Spanair 5022 Off Runway In Madid on T/O

[Gabriel]

Until someone tells us in what position the flap control(s) were found, we don't know if the crew failed to set the control(s) to the correct position or whether they set the control(s) properly but the flaps/slats didn't move to the selected position.
The flaps and slats indicator are linked to the flaps and slats themself, not to the lever, so the indication matches the actual flaps/slats position and not that of the lever.

At the very least, they didn't checked (effecivelly) the setting during the required checklist.

However, while not impossible, I think it's unlikely that the crew set the flaps lever but the flaps/slats failed to position on the selected seting. That would requiere yet one additional independent failure.

[andrasz]

However, while not impossible, I think it's unlikely that the crew set the flaps lever but the flaps/slats failed to position on the selected seting. That would requiere yet one additional independent failure.

I don't know about the flaps, but regarding slats, this is what I picked up elsewhere:

On the MD 80 series the slats are controlled by a single drum dual actuator mechanism mounted in the aft center wing box section.This unit is linked to each slat section with VERY thick control cables that pull the the slats either in or out, and they are all linked together unlike say a 737. That is why no separate slat section indicator is needed like on the 737.If one is going out ALL is going out, unless a catastriophic failure occured to the cotrolcable system. Should the spring-loaded cables from the lever to the actuator become severed (again very unlikely) that would immediately be noticed on the free movement of the lever.

Also as Gabriel pointed out earlier, in case the lever was properly set, there would have been a flaps/slats disagree warning, which is separate from the TCW.

[Crunk415balla]

Correct me if I am wrong, but may this not mean that the pilots forgot to set the flaps in the T/O config, but instead that they did lower the flaps, but the flaps didn't extend, seeing as warnings failed anyway?

[andrasz]

Correct me if I am wrong, but may this not mean that the pilots forgot to set the flaps in the T/O config, but instead that they did lower the flaps, but the flaps didn't extend, seeing as warnings failed anyway?

This is a VERY remote possibility, requiring independent failure of slat/flap actuator mechanism AND slats/flaps disagree warning (note: in this case there would be no take-off configuration warning, because that is fed from the position of the slat/flap lever). A much more likely scenario is that the crew did not set the flaps, and the TOCW failed (requires just one failure).

[Smilin_Ed]

However, while not impossible, I think it's unlikely that the crew set the flaps lever but the flaps/slats failed to position on the selected seting. That would requiere yet one additional independent failure.


Gabriel, I agree that this kind of failure is very remote, but I learned long ago in accident investigation that you need to consider everything. I would like to know what position the flap/slat control levers were found in.

[Dmmoore]

Gabriel, I agree that this kind of failure is very remote, but I learned long ago in accident investigation that you need to consider everything. I would like to know what position the flap/slat control levers were found in.

Ed,
We may not be told where the levers were actually positioned until the final report is published.

We can infer the following:
The flaps / slats were not extended.
It the flaps were commanded to extend but did not, a flap / disagree warning (separate from the indication and takeoff warning system) would have illuminated and the flight crew would have missed another check list item.

Again, we may not have the final answer until the final report is published. However, I see nothing to suggest an aircraft malfunction (other than the takeoff warning) or the regulatory agencies would be all over the problem.

[andrasz]

Security camera video now online:

http://www.elpais.com/articulo/espana/Ha/habido/accidente/cabecera/33/izquierda/creemos/aviacion/elpepuesp/20080918elpepunac_12/Tes

[scottkin]

Wow!!!!! From I can see in the video it looks like the plane just belly flops back to the runway and slides into the dirt until it explodes.

[Carl E. Hackert]

Judging by the plume of dirt, the newly released security video seems to begin with A/C already into the turf adjacent to the runway. I looks like they never got very high. Its hard to imagine the crew attempting to land the A/C back onto the runway when the realized the A/C would not fly out of ground effect. The irresistable force of the stall more likely put them back onto terra firma while the crew and their ill-fated passengers were along for the horrible ride into aviation history. Its possible that the warning systems were accidentally disabled yet, according to the reports of last month recorded in this Forum, the crew in the takeoff checklist did say "OK" to the call for flaps and slats. This is not the standard language since settings were never annunciated.

So, it comes down to the FDR info to determine the flaps & slats settings... God rest their souls...

[scottkin]

Carl I think you may be onto something there. Watch the dust come into the picture at the start of the video. I think the trouble started before the plane came into the picture. Also check out the piece of debris that flies off the plane after the explosion that flies back toward the direction of takeoff.

[aardvark2zz]

.
The fuselage angle relative to the runway is 23 degrees.

The angle of attack shown in that frame would have been slightly larger than the pitch angle due to (probably) descending airplane.

