TAM A320 crash at Congonhas, Brazil
I support PPRuNe
Join Date: Jul 2007
Location: Belo Horizonte, Brazil
Posts: 162
Likes: 0
Received 0 Likes
on
0 Posts
Airbus Telegram/Document
The newspaper Folha de São Paulo published that the Airbus Industry issued a Telegram/Document to all using their products stating that the black boxes in the TAM accident revealed that there was nothing wrong with the airplane. Previous reports from the pilots who flew the plane in the last few days said that there was a problem in the left engine and that when descending from 260 the pilot felt a strong vibration in the engine (s?) on neutral that went away when he accelerated the plane. TAM also said that in the last 90 days there was nothing wrong with the plane except the right reverser being locked up.
PPRuNe supporter
Join Date: Dec 2003
Location: Planet Earth
Posts: 1,677
Likes: 0
Received 0 Likes
on
0 Posts
So how does one manually extend the GS on an A320? The same lever as used to arm them? by Mad (flt) Scientist
TripleBravo;
The DFDR lat/long parameters in and of themselves are not sufficiently accurate to determine the touchdown point. Lat/Long are normally supplied by the IRS's and as you know, position is determined by DME/DME etc and a bias is supplied to the IRS's, (the IRS's actually never are "updated" themselves).
The only way that a touchdown point can be determined with any accuracy is if the aircraft had GPS and the DFDR dataframe was sufficiently sophisticated to capture these parameters. Even thought the Airbus 320 DFDR is a good one, not all DFDRs capture huge amounts of data. FOQA/FDA recorders typically capture up to four times the amount of data that the DFDR does and often at higher sample rates. Even if this aircraft were fitted with such a second recording system however, it would not have survived the crash as the FDIMU/FDAU (whatever is installed) is in the EE compartment.
The parameters which are typically used to determine the touchdown point are VACC, (vertical acceleration), main gear oleo compression, IVSI all of which are combined in an analysis by flight data software to determine the point of first touchdown. You are correct in observing that this determination is only as accurate as the sample rate, typically once-per-second and that critical information can be missing in such a slow rate. VACC is typically sampled at eight times per second as are the oleo compression parameters but this is less the case with many DFDRs. I suspect this DFDR samples these parameters at these higher rates however. Both interpretation and interpolation are crucial aspects of this kind of analysis and so the process is not as "direct" or straightforward as it may first seem.
The other way touchdown point might (emphasize might) be determined is to animate the approach, path on the runway and excursion to the left. The wind vector is known and other basic parameters can be sufficient to provide a very good although elementary animation even with a slight glidepath duck-under. The process may or may not yield accurate information depending upon a number of factors but it can be done. I think the airport video is probably too blurred to be of accurate use.
PJ2
Quote:
I'm not sure anyone will ever know. I'm not aware of any data on the FDR that will tell.
That should be an easy one, if you have the DFDR data (and the coordinates of the runway). LON / LAT is recorded, the air / ground state as well as radio alt, so that you can determine the position where the aircraft settled within GPS precision (limited by the time step interval of the recording and the precision of the aircraft's transducers).
I'm not sure anyone will ever know. I'm not aware of any data on the FDR that will tell.
That should be an easy one, if you have the DFDR data (and the coordinates of the runway). LON / LAT is recorded, the air / ground state as well as radio alt, so that you can determine the position where the aircraft settled within GPS precision (limited by the time step interval of the recording and the precision of the aircraft's transducers).
The only way that a touchdown point can be determined with any accuracy is if the aircraft had GPS and the DFDR dataframe was sufficiently sophisticated to capture these parameters. Even thought the Airbus 320 DFDR is a good one, not all DFDRs capture huge amounts of data. FOQA/FDA recorders typically capture up to four times the amount of data that the DFDR does and often at higher sample rates. Even if this aircraft were fitted with such a second recording system however, it would not have survived the crash as the FDIMU/FDAU (whatever is installed) is in the EE compartment.
