Turkish airliner crashes at Schiphol
Join Date: Dec 2005
Location: At home
Posts: 244
Likes: 0
Received 0 Likes
on
0 Posts
Cloud base & FD
Many posters have wondered what could possibly have distracted the crew so badly as to leave the speed decay unnoticed.
Here's one theory. The ADS-B data shows that the aircraft passed 700 ft i.e. the cloud base level, almost exactly at the same time when the speed had decayed to Vref and the throttles should (latest) have been advanced from idle, either courtesy of the A/T (which we know was inhibited by the faulty RA1) or manually.
Maybe we have a situation with 3 pairs of eyes looking out the windscreen to try to find the runway in the haze, due to "unexpected removal of the Flight Director Command Bars during approach" which is mentioned as one consequence of the RA fault? And nobody monitoring the speed tape or paying attention to the throttles at this critical time.
Comments? Sorry if this has been posted before.
Here's one theory. The ADS-B data shows that the aircraft passed 700 ft i.e. the cloud base level, almost exactly at the same time when the speed had decayed to Vref and the throttles should (latest) have been advanced from idle, either courtesy of the A/T (which we know was inhibited by the faulty RA1) or manually.
Maybe we have a situation with 3 pairs of eyes looking out the windscreen to try to find the runway in the haze, due to "unexpected removal of the Flight Director Command Bars during approach" which is mentioned as one consequence of the RA fault? And nobody monitoring the speed tape or paying attention to the throttles at this critical time.
Comments? Sorry if this has been posted before.
Join Date: Sep 2008
Location: MI
Posts: 570
Likes: 0
Received 0 Likes
on
0 Posts
Stall Recovery
There is a "fault" in the Stall Recovery Procedure as far as I'm concerned. especially with the 737-200. Not sure about the -800, but from reading here and elsewhere it seems to be the same. And that is applying FULL (or TO/GA) power at the recognition of a stall. I learned the hard way, that it is NOT necessary to use full throttle. More than 1/2, yes, but not 100%. The nose just pitches up too much further delaying the increase in speed needed to recover. Perhaps the procedure might be amended.
Join Date: Dec 1999
Location: Both sides of 40W
Posts: 219
Likes: 0
Received 0 Likes
on
0 Posts
Rainboe:
Sorry about that but, I have been away for a while!! Iīll try to read the 78 odd pages on this thread!! But, their seems to be a lot said about this A/T theory which in my opinion is one of the factorīs in this whole tragic event. Letīs hope that the investigation brings out the contributing facts to this accident.
Sorry about that but, I have been away for a while!! Iīll try to read the 78 odd pages on this thread!! But, their seems to be a lot said about this A/T theory which in my opinion is one of the factorīs in this whole tragic event. Letīs hope that the investigation brings out the contributing facts to this accident.
Join Date: Mar 2009
Location: oxford
Posts: 1
Likes: 0
Received 0 Likes
on
0 Posts
Jackharr
I was just about to ask the same question. Engines below centre of gravity and centre of drag, so by appling thrust forces the nose up (as has been stated before). Low speed, so control surfaces less effective in offsetting this pitching moment. Increased AoT giving increased drag, so speed can decrease despite the additional thrust thus accentuating the stall. Would it have been possible to have cashed in almost all of the 450 ft height for speed whilst gently increasing thrust? (The captain was an F4 pilot - thrust on the centre line - different scenario!).
Join Date: Mar 2009
Location: 495 ft
Age: 51
Posts: 1
Likes: 0
Received 0 Likes
on
0 Posts
RA intermittent fault
Maybe releted.
I've been flying the bird for about 7000hrs (4000 NG). I've never had a problem like this until last month. Than I had the same issue they experinced(It was a -700 ). Doing a Visual APP manually, I've noticed there is no automated alt callout at 500'. Felt strange. By the time I looked in the RA reading was normal. But then there was no 50' 40' 30' 20' 10' callout ( option on our planes) either and the left RA was showing -7'. On the way back to base everything was normal. Maintanance checked if there is any recorded fault but there were none. Flying the same ship later to other DEST no probs. Two weeks later same ship, same DEST and it happened again. Maintanance answer: No recorded fault, system works normally.
Tought might be interesting to see if any other NG drivers had this before.
I've been flying the bird for about 7000hrs (4000 NG). I've never had a problem like this until last month. Than I had the same issue they experinced(It was a -700 ). Doing a Visual APP manually, I've noticed there is no automated alt callout at 500'. Felt strange. By the time I looked in the RA reading was normal. But then there was no 50' 40' 30' 20' 10' callout ( option on our planes) either and the left RA was showing -7'. On the way back to base everything was normal. Maintanance checked if there is any recorded fault but there were none. Flying the same ship later to other DEST no probs. Two weeks later same ship, same DEST and it happened again. Maintanance answer: No recorded fault, system works normally.