(pic from pprune)

http://i420.photobucket.com/albums/pp286/aardvark2zz/spanair.jpg
.

[Dohc]

JK5022. Accident Aircraft Provisional Brief:

http://www.elmundo.es/documentos/2008/09/16/borrador_barajas.pdf

(obviously in Spanish)

[Gabriel]

.The fuselage angle relative to the runway is 23 degrees.

The angle of attack shown in that frame would have been slightly larger than the pitch angle due to (probably) descending airplane.
Did you take into account perspective (i.e. that the fuselage is not perpendicular to the camera)? That would make the "flat" angle measured on the picture larger than the actual pitch angle.

[Gabriel]

JK5022. Accident Aircraft Provisional Brief:

http://www.elmundo.es/documentos/2008/09/16/borrador_barajas.pdf

(obviously in Spanish)
This is not an officially released initial report. It's a draft of such report.

[Dohc]

I know, but as you can see the small and provisional brief contains more information than any valid hypothesis or theory that has been commented so far in this forum, especially when it comes to information extracted directly from the lost aircraft.

In the report, researchers spoke about the malfunction of the TOWS system. As you know, this TOWS failure in MD-82 was the reason of the similar accident at Detroit in 1987:

http://www.airdisaster.com/cgi-bin/view_details.cgi?date=08161987&reg=N312RC&airline=Northwest+Airlines

I hope that the full report definitively clarify the real causes of this sad incident.

[AVION1]

1) this camera is almost at the end of the runway
2) we don't see the whole event from this video, only the last moments of the crash
3) the aircraft crash landed on the grass (that's why we see dust coming out)
4) this video can not tell us what happened before....
:(

[Dmmoore]

"IF" the cause of this event "IS" failure of the flight crew to configure the aircraft properly, then the probable cause will be "Pilot Error" with contributing factors being TOWS and any factory that led to the crew error.

Again, "IF" the cause is other than the above, all bets are off.

[Dmmoore]

While the video is not conclusive evidence, the aircraft is airborne and most likely off the runway's left edge when the aircraft comes into view. A few frames later the aircraft is nose down and on the ground. I would have expected the aircraft to contact the ground nose up and if the main gear was supporting the aircraft to remain nose up for a few frames.

The rapid lowering of the nose suggests the main gear failed at touchdown or the drag from the main gear was higher than the elevator could overcome.

In any event, the video does not provide much into detecting the cause of the event.

[Theoddkiwi]

It does show that had the river valley not been there, eg the flat area had continued for a few hundred meters more they aircraft would have probably stayed intact with a much lower loss of life, though i am sure there would have been some pretty bad spinal injuries.

I am still in favour of a misconfigured aircraft being the primary cause and the TOW being accidentaly disabled during the maintenance action.

I suspect the crew did all their checks correctly the first time they attempted to take off, but did not on the second time and thus missed vital steps before they hit the throttles.

[scottkin]

I'm not sure if I missed something but has anyone heard anything about the CVR or the FDR yet?? That has got to of shed some light on this.

[Uncle Jay]

Looking at the video, the aircraft is airborne at the 1000-foot runway markings and on the runway centerline, yet the video shows (a frame later) a dust trail starting BEFORE the 1000-foot runway markings, 60 meters right of the centerline, and the airplane flat on the ground. No way. 2 different airplanes.

transcript:

-First ground marks: 60m away from runaway, measured perpendicular to the runaway's axis, and at the 3207,5 meter mark, measured from the direction of the runaway (total runaway length in Madrid, some 4400m and is also 60m wide itself).
-It rolled on the ground for 448m until reaching the limit (of that terrain), almost in a straight manner, at an angle about 16º with the runaway.
-Lost contact with the ground (terrain drops there) and recovers contact with the ground 150m away from the peripheral road that surrounds the runaway, which is 5,5m below in elevation from the first terrain touched by the plane.
-It continued on the ground on irregular terrain until it found the creek where the main remains are found, already quite disintegrated.

[mzk49o1]

Looking at the video, the aircraft is airborne at the 1000-foot runway markings and on the runway centerline, yet the video shows (a frame later) a dust trail starting BEFORE the 1000-foot runway markings, 60 meters right of the centerline, and the airplane flat on the ground. No way. 2 different airplanes.

Glad I'm not the only one who noticed this.

[ultraflight]

Link to the Spanish report. (http://www.fomento.es/NR/rdonlyres/97C78A32-1428-4BC4-B70C-18DA3474E92C/38073/2008_032_A_preliminar.pdf)


Tried to get an English version, but it seems not to have been posted yet.