The parameters which are typically used to determine the touchdown point are VACC, (vertical acceleration), main gear oleo compression, IVSI all of which are combined in an analysis by flight data software to determine the point of first touchdown. You are correct in observing that this determination is only as accurate as the sample rate, typically once-per-second and that critical information can be missing in such a slow rate. VACC is typically sampled at eight times per second as are the oleo compression parameters but this is less the case with many DFDRs. I suspect this DFDR samples these parameters at these higher rates however. Both interpretation and interpolation are crucial aspects of this kind of analysis and so the process is not as "direct" or straightforward as it may first seem.
The other way touchdown point might (emphasize might) be determined is to animate the approach, path on the runway and excursion to the left. The wind vector is known and other basic parameters can be sufficient to provide a very good although elementary animation even with a slight glidepath duck-under. The process may or may not yield accurate information depending upon a number of factors but it can be done. I think the airport video is probably too blurred to be of accurate use.
PJ2
Join Date: Jul 2007
Location: Germany
Posts: 556
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by bomarc
on the airbus:
can you reverse with nosewheel off the ground?
can you deploy ground spoilers with nosewheel off the ground?
can you reverse with nosewheel off the ground?
can you deploy ground spoilers with nosewheel off the ground?
Originally Posted by hetfield
@bomarc
Yes you can, but it's not recommended.
Yes you can, but it's not recommended.
FCOM 3.03.22: SOP
LANDING
Select MAX REV immediately after main landing gear touches down
Wheel braking is another matter. To quote said FCOM again:
Braking may be commenced before nosewheel is down, if required for performance reasons; but when comfort is the priority, it should be delayed until the nosewheel has touched down.
However it should be noted, that the PIC (pilot in command) always has full authority to do what he deems necessary for the safety of the passengers, crew, people on the ground and aircraft.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Surely we are wasting our electronic breath here discussing touch-down point? I cannot see how a difference of several 100m in the location would have made much difference in view of what we now think happened. I feel we should allow the crew credit in that they would probably have got that bit right.
Please - also note academic certification details are not really relevant to a practical crew working from published LDRs in an on board book which, as I have said before, on the 737NG ASSUME both reversers used at no2 detent on a slippery runway. Whether AB does I know not
Please - also note academic certification details are not really relevant to a practical crew working from published LDRs in an on board book which, as I have said before, on the 737NG ASSUME both reversers used at no2 detent on a slippery runway. Whether AB does I know not
alf5071h,
I feel you've distilled most of what's important in your recent post. This accident is a real wake-up call for those who do or might have to land on 'limiting' runways... which is all of us.
When you're "close to the edge" in performance terms, virtually everything you need to do becomes a "single point of failure" scenario, with limited or no chance of recovery due to time and workload constraints. In the case under discussion (wet, slippery, MEL items, CRM issues), everything has to go right for a successful outcome: landing not in TDZ = G/A or crash; RWY state worse than advertised = G/A or crash; groundspeed too high = G/A or crash; spoilers not deployed = G/A or crash; autobrakes not engaged = G/A or crash; engine(s) not at idle = G/A or crash, etc. By the time a critical failure has been recognised, you'll probably be outside the normal operating envelope and into the "hope this works and we get away with it" area. How many landings around the globe, every month, fall into the "got away with it" category?
It doesn't take much to turn an assured landing in an assured accident, when you're operating with low margins. Maybe more emphasis needs to be put on "rejected landings" in these sort of cases? After all, if you're on an instrument let down with a DH of 500' and the cloud is BKN @ 400', OVC @ 500', you shouldn't be suprised if it ends up as a missed approach. In my last sim detail we practiced rejected landings as part of the lesson plan and very useful it was, IMHO.
Statistically, there are variations in all the parameters involved with landing an aircraft and bringing it to a halt and these create a probability distribution for the actual stopping distance. On a non-limiting, dry runway you can be well up in the percentiles yet be perfectly safe. Also, there is plenty of time to recognise a gross error, such as lack of braking or an extremely long float. For cases like the one under discussion, small changes in external factors (let alone internal ones) can put you off the end before you know it; it can be a subtle combination too: wind not as advertised, runway braking coefficient lower, slightly prolonged flare, delay in reverser operation... it all adds up to a potential disaster but as you were expecting things to be a bit more marginal than normal operations, it becomes difficult/impossible to spot something that takes you "over the edge".
