Spanair accident at Madrid

[bonatti]

Whilst i find it difficult to attribute some of the blame on the engineers involved that day, ( the biggest mistake, but not the only, was the pilots failure to follow their checklist) there is one thing i feel uneasy about.


The MEL stated the RAT heater,IF INOPERATIVE, may be isolated and deferred. The RAT heater WAS NOT inoperative , and, as an engineer that would have had me asking questions !

The MEL is a straigt forward document that , unfortunately , is open to abuse should you choose to mis-interperate it.
It is down to the engineers and crews personnel integrity not to abuse the MEL, regardless of commercial pressure. Management and operations can be easily baffled by some technical garb to buy you some time to go and look at the schematics. A further 15 minutes delay isnt going to change anyones day.

[SPA83]

The « punctuality first » logic :

- “Hi ground engineer! I have the engine n°2 fire warning light illuminated!”
- “No problem captain, I’m going to remove the lightbulb!”
- “Thank you very much”

[Joetom]

MEL examples.

Nbr 2 tank fwd boost pump will not switch off ?

Capt wiper will not switch off ?

Left taxi lamp will not switch off ?

F/O seat moves on elects with no request ?

Storm light will not switch off ?

Nbr 1 Rev stuck in deploy ?

Nbr 3 eng oil ind shows full(flt deck), but oil level is half full on glass ?

Nbr 2 pack on when switched off ?

TAT probe inop ?

These are normal defects that the MEL allows for.

A question I have asked before but still no answer on here, I would suspect that the MEL item in question has been used before and again would of expected TOWS to be found INOP before MEL item was fixed, and so would of expected some form of feedback that may have changed the conditions for this item inside the MEL ?

Or may be, this was the first time this event was found ?

Can anybody answer my question ? Tks in adv.

[NigelOnDraft]

A question I have asked before but still no answer on here, I would suspect that the MEL item in question has been used before and again would of expected TOWS to be found INOP before MEL item was fixed, and so would of expected some form of feedback that may have changed the conditions for this item inside the MEL ?

It would seem the use of the MEL/CB had nothing to do with the non-functioning of the TOWS. If the theories quoted here are correct, then pulling the RAT Heater CB leaves the TOWS functioning. The question is whether the RAT Heater "on" at an inappropriate time should have led to futher trouble shooting, and not quick use of the MEL, as per:

The MEL stated the RAT heater,IF INOPERATIVE, may be isolated and deferred. The RAT heater WAS NOT inoperative , and, as an engineer that would have had me asking questions !

Now bonatti says he would have been asking further questions. Good on him, and his company, if that is encouraged / permitted etc.

Regrettably, I would say most companies in practice would have done what was done in the Spanair accident. Can we make the MEL fit? If so, then do it... It does not matter if the "component" is stuck On, Off, Intermittant etc. Often (would not be appropriate here) no CB is pulled - you are just left with an intermittant system and repeated failure messages. On other occasions the CB is pulled - and whether or not the system was strictly "inoperative" before, it is now, and so the MEL can be used.

I'm not saying I like it - just that is in practice what happens

NoD

[Joetom]

Before this accident.

Has a RAT ever been made inop as per the MEL, then found the TOWS also inop because relay R2-5 was problem, then found R2-5 relay change also fixed the RAT ? (or never understood until this accident)

My thinking is if the above has previous, did any form of feedback not change the conditions on the MEL for the RAT item ?

[bubbers44]

Since R2-5 is energized any time the ac is on the ground with electrical power they must have a regular failure rate over 10 or 20 years of operation so this must have happened dozens or 100's of times prior to this time.

[Phalconphixer]

Re-Killer Checks;

I was directed to this video by someone on another quite unrelated forum and it would appear that despite comments made elsewhere in this thread about a flap / slat check being a required item on three of the departure check lists it is not necessarily complied with.
Here is one example…At least the very last minute Killer check worked or maybe it was advancing the throttles to get him around the corner onto the runway kicked off the TOWS?
Aviation Video: McDonnell Douglas MD-83 - Spanair

One thing I cannot get my head around however; reading the report of the MD-82 Detroit accident. As I understand it V1, Vr and V2 are calculated for a given set of conditions associated with weight, temperature, and wing condition ie recommended flaps and slats settings. In the case of the Detroit accident a clean wing would have given a stall speed of 170KIAS and I assume that the Spanair aircraft would have had a roughly similar calculated stall speed given the same loading.

According to the NTSB Accident Report, extending the slats to their normal take off setting would have reduced the stall speed to around 130KIAS and the flap extension would have reduced it to around 122KIAS. Now assuming VR=1.25xVs we get 152kts.