[cja]

Link to the Spanish report. (http://www.fomento.es/NR/rdonlyres/97C78A32-1428-4BC4-B70C-18DA3474E92C/38073/2008_032_A_preliminar.pdf)


Tried to get an English version, but it seems not to have been posted yet.


Try here for an English Synopsis from Flight International (http://www.flightglobal.com/articles/2008/10/10/317285/first-formal-spanair-crash-data-sheds-little-light-on-flap-configuration-mystery.html)

[ultraflight]

http://www.elpais.com/recorte/20081005elpepinac_7/XLCO/Ies/20081005elpepinac_7.jpg

From El Pais website (http://www.elpais.com/articulo/espana/MD/rozo/tragedia/2007/fallos/similares/Barajas/elpepiesp/20080921elpepinac_1/Tes)..:

This picture was taken by the pilot of another plane. It shows an MD-83 with registration OE-LMM (Mapjet), on the 5th of June 2007, just before take-off from Lanzarote airport (Canary-islands). (The small pic is to show to the readers what flaps are).
No alarms regarding the incorrect configuration sounded in the cockpit, but the pilots succeeded in controlling the A/C. It was extremely instable after rotation, rolling violently and just avoided hitting buildings in the extension of the runway.

The article claims that perhaps the more powerful engines of the MD-83 saved the day, but there are a lot more parameters such as take-off weight to be taken into account in order to compare the two cases.

The incident is under investigation (by CIAIAC, Ministerio de Fomento)

[dr_cfii]

http://www.elpais.com/recorte/20081005elpepinac_7/XLCO/Ies/20081005elpepinac_7.jpg

From El Pais website (http://www.elpais.com/articulo/espana/MD/rozo/tragedia/2007/fallos/similares/Barajas/elpepiesp/20080921elpepinac_1/Tes)..:

This picture was taken by the pilot of another plane. It shows an MD-83 with registration OE-LMM (Mapjet), on the 5th of June 2007, just before take-off from Lanzarote airport (Canary-islands). (The small pic is to show to the readers what flaps are).
No alarms regarding the incorrect configuration sounded in the cockpit, but the pilots succeeded in controlling the A/C. It was extremely instable after rotation, rolling violently and just avoided hitting buildings in the extension of the runway.

The article claims that perhaps the more powerful engines of the MD-83 saved the day, but there are a lot more parameters such as take-off weight to be taken into account in order to compare the two cases.

The incident is under investigation (by CIAIAC, Ministerio de Fomento)

Interesting...instead of taking pictures, why didn't the other pilot jump on and alert them about the wront TO config???

[scottkin]

Interesting...instead of taking pictures, why didn't the other pilot jump on and alert them about the wront TO config???

Great question...

[ultraflight]

.
The fuselage angle relative to the runway is 23 degrees.

The angle of attack shown in that frame would have been slightly larger than the pitch angle due to (probably) descending airplane.

(pic from pprune)

http://i420.photobucket.com/albums/pp286/aardvark2zz/spanair.jpg
.

Maximum pitch 18 degrees, according to the provisional report.

[juan23]

....


<TABLE class=storycontent cellSpacing=0 cellPadding=0><TBODY><TR><TD colSpan=2>Spanair mechanics face crash quiz


</TD></TR><TR><TD class=storybody><!-- S BO --><!-- S IIMA --><TABLE cellSpacing=0 cellPadding=0 width=226 align=right border=0><TBODY><TR><TD>http://newsimg.bbc.co.uk/media/images/44946000/jpg/_44946501_44946052.jpg A preliminary report focused on the plane's wing flaps and warning system

</TD></TR></TBODY></TABLE><!-- E IIMA --><!-- S SF -->
The judge investigating the Madrid plane crash that killed 154 people in August is to question three mechanics on suspicion of manslaughter.
The two mechanics who checked the plane before take-off and Spanair's head of maintenance at Barajas will appear before the judge, court officials said.
The Spanish passenger jet crashed and burst into flames shortly after take-off at Madrid's Barajas airport.
An investigation said the plane's wing flaps had been set incorrectly. <!-- E SF -->
Human error
Investigators say the two mechanics had deactivated a faulty temperature gauge, but failed to spot a problem with the aircraft's take-off warning system, which was operating on the same electronic relay.
Less than half an hour later, Spanair flight 5022 crashed after take-off, killing all but 18 passengers on board.
Investigators say the wing flaps - which should have provided lift - had not been deployed, and that the warning system failed to sound in the cockpit.
The suspicion, set out in court documents, is that the mechanics failed to pick up on a broader electrical fault, which would prove fatal, says the BBC's Steven Kingstone in Madrid. <!-- Inline Embbeded Media --><!-- This is the embedded player component -->
<OBJECT id=bbc_emp_fmtj_embed_obj height=287 width=448 classid=clsid:D27CDB6E-AE6D-11cf-96B8-444553540000>
