I'm reminded of the overrun accident @ Chicago and many others. It doesn't have to be a short runway - you can go off the end of 3,000m in the wrong conditions - but there is so little time to recognise & correct the situation when tarmac is at a premium...
I feel you've distilled most of what's important in your recent post. This accident is a real wake-up call for those who do or might have to land on 'limiting' runways... which is all of us.
When you're "close to the edge" in performance terms, virtually everything you need to do becomes a "single point of failure" scenario, with limited or no chance of recovery due to time and workload constraints. In the case under discussion (wet, slippery, MEL items, CRM issues), everything has to go right for a successful outcome: landing not in TDZ = G/A or crash; RWY state worse than advertised = G/A or crash; groundspeed too high = G/A or crash; spoilers not deployed = G/A or crash; autobrakes not engaged = G/A or crash; engine(s) not at idle = G/A or crash, etc. By the time a critical failure has been recognised, you'll probably be outside the normal operating envelope and into the "hope this works and we get away with it" area. How many landings around the globe, every month, fall into the "got away with it" category?
It doesn't take much to turn an assured landing in an assured accident, when you're operating with low margins. Maybe more emphasis needs to be put on "rejected landings" in these sort of cases? After all, if you're on an instrument let down with a DH of 500' and the cloud is BKN @ 400', OVC @ 500', you shouldn't be suprised if it ends up as a missed approach. In my last sim detail we practiced rejected landings as part of the lesson plan and very useful it was, IMHO.
Statistically, there are variations in all the parameters involved with landing an aircraft and bringing it to a halt and these create a probability distribution for the actual stopping distance. On a non-limiting, dry runway you can be well up in the percentiles yet be perfectly safe. Also, there is plenty of time to recognise a gross error, such as lack of braking or an extremely long float. For cases like the one under discussion, small changes in external factors (let alone internal ones) can put you off the end before you know it; it can be a subtle combination too: wind not as advertised, runway braking coefficient lower, slightly prolonged flare, delay in reverser operation... it all adds up to a potential disaster but as you were expecting things to be a bit more marginal than normal operations, it becomes difficult/impossible to spot something that takes you "over the edge".
I'm reminded of the overrun accident @ Chicago and many others. It doesn't have to be a short runway - you can go off the end of 3,000m in the wrong conditions - but there is so little time to recognise & correct the situation when tarmac is at a premium...
Returning for a moment to the human side of things & the question "why" the #2 T/L wasn't retarded, could it be that the Hot-2 comment "Reverse number one only" made in the critical couple of seconds after touchdown & the "Retard" announcements was an instruction (albeit incorrect) to the handling pilot which had the effect of distracting &/or deterring the poor chap away from a twin T/L action and focusing all action on the #1 T/L?
http://www.flightglobal.com/articles...-recorder.html
http://www.flightglobal.com/articles...-recorder.html
Last edited by Max Tow; 4th Aug 2007 at 11:28. Reason: to incl cvr
The Reverend
Join Date: Oct 1999
Location: Sydney,NSW,Australia
Posts: 2,020
Likes: 0
Received 0 Likes
on
0 Posts
Unfortunately, if they had elected to do a missed approach, the crew would have been put through an inquisition as other aircraft landed successfully (more or less) previously and with the very same aircraft the preceding day. Not easy to have a win. As history has proved, that runway at Conhongas was not the safest when wet and ungrooved. At least two other 737s could have shared the fate of Tam 3054. TAM has decided not to despatch any flights to Conhongas in runway wet conditions.
Join Date: Sep 1999
Posts: 541
Likes: 0
Received 0 Likes
on
0 Posts
I am trying hard to understand this:
If you leave one of the TL's in CL on touchdown you get:
i)AT disarms when you select the other TL to rev idle
ii)a repetitive RETARD(this is the only warning that you have forgotten to retard the other TL-no stopgap interlock to prevent you having fwd &rev thrust simultaneously??))
iii)no moving spoiler handle-only by referring to the screen do you get confirmation of spoiler actuation(tactile feedback problem..in a conventional a/c the spoiler handle physically moves-you feel it,you hear it,you see it with peripheral vision)
iv)the forgotten TL stays in the position you left it when AT disarmed and the FADEC commands forward thrust at that setting(even though the other eng is now commanding rev thrust???)..huge design problem here..
v)when do you get gnd spoilers?LEFT squat switch and both TL's in IDLE???
vi)does autobrake look for the same logic conditions as gnd spoilers?
