraded equipment on an A/C. As most pilots does`t know the contents of the MEL by recall/Memory.

:With radio altimeter(s) inoperative, do not use the associated autopilot or Autothrottle for
approach and landing.


Still we do not know if this particulare fault had been adressed by the flight crew in this accident before or during the approach. It might even been this fault that had pussled them, and taken their attention away.

Norwegian - you are way off the rails. They were NOT using the associated autopilot. There is no restriction in the MEL to which I am attuned to use of autothrottle. Perhaps there should be, although it would not have affected this accident since the radalt was not an MEL item.

You don't get it, SC, do you? It was NOT an MEL item. Read the thread. In this case the MEL is not relevant.

Even if it had been, you above all should know that an MEL is company specific and some NG MELs say

3. With radio altimeter(s) inoperative, do not use the associated autopilot for approach and landing.

Note the missing bit! There's a job you should get your teeth into.

missy

I believe this has been mentioned earlier, however it assumes that you have little else to do other than physically watch ever single arrival. Surely at an airfield such as EHAM this is just not feasable.

Out of interest what would have you said? , bearing in mind it may further confuse an already confused and disorientated cockpit.

Rainboe:

Simple question: Why do we have MSAW?

Ah Sorry! Of course, you have no idea what MSAW is! It's NOT the controllers job to fly the aircraft. Hence, MSAW is unknown to you, or forgotten because it's a very, very silly system).

Let me refresh you memory (either you like it or not):

MSAW=Minimum Safe Altitude Warning systems

From SKYbrary - MSAW i quote:

Minimum Safe Altitude Warning (MSAW) is a ground-based safety net function of an ATC radar data processing system. MSAW is intended to assist in the prevention of controlled flight into terrain accidents by generating, in a timely manner, an alert of a potential or actual infringement of a minimum safe altitude. The MSAW function, using software in ATC computers, alerts ATCOs when radar returns indicate an aircraft penetrating or about to penetrate the minimum safe altitude.

--End of quote--

MSAW is of course a total waste of money as the ATC operator is not supposed to fly or tell the pilots how to fly. Go figure! Heck, the next they'll come up with is MSSW (Minimum Safe Speed Warning).

SC Let's, at least, get this right.

ARINC 707-6 Document Information: Title, Radio Altimeter. Aeronautical Radio Inc. Publication Date: Oct 30, 1998

And this, typical -

1. PRODUCT OVERVIEW
The ALA-52B Radio Altimeter (RALT), designed and manufactured by Honeywell’s Aerospace Electronics Systems. The ALA-52B RALT, a lightweight, solid-state, airborne altimeter provides accurate digital height measurements above terrain during aircraft approach, landing, and climb-out phases. The operating range is –20 feet to +8000 feet.


http://www.honeywellcommnav.com/docu...l/52b_tech.pdf

SC :ugh: I suspect that I've installed, from scratch, more RadAlts than you. The zero reading should be seen with oleos extended, bogies in trail, and the trailing wheels just touching. Understand?

When we have finished redesigning the Radio Alt. we had better turn our attention to the Auto Throttle. If it had malfunctioned in such a way that it set a low RPM or throttle closed the result would have been the same.

Things fail – We usually sort the problem. (Usually, since there is no denying that a wing coming off would be challenging) This was a case where there really wasn’t much of a problem to sort. SOPs/Standard calls should have taken care of it. (Assuming that the sequence of events was as reported.)


Yes I have read the whole thing

Each autopilot requires its own radio altimeter by design requirements. They are independent sensors linked to their own autopilot and are not allowed to be crossfed into other autopilots.
Also rad alts can have a negative reading on the ground. They are calibrated to read zero when the main landing gear contacts the runway in the landing configuration. Once the aircraft settles on the ground and the landing gear oleos are compressed, you will have a negative reading.

Takeoff: LNAV is available and commonly selected above 400'RA. Ist clue.
In descent "2500" EGPWS call from RA. 2nd clue.
On finals "1000" call from RA.