<embed type="application/x-shockwave-flash" src="http://news.bbc.co.uk/player/emp/2_6_5222/player.swf" width="448" height="287" id="bbc_emp_fmtj_embed_emb" name="embeddedPlayer_7575006" wmode="default" allowFullScreen="true" name="embeddedPlayer_7575006" flashvars="config=http://news.bbc.co.uk/player/emp/config/default.xml?v10&companionSize=300x30&companionType=adi&preroll=http://ad.doubleclick.net/pfadx/bbccom.live.site.news/news_europe_content;sectn=news;ctype=content;news=europe;slot=companion;sz=512x288;tile=6&companionId=bbccom_companion_7575006&config_settings_autoPlay=false&config_settings_showPopoutButton=false&playlist=http%3A%2F%2Fnews.bbc.co.uk%2Fmedia%2Femp%2F7570000%2F7575000%2F7575006.xml&config_plugin_fmtjLiveStats_pageType=eav2&embedReferer=http://news.bbc.co.uk/2/hi/europe/default.stm&config_plugin_fmtjLiveStats_edition=International&embedPageUrl=/2/hi/europe/7673316.stm&"/></OBJECT>
<!-- companion banner -->Advertisement

<!-- END - companion banner --><!-- caption -->Mobile phone video of the immediate aftermath of the crash
<!-- END - caption -->
<!-- end of the embedded player component --><!-- END of Inline Embedded Media -->
The three men have formally been cited for the manslaughter of 154 people and for the injuries suffered by 18 survivors.
They are expected to give evidence in court next month.
Separately, the judge has set up a second investigation committee - consisting of pilots, flight engineers and mechanics - to run in parallel with an ongoing government-led inquiry.
The loss-making Spanish airline is owned by Scandinavia's SAS. Earlier this month, it announced it would shed more than 1,000 jobs and cut capacity by 25%. Spanair lost 515 million Swedish crowns ($71.72m; £39.69m) in the first half of 2008. <!-- E BO -->


</TD></TR></TBODY></TABLE>

[Dmmoore]

Assuming the cause of the crash was a misconfiguration of the aircraft, the primary cause must lay with the flight crew. They were the ones that failed to properly configure the aircraft.
A contributing factor would be the TOW system.
If the mechanics followed the approved procedure when deactivating the RAT indicator, they should not have any problem defending themselves.
If they took a shortcut, they better get a couple of good lawyers.

[Crunk415balla]

Interesting...instead of taking pictures, why didn't the other pilot jump on and alert them about the wront TO config???
Thats what I'm thinking, if he saw that why not say get on ATC and tell him to check his flaps settings.

[ultraflight]

Looking at an image of the mid-console on the flightdeck of the MD-82, the flaps lever is huge and sitting right next to the throttles.
It seems so unlikely that such an important lever being in such a wrong position, right next to your hand when you advance the throttles, goes unnoticed by both pilots.

PICTURE (http://images2.jetphotos.net/img/1/2/0/1/88935_1220884102.jpg)

[andrasz]

It seems so unlikely that such an important lever being in such a wrong position, right next to your hand when you advance the throttles, goes unnoticed by both pilots.


With a Flaps 1 or 5 setting, the lever position is very close to Flaps zero. You would need to actually look at which slot it is placed in, the difference is not readily perceptible by 'feel' until the 10-15 range.

[dr_cfii]

Since we're talking about the TO Config, and setting the flaps/slats...
WHEN is the TO checklist done???
Yesterday, I flew up to ORD and back to MIA on AA (757-200). On the way up, the slats/flaps were not set until we taxied into POSITION and hold!!!! I almost sh*t in my pants. The captain told the crew to prepare for TO, we're #2, then the other guy rolled, and as we were taxiing into position they extended the slats/flaps. Made me feel really uncomfortable! I asked the captain on the way out, he said that they kept changing the RWY on us, so they held the TO configuration till the last minute...

On the way back, the TO config was set as soon as we started taxiing, almost right after the engines started after pushback...

I flew on the 777 twice a month last year (to China) and recall they would always set the wings during early taxi, or even as soon as pushback is done, engines on, set the flaps/slats...

[Dmmoore]

Since we're talking about the TO Config, and setting the flaps/slats...
WHEN is the TO checklist done???
Yesterday, I flew up to ORD and back to MIA on AA (757-200). On the way up, the slats/flaps were not set until we taxied into POSITION and hold!!!! I almost sh*t in my pants. The captain told the crew to prepare for TO, we're #2, then the other guy rolled, and as we were taxiing into position they extended the slats/flaps. Made me feel really uncomfortable! I asked the captain on the way out, he said that they kept changing the RWY on us, so they held the TO configuration till the last minute...