On a conventional a/c,the logic conditions are:
a)flight spoilers on wheel spin up
b)gnd spoilers on squat switch
c)rev thrust below 10'RA BUT BOTH TL's at idle(this is the key interlock..you cant have one eng producing forward thrust and one reverse thrust)
If a pilot in his haste to stop on a short wet runway forgets to retard a TL but the aircraft allows him to engage the retarded TL in rev and denies him spoilers/autobrake,is this pilot or design error?Why let the pilot get stuck between a rock and a hard place?If hes forgotten to retard,deny him everything,that way he wont get swing(how can you brake manually with asymmetric thrust-you'll need rudder),he wont get retardation and he must abort the landing.
If you leave one of the TL's in CL on touchdown you get:
i)AT disarms when you select the other TL to rev idle
ii)a repetitive RETARD(this is the only warning that you have forgotten to retard the other TL-no stopgap interlock to prevent you having fwd &rev thrust simultaneously??))
iii)no moving spoiler handle-only by referring to the screen do you get confirmation of spoiler actuation(tactile feedback problem..in a conventional a/c the spoiler handle physically moves-you feel it,you hear it,you see it with peripheral vision)
iv)the forgotten TL stays in the position you left it when AT disarmed and the FADEC commands forward thrust at that setting(even though the other eng is now commanding rev thrust???)..huge design problem here..
v)when do you get gnd spoilers?LEFT squat switch and both TL's in IDLE???
vi)does autobrake look for the same logic conditions as gnd spoilers?
On a conventional a/c,the logic conditions are:
a)flight spoilers on wheel spin up
b)gnd spoilers on squat switch
c)rev thrust below 10'RA BUT BOTH TL's at idle(this is the key interlock..you cant have one eng producing forward thrust and one reverse thrust)
If a pilot in his haste to stop on a short wet runway forgets to retard a TL but the aircraft allows him to engage the retarded TL in rev and denies him spoilers/autobrake,is this pilot or design error?Why let the pilot get stuck between a rock and a hard place?If hes forgotten to retard,deny him everything,that way he wont get swing(how can you brake manually with asymmetric thrust-you'll need rudder),he wont get retardation and he must abort the landing.
Hot Dog
Who has ever mentioned a "missed approach"? If you mean a hypothetical consideration in those last few seconds of taking off again, I very much doubt whether the previous day's landing performance, or your unsubstantiated suggestion of an "inquistion", would have been part of the thought process. In any event, there's no evidence on the CVR of anything other than a continuing attempt to stop.
Who has ever mentioned a "missed approach"? If you mean a hypothetical consideration in those last few seconds of taking off again, I very much doubt whether the previous day's landing performance, or your unsubstantiated suggestion of an "inquistion", would have been part of the thought process. In any event, there's no evidence on the CVR of anything other than a continuing attempt to stop.
Allright everyone, take a look at this:
(taken from the flightglobal website. My bolding)
HOT-1 okay.
18:48:21.0
FWC twenty.
18:48:21.6
FWC retard.
18:48:23.0
FWC retard.
18:48:24.5
CAM [sound of thrust lever movement]
18:48:24.9
CAM [sound of increasing engine noise]
18:48:25.5
GPWS retard
18:48:26.3
CAM [sound similar to touchdown]
18:48:26.7
HOT-2 reverse number one only.
18:48:29.5
HOT-2 spoilers nothing.
18:48:30.8
HOT-1 aaiii. [sigh]
The FWC says RETARD twice, then there is an INCREASE in engine noise??? After which the GPWS says RETARD (obviously)
So how am I supposed to read this? The computers want the thrust levers to be retarded, but instead there is can INCREASE in engine noise!?! The significant thing here is that all this happens when the aircraft is still airborne. Touchdown happens at 18:48:26.3
There is an increase in engine noise 1.5 seconds prior touchdown
Therefore the increase in engine noise can not be attributed to deployment of the operative thrust reverser.
Therefore the increase in engine noise is due to forward thrust, go-around power setting.....
(taken from the flightglobal website. My bolding)
HOT-1 okay.