I think all these come from LHS RA. There may be different customer options for call outs, but the RA can be seen to be in use before the final approach. There might not be a u/s flag, but there are various scenarios where it can be seen to be correct yes/no. If so, is the concentrating on lack of tech log entry being so significant not a slightly smelly fish.

Maximum and biggles225, I am so glad you have got your coats. Could you kindly persuade most of the other posters here to get theirs as well, so that the Mods can lock the door behind them and the rest of us can have a proper discussion on how 3 pretty experienced pilots can fly an aircraft, with no major defects, into the ground and kill themselves. Thank You.

Too Many Straws
 

We all have heard of the straw the broke the camel's back. This appears to be another case of too many straws:

Rushed approach in IMC.

Turned onto final approach above glidepath, requiring throttle-off descent.

Radalt #1 locks onto false signal path and puts out -8 feet, valid.

GPWS sounds off with "Too Low, Gear", due to radalt #1.

Capt. EADI displays -8 feet, probably in a yellow meatball.

AT goes unnoticed into Retard mode, the throttles already pulled back in dispatched descent to catch the glidepath.

What we don't know yet is the "conversation" and other distractions being loaded onto the PF's brain as well, to the point he fails to react to decaying airspeed.

GB

My FCOM2 is dated May 2008

It doesn't say Min Sped Reversion is not available on GS, contralyly to another older book I've been shown

Mine also specifically mentions "alpha floor"

On top of that FCTM says:

The primary feature the autothrottle ARM mode provides is min speed protection in the event the airplane slows to min maneuvering speed.

Whether there is speed reversion on GS or not, I'll surely check it next time I go to the sim, I disagree with:

IMHO it only takes the A/T to add power to regain speed, so no A/P "on speed mode" (pitching down) on GS of course.

... and regarding the A/P wiring comments some posts ago, are you sure the A/P automatically disconnects on stick shaker or is it more like A/P cannot be engaged with stick shaker on?

see FCTM:

If an approach to a stall is encountered with the A/P engaged, ... If autopilot response is not acceptable, it should be disengaged

Further down:

Anytime the airplane enters a fully developped stall, the A/P should be disengaged ...

This warning does not apply to decaying airspeed or ground speed [Propellor planes fly slower than jets]; and when the airplane went more than 1 dot below glide path and had dropped out of the sky less than 2 miles short of the runway, this system couldn't alert the tower controller of an adverse flight profile trend in time to have changed the outcome of this crash.

Let me start provocative:

Being a pilot is great. They wear nice uniforms, enter the cockpit, press some bottons, start, press some more bottons, land and that's it. Really? Partly yes (don't shout too much before reading to the end) - as long as everything is working fine.

Real life is different. An aircraft has today a lot of automatics as a result of thousands of man years of engineering - and it is getting more and more. All these automation tools make the plane easier to fly and safer to fly - as long as everything is working fine.

The pilots have to control these tools, to check weather the output of these tools (e. g. thrust level) is consistand with the actual situation. Is one of these tools inoperative - no problem, this will be flagged and other systems (including pilots) can react. But when they are still operative and produce nonsense it is getting more complicated. To realize there is a problem and to come to the right conclusions within a few seconds... that is the problem and that is why the guys in row 0 don't have to pay for their seats but are payed.


Rainboe put it perfecly into very few words some days ago: "Pilots have no right to 'rely' on automatics."

Seems he knows Tom's first principal theorem of enineering: "Only nonexistent technical equipment will not fail" (Originally: "Nur Technik, die nicht vorhanden ist, wird nicht ausfallen.") That means pilots always should expect a failure.

A last try to stop the discussion about different radalt readings:

Of course it would be possible to check the two radalt if the reading is aproximately the same. Ask Mr. Boeing or one of his enineers why they don't check for discrepancies. I'm sure they have one or more good reasons.

... or maybe not. Sometimes it takes an sad event like this one for things to evolve in the right direction.

This aircraft type is over 40 years old and apparently has evolved very little over that period. Boeing continue with their philosophy of putting a sleepy fallable human in charge of a poorly designed machine that will not save you when you screw up.