On the way back, the TO config was set as soon as we started taxiing, almost right after the engines started after pushback...

I flew on the 777 twice a month last year (to China) and recall they would always set the wings during early taxi, or even as soon as pushback is done, engines on, set the flaps/slats...


90% of the time, the aircraft is configured at or shortly after the start of taxi. There are reasons for delaying flap deployment but not many. They are usually related to weather.

[dr_cfii]

90% of the time, the aircraft is configured at or shortly after the start of taxi. There are reasons for delaying flap deployment but not many. They are usually related to weather.


THanks Don, I guess this was unusually late...

[Spad13]

I asked the captain on the way out, he said that they kept changing the RWY on us, so they held the TO configuration till the last minute...



I call BS.

[AJ]

Boeing SOP calls for the flaps to be set in the after start procedure (before taxi). They are then checked in the Before Takeoff Checklist.

Airlines do modify Boeing SOP to their own procedures. At Qantas we are very closely aligned with Boeing procedure on the 737, 747, 744 and 767.

As Don mentioned the exception would be when taxying in slush or snow to avoid contamination.

[Gabriel]

Pretty complete interim report:

http://www.fomento.es/NR/rdonlyres/9AA08FE0-1151-41A0-9A57-9C5DF8BF6F8E/52772/2008_032_A_INTERINO_01_ENG.pdf (http://www.fomento.es/NR/rdonlyres/9AA08FE0-1151-41A0-9A57-9C5DF8BF6F8E/52772/2008_032_A_INTERINO_01_ENG.pdf)

Most theings we already knew:

- The crew did not check the TAWS because it was not the first flight of the day (now required before each flight).
- The TAWS was not working because one of the circuits that sense air/ground (nose gear strut compresion) was not working.
- That same failure made the RAT heat work on ground (it is supposed to work only when the airplane is in air mode)
- The crew taxied to the runway, but then returned to the ramp to check the RAT which was overheating.
- Without checking the reason for the filure, the maint. personel and the captain agreed to MEL the RAT heat, and opened its ciercuit braker.
- On the second taxi, the FLAPS item in the checklist was ommited. The checklist had been interrupted in the middle to contact ATC for clearance, and when they continued it they skipped the FLAPS item.
- They never set the flaps for take-off in this second attempt.
- The take-off roll was normal, they rotated at the propper speed (had the slats/flaps been set). The TAWS didn't alert abt the misconfig because of the reasons explained above.
- A few seconds after rotation the stick shaker activated, then the stall warning.
- With the airplane banking right and back to level a few times, the speed and altitude kept increasing slowly but the pitch attitude too until the plane fully stalled, went down, crashed and burned.

One thing I wuld have liked being included in the report, which wasn't and judging from what they list as ongoing investigation won't ever be, is whether the flight was recoverable after the stickshaker activation had the pilot used proper stall recovery techniques. But I know the answer to that anyway.

[ihatemondays]

Since we're talking about the TO Config, and setting the flaps/slats...
WHEN is the TO checklist done???
Yesterday, I flew up to ORD and back to MIA on AA (757-200). On the way up, the slats/flaps were not set until we taxied into POSITION and hold!!!! I almost sh*t in my pants. The captain told the crew to prepare for TO, we're #2, then the other guy rolled, and as we were taxiing into position they extended the slats/flaps. Made me feel really uncomfortable! I asked the captain on the way out, he said that they kept changing the RWY on us, so they held the TO configuration till the last minute...

On the way back, the TO config was set as soon as we started taxiing, almost right after the engines started after pushback...

I flew on the 777 twice a month last year (to China) and recall they would always set the wings during early taxi, or even as soon as pushback is done, engines on, set the flaps/slats...

It's maybe a little o/T, but DLH just retired their last A300-600's last month. On short hauls and with light loads, their t/o config sometimes was SLATS ONLY. Made you wonder what the heck...

But I agree, on most flights I had, the flaps were set early.

ToM

[Theoddkiwi]

I know on our my airline flap position is set after engine start during the before taxi checklist.

Would be interesting to see if they could have gotten up with out flaps when stick sharker had activated. really flatten off to rebuild speed and slowly climb out.

[Gabriel]

Would be interesting to see if they could have gotten up with out flaps when stick sharker had activated. really flatten off to rebuild speed and slowly climb out.
But it looks we'll never officially know. The preliminary report says nothing about that and it is not listed in the "to do" list for the ongiong investigation.

I've said "officially" because, extra-officially, I know they could have done it.