Tought might be interesting to see if any other NG drivers had this before.
Join Date: Jun 2001
Location: UTC +8
Posts: 2,626
Likes: 0
Received 0 Likes
on
0 Posts
Finman . . . if you had read this thread you would have known that your "food for thought" quote from Boeing had been posted at least twice.
So, are you suggesting that a copilot who is "slow" with his instrument scan and had failed to monitor his airspeed and flown into the stick shaker should also "be expected to react correctly" to effect a timely recovery from an approaching stall . . . ? Not in my cockpit. Nor is the F/O ever a PIC where I fly. 
Propellerhead: . . ."The captain should NOT take over in the event of a stall normally (maybe different on a training flight if the trainee fails to react). The F/O is PIC/US and is fully trained and should be expected to react correctly to a stall warning."
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
My reading of the reports so far (the genuine ones) is that the RadAlt malfunctioned at or before 2000' when the GPWS warning was issued. I think by 500'; or so the a/c was well into decayed speed and near stall and had been without 'normal' power for around 100 seconds, although normally some of that time would be expected to be throttles closed anyway as it slows up with flap and gear extension. The situation was therefore masked. I can understand how this developed that far. Missing the 'RETARD caption is easy and 'green arrows' tell you nothing at that point. What I cannot understand is the lack of monitoring thereafter. Apart from PF being aware of what the a/c is doing, PNF's job is primarily to monitor. A TC as PNF has an even greater demand on him/her and I would expect to be even more attentive. Then there is the 'recovery' itself. Apparently wrong and muddled. I would expect strong calls of 'SPEED' from my PNF, rising in tempo and volume until I hear 'I HAVE CONTROL!'. I have lost the plot now on the height at which the stall warning occurred, but with the power available a recovery would certainly have been possible had it been flown correctly.
Hetfield - lawyers aside, the aviation doctrine revolves around 'Fail' as a finite situation. Fluctuating or wrongly reading do not necessarily constitute a 'failure' in terms of warnings or software interaction, and as someone else has pointed out, A/T and RadAlt #1 are NOT needed for a successful A/P B approach. Also a manual approach can be flown by the LHS with those 'failures' if desired.
Hetfield - lawyers aside, the aviation doctrine revolves around 'Fail' as a finite situation. Fluctuating or wrongly reading do not necessarily constitute a 'failure' in terms of warnings or software interaction, and as someone else has pointed out, A/T and RadAlt #1 are NOT needed for a successful A/P B approach. Also a manual approach can be flown by the LHS with those 'failures' if desired.
Last edited by BOAC; 5th Mar 2009 at 19:10.
Join Date: Oct 2007
Location: Location, Location
Posts: 743
Likes: 0
Received 0 Likes
on
0 Posts
QUOTE: top: 'thoughts of other NG drivers'...
Yes.
Flew the 737 for 5300 hours (about 3500 on NG) and had several anomalies with the rad alts (both left and right). Was not uncommon to be unable to reproduce the problem on the ground. Probably had an MEL'd inop Rad Alt about 5 times (more commonly the right RA funnily enough) on the NG as well.
No big issue really if you are fully aware of what you won't have available to you as a result.
Yes.
Flew the 737 for 5300 hours (about 3500 on NG) and had several anomalies with the rad alts (both left and right). Was not uncommon to be unable to reproduce the problem on the ground. Probably had an MEL'd inop Rad Alt about 5 times (more commonly the right RA funnily enough) on the NG as well.
No big issue really if you are fully aware of what you won't have available to you as a result.
Join Date: Aug 2006
Location: Nairn, Highland
Age: 85
Posts: 159
Likes: 0
Received 0 Likes
on
0 Posts
Am I correct in understanding that with B737, when full power is applied, the elevator lacks the authority to counter the pitch up trim change? As an ex C130, Bae146 pilot, etc, that strikes me as an amazing design fault.
Join Date: Sep 2005
Location: here and everywhere
Age: 50
Posts: 56
Likes: 0
Received 0 Likes
on
0 Posts
RA problem
At my company we fly the 800 and 4 of our birds are same age as the THY
(2002).
We have been straggling with RA problems the last 5 years, each time with a different A/C.
RA showing -7, GPWS lng gr warning, NO AUTO ALTITUDE CALLOUT, Roll bars on FD of the affected RA dissappearing on VORLOC capture and pitch bars dissapperaring on GS capture.
Finally after landing most of the times we got PSEU light that went off once engines were shut down and PRG BRK was set!
Fortunately no uncommanded RETARD by AT ever occured!
Lastly my 2p regarding this issue.
At my company we used to operate Airbus were any FMA change had to be announced by the PF! This procedure was kept when the Boeing was introduced!
I find that this procedure can close gabs and reduce cock ups!
(2002).