...Lay enough traps in my path and I am sure to fall into one of them. I'm only human.

Do you honestly think there is a problem with the B737 design?

AirDisaster.Com: Statistics

The Boeing 737 has one of the best safety records of any aircraft flying today.

Turkish Airlines has one of the worst safety records of any airline flying today.

While the Boeing design may be not perfect and there is room for improvement, it has evolved quite significantly. I'm afraid (AFAYC) and glad (AFAIC) that the fallable human will remain in the cockpit in the foreseeable future.

While traps have to be identified and corrected, crew performance standards have to be maintained.

"The ultimate result of shielding men from the effects of folly is to fill the world with fools."
Herbert Spencer 1820-1903

No pilot should use automation as a backup for being stupid. The automation is there to help you make your job easier so you can pay attention to other things while it is flying your aircraft. When you blame the automation for not taking care of your negligence you put way too much responsibility on something that you were supposed to be doing and didn't. Just fly the f-----g airplane.

admittedly, it's been a while since i flew a boeing in an airline op.....

however, i remember quite clearly that:

1. you can always go around.

2. if there's any doubt, there's little harm in spooling up the engines just a tad -- maybe not on a fully coupled approach to mins or a cat iii in 0-0 [obviously] -- but in this case, the wx wasn't all that bad -- [not bad to me = > 500' & 1/2 mi or so] -- the poldervan is about 2x what the 73-8 needs to roll out and turn off sans snow / ice.

and if that little spoolup makes u too long -- refer to # 1.

and yes -- if i really needed to be at 0 thrust, i still had my hand resting there, ready....[not gonna miss pubbing at the brown bars in ams cuz my hand is elsewhere than the throttle quadrant -- :) ]

3. as an instructor i set my priorities of safety over instructing -- hard to instruct dead students -- plus my unblemished ticket is more important than showing you what you should have already learned in the sim...

4. we only used autopilot / a/t etc when really necessary -- much preferred clicking it off once established and hand flying -- why? -- we knew at least once and usually twice / yr we'd be going to that torture chamber they called the sim -- and how likely was it that they'd welcome us there and say -- ok -- we want to see you pass this check using the autopilot?

nope -- the a/l i graduated said: in the sim you're gonna fly the whole time with an engine or 2 out, maybe one on fire, and a hyd or elec failure [at least] -- we know that if you can do that, you'll be fine flying the line with most everything working.

plus -- all flight checks were given somewheres between 2 and 5 am -- often with the faa watching....

frankly, i was glad for it -- till i read bout this accident [and the colgan, right here in my home town], i had no idea it's not that way everywhere -- so there is a difference between airlines!

5. #1 a/t inop? no worry! there's always the #2 a/t -- it's called the f/o [or s/o, if anyone knows what those were...]

cheers!

dt

WWW.AIRLINE-CRASH-ANALYSIS.COM

I don't want to take sides in this argument (others are doing a good job of illustrating the issues).

But one thing to keep in mind is best illustrated in the CVR when it becomes available. It's a tough thing to hear a CVR where the machine is about to punish the human for making a mistake. After hearing one you want to go out and beat the aircraft with a hammer.

lomapaseo;
Yes I hear exactly what you are saying.

But when we beat the aircraft, we beat ourselves; - the metal doesn't care and will do it again just as soon as it can.

Just a thought
 

From the moment the accident aircraft descended through 1000' AGL on the EHAM RWY18R G/S the crew were guilty of breaching widely established stabilized approach criteria that should have required them to perform a go around. Which successfully executed would have prevented the accident. Regrettably, the crew do not appear to have been aware of the breach. Perhaps because they were doing such a poor job of monitoring the automation, which also unbeknown to them had laid a trap involving the malfunction of a relatively minor piece of equipment that would come to bite them not once but twice, just a few seconds later and further down the G/S.

Unfortunately for those who perished alongside the flight crew, under the current state of the art in commercial aviation, it is the pilots who are left to determine if and when stabilized approach criteria have been breached and the decision to go round or continue the approach similarly is left to the discretion of pilots.