[mosteo]

Técnicos de mantenimiento denuncian a Boeing por el accidente del avión de Spanair

http://www.20minutos.es/noticia/548759/0/denuncia/spanair/accidente/

Rough and abridged translation (mine):

Technicians blame Boeing for Spanair plane accident.

Spanish association of airplane maintenance technicians (Asetma) made in the Magistrate's Court in Madrid a criminal complaint against Boeing (...).

Delgado (Asetma's president) explained that Boeing either didn't respect or failed to communicate to airlines the necessity of fulfilling some mandatory rulings, thus making it "responsible" or "an influencing party" in the accident. (...)

Delgado said that Boeing didn't respect obligatory normative stablished by the FAA.Sorry, I can't do better with specialized words. May be google trans is better.

[TeeVee]

not much technical stuff here. basically, they are trying to say the Boeing failed to provide a warning light to backup the alarm that is supposed to notify the crew of an improper takeoff configuration. i'm pretty sure that boeings have a chime or beep warning system to notify the crew if the aircraft is not configured for takeoff, i.e., flaps not set...

sounds to me like the airline is trying to drag boeing's deep pockets into the fray and further trying to deflect the absolute failure of the flight crew.

also, unlike here in the US, they have actually charged two maintenance guys with 154 counts of "reckless homicide" and 18 counts of "reckless injury"

[Gabriel]

i'm pretty sure that boeings have a chime or beep warning system to notify the crew if the aircraft is not configured for takeoff, i.e., flaps not set...
The fully Boeing planes (like the 737) have just an intermitent toot-toot-toot sound warning.

The ex-MD planes (now Boeing) have a sound of bells and the word of what is miscinfigured spoken in clear voice (Flaps!, Slats!, Spoilers! Trim! Parking brakes!). However, this system was not working during this accident. A back-up light would have not worked either because the probelm was that the plane "thought" that it was on the air because a switch attached to the nose landing gear strut failed, so infact the plane didn't thought it was taking off as to trigger a take-off warning to begin with. Also Spanair wasn't following Boeing's procedure of checking this warning system before each flight (the check is simpy to, before engines start, advance throttles with the plane intentionally misconfigured). They were checking it before the first flight each day (and this was not the first flight). That was the original procedure, but MD (now Boeing) changed it from "before first flight" to "before every flight" after an almost copy-paste accident that happened in Detroit several years ago. Lessons not learned in this case. The news article doesn't say anything about this, but maybe they are blaming Boeing for not sending them the revised procedure?

Additionally, the failure of the air/ground mode system is what caused a RAT probe overheat that made the crew return to the gate. The RAT probe heat is supposed to be automatically on with the plane flying and off with the plane on the ground, but because since the plane "thought" that it was in air mode it turned the RAT heat on, which with the lack of cooling airspeed overheated. But when they returned to the gate neither the crew nor the maint mechanic thought of this. Instead of finding the source they just pulled the circuit braker of the RAT heat to leave it out of service (it was not a required equipment for this flight in good weather).

This pretty much sumarizes the accident except the crew and plane performance during the actual take off. Was the plane able to complete the take-off and climb out safely even with the flaps and slats unintentionally left retracted? Yes. Absolutely. (you won't read this in the accident report, but still) Would that have required any exeptional skills from the pilot? No. Just to honor the stall warnings.

[TeeVee]

question: if the plane could have made the takeoff without flaps but with pilots acting properly, is this the same for the LAPA flight that crashed for similar reasons back in 1990 or so? that was a 737 with no flaps that never made it.

[Gabriel]

Very likely yes. Spanair is no doubt. The difference is that Spanair was taking off from a 14000ft runway, while LAPA had less than half of that.

I had the very rare opportunity to re-enact the LAPA take-off in a real full flight / full motion B-737-200 simulator in the same airport and was able to do it, pretty easily I would add, and I'm just a low time PPL that never flew something bigger than a 4 seats GA piston single weighting less than 2500lbs. Admittedly, I had the benefit of hindsight, but still.

Nobody doubts that an airplane is able to fly with the flaps and slats retracted. It happens all the time, in every flight.

The take-off profile is such that with an engine failing at V1 and the pilot rotating at Vr the plane will lift off with a speed that has a margin over the stall and will achieve V2 and 15ft above the opposite runway end, if the plane is taking off from the shortest runway that is legal for the wight, or the heaviest weight that is legal for the runway.

The plane stalls faster with the flaps and slats retracted. If you know you'll be taking off in that condition, that's no problem. You'll use the appropriate V1, Vr and V2 speeds for that condition and you'll be required to use a runway that is long enough for the airplane weight, or an airplane wight light enough for the available runway, so as to meet the performance requirements of the preceding paragraph.