18:48:21.0
FWC twenty.
18:48:21.6
FWC retard.
18:48:23.0
FWC retard.
18:48:24.5
CAM [sound of thrust lever movement]
18:48:24.9
CAM [sound of increasing engine noise]
18:48:25.5
GPWS retard
18:48:26.3
CAM [sound similar to touchdown]
18:48:26.7
HOT-2 reverse number one only.
18:48:29.5
HOT-2 spoilers nothing.
18:48:30.8
HOT-1 aaiii. [sigh]
The FWC says RETARD twice, then there is an INCREASE in engine noise??? After which the GPWS says RETARD (obviously)
So how am I supposed to read this? The computers want the thrust levers to be retarded, but instead there is can INCREASE in engine noise!?! The significant thing here is that all this happens when the aircraft is still airborne. Touchdown happens at 18:48:26.3
There is an increase in engine noise 1.5 seconds prior touchdown
Therefore the increase in engine noise can not be attributed to deployment of the operative thrust reverser.
Therefore the increase in engine noise is due to forward thrust, go-around power setting.....
Join Date: Jan 2001
Location: UK
Posts: 2,044
Likes: 0
Received 0 Likes
on
0 Posts
Therefore the increase in engine noise is due to forward thrust, go-around power setting.....
How on earth have you determined GA Thrust
Nobody has even begun to suggest they selected TOGA on the other TL 
Nigel on draft said:
Sorry mate, I stand corrected on that one. I meant to say that as there is clearly an increase in engine noise, the thrust is thus apparently increased from an approach setting to something higher. I am trying to understand Airbus autothrottle systems but my brain only works in boeing modes.
Therefore I used the wrong term here and should have typed CLB.
How on earth have you determined GA Thrust Nobody has even begun to suggest they selected TOGA on the other TL
Therefore I used the wrong term here and should have typed CLB.
Reverser No1
Max Tow:
, I thought may have been a command call. See previous post of mine. However, the retard call has gone 'out', which happens when a reverser is pulled. Also the next call is 'Spoilers' only a short time later. It would seem highly likely that this was an 'Indication' call.
It will be interesting to see exactly what the No2 engine EPR was reading.
In the TW incident Thrust was frozen at 1.08.
Incidently there was no 'Swing' as someone mentioned earlier. If you think about - if there was instant climb power on the right engine then the a/c would have vacated the runway alot sooner than it did. Even with full right rudder. - Brakes ineffective.
The aircraft appeared to be turned off the runway. 'Turn Turn'.
The reversing engine produces NO swing.
Jim
Reverse No1 only
It will be interesting to see exactly what the No2 engine EPR was reading.
In the TW incident Thrust was frozen at 1.08.
Incidently there was no 'Swing' as someone mentioned earlier. If you think about - if there was instant climb power on the right engine then the a/c would have vacated the runway alot sooner than it did. Even with full right rudder. - Brakes ineffective.
The aircraft appeared to be turned off the runway. 'Turn Turn'.
The reversing engine produces NO swing.
Jim
Last edited by James7; 4th Aug 2007 at 12:58.
James7:
I expect you're right though the gap between first & second "Retard" is 1.4 secs & between second & third seems 2.5 so I'm not sure when next one would have been due. No doubt FDR comparison will tell all. I wonder whether FDR will also show whether one or both TR applied by crew on previous POA landing. Haven't noticed anyone reporting when the TAM MEL changed - as someone suggested, would be handy if such op procedure changes actively notified (i.e. not just a paperwork update) & reasons given. I would assume that two experienced guys on this flight would have discussed TR lockout implications at least before first sector landing. Was the o/b MEL correct, one wonders?
I expect you're right though the gap between first & second "Retard" is 1.4 secs & between second & third seems 2.5 so I'm not sure when next one would have been due. No doubt FDR comparison will tell all. I wonder whether FDR will also show whether one or both TR applied by crew on previous POA landing. Haven't noticed anyone reporting when the TAM MEL changed - as someone suggested, would be handy if such op procedure changes actively notified (i.e. not just a paperwork update) & reasons given. I would assume that two experienced guys on this flight would have discussed TR lockout implications at least before first sector landing. Was the o/b MEL correct, one wonders?