But when accident report after accident report would seem to regularly fault crews for not monitoring the automation as required, then perhaps its time to recognize that humans including highly trained professionals such as airline pilots, may be badly suited to monitoring activities and resolve to take the decision to go around when stabilized approach criteria have been breached out of the hands of crews and give it to the automation. This could be achieved by yet another suite of aural warnings; Unstable UNSTABLE UNSTABLE! - GO AROUND! GO AROUND! Or less subtly, the automation could be given the authority to invoke TOGA and place the aircraft on a pre-programmed missed approach procedure without human intervention.

Such a system would inevitably require RA input in order to function correctly and build in safeguards to prevent nuisance alerts, but I suggest we not put it on just the one RA source without some form of comparator.

Comments? suggestions?

Only partly TIC.

Pilots are very good at fixing bad approaches. Many accidents have been prevented because of their alertness and ability to prevent crashes. Sometimes things go wrong so this happened. Hopefully this incident will not be repeated.

Norwegian,
You are quite right about the MEL.Unfortunately though,the RA problem wasnt reported so the dispatch of the aircraft was never subject to the MEL.Did other crews verbally warn this crew of the problem?If so,then the MEL should have been consulted.The MEL issue will figure bigtime in the report.

MSAW;not applicable here for obvious reasons as already pointed out.

You might want to rethink that one.It sounds crazy to me.But maybe the Airbus people will take you up on it.

bubbers44 - hope isn't a plan.

Rananim: Sounds crazy to me too. But if problems with the man/machine interface continue to crop up in accident reports of this sort with some regularity, then at some point something will have to be done. At least with my proposed solution we would merely be ensuring actions that should be taken perhaps more often, actually get done when required by handing responsibility to the automation. As a bonus, I think the threat of the automation breaking them out of an approach if they exceed parameters will likely tend to improve crew monitoring of the automation, resulting in more stable approaches generally and fewer incidents or accidents of this type.

Not a very logical point. You can apply it to cars, motor bikes, horse and carts, ships, billy-carts and virtually any form of transport including space shuttle.

Regrettably, the problem may be cultural. The Capt. was possibly a smug ex-Turkish Air Force pilot, and the F/O and jumpseat were suitably subservient, to the point of watching death approach, but too scared to object (fearing pointing out dangers would cast doubt on the Captain's Command Authority). The Turkish culture is rich, but, aviationally speaking, much to deferential to authority. Maybe I am pre-mature, but I see no other explanation now. This is an inherent weakness in Turkish culture.

On the contrary, rather than say the crew were probably unaware of breaking the rules, I would hazard a guess that in this case and the majority of accidents similar to what we have read here, crews were well aware they were pressing on, regardless of breaking established safety norms. Often it is ethnic culture and/or just plain cowboy flying which they have successfully got away with on previous occasions.

MU300,

You appear to want to fix the problem of over reliance on automation-with more automation!

No automatic system will ever be perfect, and piling layer after layer of automation into the system simply means we will find a new flaw, that will cause a new type of accident.

After Ceritous we got TCAS. Great kit, but it didn't take long for it to contribute to an accident (DHL/Tupolov).


Because of CFIT we got GPWS. GREAT kit- but, IMHO has led to a down-grade in terrain awareness.

The list goes on.

SOPs and training which ensure the flight path of the aircraft is identical whether or not automatics are involved would lead to a higher situational awareness generally.

Regular hand flying of approaches and a strict culture of down-grading or dis-connecting automatics as an immediate response to anomalous behavior should be an industry standard.

Many times both in the sim and in the aircraft, I've seen the blank-eyed confusion when the automatics do something unexpected. The response is often, in a critical phase of flight, to try and solve the anomaly and get the box doing what it should, rather than switch the damn thing off, try and salvage the situation within SOP stabilization criteria, or go around and sort things out at a safe height.

I fly 777s and it is SOP that the AT is on for all phases of flight. I'm quite convinced this has degraded skills at monitoring airspeed on approach, and left a "It will look after me" mindset.

Regular approaches (in appropriate conditions) with AT off would raise the standard of monitoring when it's on.