The problem is when you think you have flaps and slats extended to a given setting, calculate V1, Vr and V2, and the airplane weight / runway length combination for that configuration, and then attempt to take off with the flaps / salts retracted. If the runway was relatively short and you are close to the maximum weight for that runway length / minimum runway length for your take-off weight, WITH FLAPS AND SLATS EXTENDED, there might be just not enough runway to achieve a speed compatible with flight in the available runway WITH FLAPS AND SLATS RETRACTED.

However, you can make good use of some margins. First, probably you won't have an engine failure so you'll be accelerating almost twice as fast after V1 as supposed for the take-off performance. Second, you will be accelerating a little faster from the beginning with the reduced drag configuration (this effect is really small because the take-off thrust is so much greater than the airplane's drag during the take-off run that a bit more or less of drag doesn't make a real difference). Third, you can get rid of the margin over stall, any speed that keeps the plane flying will work. Finally, you don't need to meet the 15ft above the runway end requirement. Even overrunning a bit before lifting off could be acceptable if that saves you.

The trick is simply "don't stall". Pull up as much as you need as long as you comply with "don't stall".

In the Detroit case (another MD-80, almost identical case than Spanair) the plane indeed managed to lift off and keep a positive but marginal rate of climb, which was not good enough to prevent hitting some light poles in a parking lot several hundred fetes from the runway. If fact, the pilot here made a pretty good and successful effort to fly at the limit of the stall but not beyond.

The MD-80 has two stall warnings: the stick shaker activates when the plane approaches the stall, and an aural stall warning (a horn and a yell "Stall!") at the limit of the stall. The stick shaker activated immediately during the rotation and the aural warning about a second later. The stick shaker remained active for the remaining of the flight, but the pilot kept lowering the nose a bit every time the aural warning triggered, and then pulling up a bit again until the aural warning triggered activated again, trying to fly exactly at the onset of the aural warning. In this way, the plane did not stall and fall from the sky, but it kept flying and climbing, albeit barely.

The problem here was not as much stall by itself, as roll stability and control. The ailerons become highly ineffective at high AoA. Also the adverse yaw (the tendency of the plane to yaw to the opposite direction of the aileron input) also is highly increased at high AoA. Most importantly, the roll damp (the tendency of the plane to achieve a constant roll rate for a given aileron input and to stop the roll when the ailerons are neutralized) almost disappears at the limit of the stall. This makes controlling the roll a very difficult stuff that requires constant and large aileron inputs and that usually result in pilot induced oscillations (the plane rolls to one side, the pilot applies opposite aileron, the aileron rolls back and the pilot centers the ailerons when the wings are back to level but the plane keeps rolling, by when the pilot realizes of that the plane has a significant ans increasing bank in the other direction so he applies full opposite ailerons and repeats the cycle). So the plane rolls to one side and the other. That is already bad enough because every time the wings are not level the lift vector tilts and you are "wasting" the part of the lift. Additionally, the frequent and large aileron deflections made by the pilot add drag. And aileron deflections past a threshold are accompanied by roll spoiler deflection, which in this case will be also large and frequent. Spoilers both increase drag and reduce lift.

This is what happened in the Detroit accident. The plane kept rocking from side to side with the pilot making full aileron (and hence roll spoilers) inputs. That spoiled the climb performance enough to prevent clearing the obstacles (the light poles).

What should the pilot have done? Exactly what he did, but at the onset of the stick shaker instead of the stall aural waring. That will have given him almost the same nominal lift but with a lot more of roll stability, hence less banks, less spoiler deflections, and hence less drag and more lift. The plane would have easily cleared the poles with room to spare. That was the conclusion by the NTSB.

There was one similar thing with the LAPA case, and it was that in the Detroit case the runway was short too (almost as short as legally possible for the airplane's weight).

In the LAPA case the crew applied the Murphy's law and did wrong everything that could have been done wrong. They forgot to set the flaps, they didn't check the flaps as required by the checklist, they didn't immediately abort the take-off when they had the take-off configuration warning, which happened as soon as they advanced the throttles for take-off. Instead, they kept trying to figure out what was the source of the warning. That's bad enough, but at least if you are approaching V1 and couldn't fix the problem, ABORT FOR GOD'S SAKE!!!. Nope. They kept going with the warning and all. They rotated at Vr, the plane didn't lift off at the usual pitch, the stick shaker activated when the plane reached a pitch of some some 15 degrees and had accelerated past V2, and THEN the pilot decided to abort, not before V1 (which is the faster you can abort), not past Vr, but past V2, and as the plane was just starting to unstick. So he reduced the throttles and pushed down. Ironically, the first effect of pushing down, before the nose actually goes down, is that the lift in the tail, and hence the net lift, increases. The plane touched down again almost at the zebra bars on the opposite runway end, departed the runway end, run over the grass for a few hundred of feet destroying the localizer antenna, went through the perimeter fence, destroyed a bus stop, crossed the intersection of two avenues destroying a couple of cars, and crashed against an embankment in a golf course across the avenue.