We have been straggling with RA problems the last 5 years, each time with a different A/C.
RA showing -7, GPWS lng gr warning, NO AUTO ALTITUDE CALLOUT, Roll bars on FD of the affected RA dissappearing on VORLOC capture and pitch bars dissapperaring on GS capture.
Finally after landing most of the times we got PSEU light that went off once engines were shut down and PRG BRK was set!
Fortunately no uncommanded RETARD by AT ever occured!
Lastly my 2p regarding this issue.
At my company we used to operate Airbus were any FMA change had to be announced by the PF! This procedure was kept when the Boeing was introduced!
I find that this procedure can close gabs and reduce cock ups!
Join Date: Jan 2008
Location: US
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
GlueBall
So, are you suggesting that a copilot who is "slow" with his instrument scan and had failed to monitor his airspeed and flown into the stick shaker should also "be expected to react correctly" to effect a timely recovery from an approaching stall . . . ?
Last edited by MU3001A; 5th Mar 2009 at 19:38.
Ka8 Flyer:
My last company (and several before) always had the 2500 ft "Rad Alt Live" followed by the exact height read out so that a comparison could be made).
To the Rest of You:
I have no experience of the the Boeing 737-800 but I am somewhat disturbed by some posters who consider hand-flying to be a CRM problem.
For example: "If you decide to hand-fly an approach, have you considered how much pressure you are putting the PNF under especially in a busy traffic environment?"
What a total load of bollox this statement is. If the PNF is unable to push buttons on the MCP and deal with the huge task of answering the radio then he should not be sitting in his seat and getting paid.
As a TRI/TRE I would usually start line-training with LVP procedures and then, once we had those squared away, I would insist on hand-flown approaches and landings until I was happy that the new addition to our company could cope with raw data, circling approaches and visual circuits.
Sadly, there are some major airlines out there right now who discourage hand-flying and to whom an NDB approach would constitute a major emergency.
We seriously need to revisit our training programmes.
My last company (and several before) always had the 2500 ft "Rad Alt Live" followed by the exact height read out so that a comparison could be made).
To the Rest of You:
I have no experience of the the Boeing 737-800 but I am somewhat disturbed by some posters who consider hand-flying to be a CRM problem.
For example: "If you decide to hand-fly an approach, have you considered how much pressure you are putting the PNF under especially in a busy traffic environment?"
What a total load of bollox this statement is. If the PNF is unable to push buttons on the MCP and deal with the huge task of answering the radio then he should not be sitting in his seat and getting paid.
As a TRI/TRE I would usually start line-training with LVP procedures and then, once we had those squared away, I would insist on hand-flown approaches and landings until I was happy that the new addition to our company could cope with raw data, circling approaches and visual circuits.
Sadly, there are some major airlines out there right now who discourage hand-flying and to whom an NDB approach would constitute a major emergency.
We seriously need to revisit our training programmes.
Join Date: Jan 2006
Location: US
Posts: 2,205
Likes: 0
Received 0 Likes
on
0 Posts
Sadly, there are some major airlines out there right now who discourage hand-flying and to whom an NDB approach would constitute a major emergency.
We seriously need to revisit our training programmes.
-----------------------------------------------------------------------
Nothing like have the final check ride include a non-precision approach to a circle to land. Do we ever do it on the line? No, but it highlights weaknesses in a hurry.
We seriously need to revisit our training programmes.
-----------------------------------------------------------------------
Nothing like have the final check ride include a non-precision approach to a circle to land. Do we ever do it on the line? No, but it highlights weaknesses in a hurry.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Correction to 'when it happened', RB, from the DSB's release - not at cloudbreak as you suggest, and hence, I feel the GPWS at 1950'?
"When descending through approximately 2000 feet the autothrottle, which uses the left radio altimeter data, transitioned to landing flare mode and retarded the throttles to the idle stop."
Interesting to hear about the single A/P a/land. I'm not sure, however, the elevator will cope with a flare from normal flap settings, since it will not autotrim, and you would have had very little requirement for a flare at F15.
"When descending through approximately 2000 feet the autothrottle, which uses the left radio altimeter data, transitioned to landing flare mode and retarded the throttles to the idle stop."
Interesting to hear about the single A/P a/land. I'm not sure, however, the elevator will cope with a flare from normal flap settings, since it will not autotrim, and you would have had very little requirement for a flare at F15.
Last edited by BOAC; 5th Mar 2009 at 20:04.
Join Date: Jan 2008
Location: US
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
When it happened
I think perhaps he was referring to airspeed decaying below V Ref ending in an imminent stall event, both of which which would have occurred below the cloud base.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
There are 2 applications of trim. I believe the second is for the flare. 
- not sure how this finished up BEFORE your post!!!
Do try and post back the results if you get the chance.
- not sure how this finished up BEFORE your post!!!Do try and post back the results if you get the chance.