While I seriously think that the take-off was possible, as it was in the Detroit case in a similar minimum runway length condition, and as I was able to do it in the simulator using the procedure explained above (reduce AoA to silence the stick shaker and then keep modulating the elevator to fly at the onset of the stick shaker), in the LAPA case the crew really screwed it up big time by not aborting immediately as the warning activated at the beginning of the take-off run, or at the very least at some point before V1. The Detroit and Spanair cases didn't have that chance because the warning failed to activate.

In the Spanair case, however, the obscene runway length gave the pilot a lot of options. The only thing that the pilot should not have done is to stall the plane, and that's what he did. Like in the Detroit case, the stick shaker activated immediately at rotation and the stall aural warning activated shortly after, and the plane started to rock from side to side with the pilot using full aileron (and hence roll spoiler) inputs. Unlike the the Detroit case, however, this pilot did not make a good, or at least a successful effort, to keep the plane from going beyond the limit of the stall. The plane stalled and hit the ground hard, unfortunately out of the runway and in a zone with obstacles (a ravine and trees).

What should the pilot have done? Again, not to stall. At the first stall warning, reduce AoA. In this case, however, the pilot could have been more aggressive with the nose down since he was not in a hurry to take-off before running out of runway as in the Detroit and LAPA cases. If reducing the AoA resulted in the plane still climbing and accelerating, great, complete the take-off. If the nose-down resulted in a touch-down back on the runway, then he might have chosen to keep accelerating on the runway for a good while and then safely take-off, or abort the take off (yes, he was well past V1 and probably past V2 too like LAPA, but unlike LAPA he still had several thousands of feet of runway ahead).

Obviously, in every case the accident would have been prevented if any one of the following two actions had been performed:
- Don't forget to set the flaps.
- Effectively execute the check-lists.
Undoubtedly these are the causes of the accidents. The warnings failing to work and the pilots failing to fly are contributory factors.

[TeeVee]

that was an awesome explanation. thanks Gabe.

now this might seem stupid, but i'm assuming that the vast majority of flights utilize flaps on takeoff. if so why not prevent the advancement of throttles past a certain point unless the wings are properly configured? sure, it may be a little "airbussy" but it would prevent accidents...

[Gabriel]

that was an awesome explanation. thanks Gabe.

now this might seem stupid, but i'm assuming that the vast majority of flights utilize flaps on takeoff. if so why not prevent the advancement of throttles past a certain point unless the wings are properly configured? sure, it may be a little "airbussy" but it would prevent accidents...
Depending on runway langth, aircraft weight, field elevation, atmospheric conditions and obstructions/terrain to be cleared during the climb, different flaps/slats configuration will be required for take-off.

For example, an MD-80 would likely take-off from La Paz with a slats only / no flaps configuration.

So the plane has to know the proper configuration first.

The MD-80 has a take-off computer where you dial in the required flaps setting and the CG position, with that the plane calculates and displays the correct trim setting. If the trim setting or the flaps setting doesn't match the values entered in or calculated by the take-off computer, the take-off config warning will go out when you advance the throttles.

Ok, additionally to sounding a warning the plane could prevent yo from advancing the throttles past a high taxi setting.

But the main probelm with the warning, at least in the Spanair and Detroit case, was that it failed. So it would have failed to "stop the throttles" too. Obviously the plane needs to know that you are attempting to take-off tho trigger the warning, stop the throttles, whatever (because you very well might want to advance the throttles with the flaps and slats retracted, for example to climb from 35000 to 37000ft).

The logic is:
- Airplane on ground
- Throttles are advanced to a certain level

If the plane fails to recognize that it is on ground, nothing else would work.

[TeeVee]

true but an easy work around might be a landing gear weight sensor...weight on gear, flap/throttle lockout...no weight on gear, throttles free

[Gabriel]

true but an easy work around might be a landing gear weight sensor...weight on gear, flap/throttle lockout...no weight on gear, throttles free
But that's exactly how it works!

A switch senses whether the landing gear's shock absorber strut is extended (no weight on wheels, airplane flying) or compressed (weight on wheels, aircraft on ground).

The sensor failed, the plane "thought" it was in "air mode". So the take-off config warning didn't work, and the throttle lock wouldn't have worked either.