Join Date: Apr 2002
Location: FUBAR
Posts: 3,348
Likes: 0
Received 0 Likes
on
0 Posts
BOAC,
Never tried it (but I will ask to now if time can be found) but I believe single
autopilot landing is no prob with norm landing flap of 30 or more appropriately 40.
The nose up trim applied during dual channel is intended I believe to give the elevators a bit less work to do for the auto go-around, which of course has to be fairly agressively instigated if flown from the Cat3 DH of 50ft.
Never tried it (but I will ask to now if time can be found) but I believe single
autopilot landing is no prob with norm landing flap of 30 or more appropriately 40.
The nose up trim applied during dual channel is intended I believe to give the elevators a bit less work to do for the auto go-around, which of course has to be fairly agressively instigated if flown from the Cat3 DH of 50ft.
Join Date: Sep 1998
Location: wherever
Age: 55
Posts: 1,616
Likes: 0
Received 0 Likes
on
0 Posts
I was a little surprised by the initial report. Not what it said but the way it said it. I seems to be written in a deliberately none technical manner. The AT mode being wrongly called flare and a lack of real data other than a couple of heights and speeds. Very different to for example a UK AAIB report.
I did note that the minimum speed mentioned matches the ADS-B data so if I look at the ADS-B trace and consider "landing speed" as quoted in the report to mean Vref then they were at Vref at about 900ft and 2.5D. That's about 30 seconds flying below Vref before the shaker fired. It seems a very long time on finals for 3 pilots not to notice. Even if it did coincide with cloud break into hazy visibility. 30 seconds is an age.
---------------added-----
Nose up trim for autoland is to ensure that if a failure causes to AP to fail the aircraft will not pitch down.
I did note that the minimum speed mentioned matches the ADS-B data so if I look at the ADS-B trace and consider "landing speed" as quoted in the report to mean Vref then they were at Vref at about 900ft and 2.5D. That's about 30 seconds flying below Vref before the shaker fired. It seems a very long time on finals for 3 pilots not to notice. Even if it did coincide with cloud break into hazy visibility. 30 seconds is an age.
---------------added-----
Nose up trim for autoland is to ensure that if a failure causes to AP to fail the aircraft will not pitch down.
Last edited by FE Hoppy; 5th Mar 2009 at 20:29. Reason: adding info about nose up trim for autoland.
Join Date: Aug 2005
Location: Cloudbase
Posts: 149
Likes: 0
Received 0 Likes
on
0 Posts
PJ2
thanks for your considerate response in 1503
Yes, that's all true for a single system. The 767, however, has three independent pitot systems, right? (two for the CPT/FO PFDs and one feeding the Backup). So while a sinlge pitot/static, if it fails without any backup, will create a major issue, having three of them will always render the failed one a minority.
The same is true for the two independent radalt systems. In that case it's at least rather trivial to determine that they are in disagreement and thus not trustworthy. If you then also look at altitude trends from the barometric altimeter, you'd again be able to rule out the faulty one.
The design of the current system does feed from a single data source, though and doesn't seem to make use of available redundancies. I'm sure there is a design decision behind this, and I guess my knowledge is insufficient to second guess the design. I'm just stating that there are options to have fallback systems.
True. There's always a point where you get to a system state that hasn't been anticipated in the design, that's the point where you have to throw it over the fence to you guys.
A nonexistent risk is, by definition, nonexistent...
thanks for your considerate response in 1503
I disagree. These two accidents bear only superficial similarities. I think it is beneficial to understand why. The main reason is, the similarity between the static/pitot sensors and a Radio Altimeter in terms of the nature of potential outcomes should one or the other fail, especially in the two accidents cited, is minimal.
An airplane with an unreliable Radio Altimeter can be flown without a problem - it is a non-event. I have seen it, flown it, seen such spikes/dropouts in flight data. I have also flown a B767 with serious pitot-static problems, (at night, over mountains) and know the potential for disorientation.
An airplane with an unreliable Radio Altimeter can be flown without a problem - it is a non-event. I have seen it, flown it, seen such spikes/dropouts in flight data. I have also flown a B767 with serious pitot-static problems, (at night, over mountains) and know the potential for disorientation.
The same is true for the two independent radalt systems. In that case it's at least rather trivial to determine that they are in disagreement and thus not trustworthy. If you then also look at altitude trends from the barometric altimeter, you'd again be able to rule out the faulty one.
The design of the current system does feed from a single data source, though and doesn't seem to make use of available redundancies. I'm sure there is a design decision behind this, and I guess my knowledge is insufficient to second guess the design. I'm just stating that there are options to have fallback systems.
There may be engineering aspects to this but this isn't fundamentally an engineering issue. It is an airmanship, piloting issue.
A nonexistent risk is, by definition, nonexistent...