The Spanair aircraft rotated at 154KIAS ie., the nose wheel was lifted at 154 but this was still well below the likely actual stall speed of around 170KIAS because the aircraft was clean wing; the stick shaker operated 4 seconds after the nose wheel was lifted and the stall warning sounded, so in short the aircraft was effectively stalled from the moment it left the ground until it fell out of the sky some 15 seconds later.

Firewalling the throttles and selecting flaps and slats at this time would have done no good at all, since with the best will in the world the crews reaction times coupled with the responding actions and deployment of the systems would have been too late.

If I am talking a load of rowlocks here someone please put me right.(But gently please!)

I am a retired Avionics Engineer with a lifetime of experience behind me but I have never been faced with the kind of pressure that the mechanics were faced with. The last fifteen years of my life were spent with the worlds largest operator of the Dassault Falcon 20 and whilst many will say that there is a world of a difference in the systems employed in this aircraft, I would disagree; for a small biz jet it is actually a mini airliner and has exactly the same dual redundancy systems fitted as many a 737 or MD-80 series aircraft and the operating procedures are the same.

But I never had Flight Ops breathing down my neck to get an unserviceable aircraft full of hot and sweaty Spanish passengers fixed. I never had to suffer the chief pilots wrath because one of our aircraft was five minutes late on task. I worked with many ex-airline staff who had however and it comes as no surprise that things get overlooked in the haste to get the aircraft back in service.

The economic problems being faced by Spanair must have played a large part in the outcome; The co-pilot who handled the second engine start was under redundancy notice and the threat of redundancy was very heavy all round. I have worked under these conditions; its not nice and mistakes do get made as a result of trying not to make a mistake.

I personally would have had second thoughts about WHY the RAT probe was working on the ground, but that is something I can say now with the benefit of hindsight and when I’m away from it all. The action taken by the Spanair mechanics to fix the reported problem was right but right for the wrong reason! Pulling the C/B achieved nothing but to isolate a serviceable probe and protect the associated wiring.

Another question evolves from the accident concerning the TOWS. Why didn’t the crew pick up on the fact that when they taxied out for the second time, presumably clean wing, the TOWS alarm didn’t sound when they opened the throttles?

I have been wracking my brain to see if there could be any other reason for the non deployment of flaps/slats on this aircraft. One report here in Spain states that this aircraft had two instances of flap/slat defects in the ten days leading up to the accident.
I am not privy to the nature of the defects, but the engineer in charge at Spanair Barajas has been questioned about them and the authorities it seems are none to happy with the answers.

The very sad thing is that someone could go to jail for this and it looks like it could be the engineers. In this litigious society we have to have someone to blame, someone to jail, someone to sue. This is so wrong. No-one wanted these poor people to die, no-one profited from their deaths; there was no crime. It was an accident; an accident not caused by a single episode but a whole series of events, which as is so often the case in our business had the inevitable result.

[sevenstrokeroll]

phalconphixer

why do you say the extension of slats/flaps and firewall power wouldn't have helped?

I've trained for that in the sim...it takes more time to write this than to do the above.

REMEMBER that a HUGE SIGN SAYING STALL and a VERY LOUD warning sound would get your attention.

[NigelOnDraft]

7sr...

There is a chance, on some types, that a rapid selection of full power and Flaps/Slats might rescue the situation.

However, it is, IMHO, the "wrong" way to analyse this accident. Firstly we need to deal with an "average crew" on a "bad day" under "pressure" and "distractions". We need to assure the Pax / Safety System we can operate safely... Requiring an instant, unco-ordinated non-SOP response I don't think matches that?

Put it another way, you are asking for a very high standard of flying, analysis, SA and reaction time from a crew who, up to that point, have (for whatever reason) exhibited none of those traits in getting an aircraft to that point without Flaps/Slats.

The "Superior Pilot... avoids use of Superior Skill..." saying springs to mind

PP
You raise some good points. You are possibly slightly incorrect abouit it being stalled from the moment it left the ground... if it had, it would never have lifted off (nor Detroit). In practice, ground effect came into play, allowing both aircraft to reach a low altitude, where the loss of ground effect, and the high drag, took over. Also, swept wings do not always stall in the classic sense - both here and Detroit the "wing rock" at high AoA become significant.

As you conclude, accident are rarely a single cause. Prosecutions help nobody, except in clear cases of malicous intent. All it achieves is make some politicians think they are doing something Errors were made, some by people alive, some probably by people not alive... As you say, the fact that the P2 was basically informed of redundancy that morning is unforgiveable, and IMHO will be a possible or contributory cause...

NoD

[sevenstrokeroll]

NIGEL

one thing that has not been determined is whether or not spanair teaches takeoff stall recovery as I have mentioned.

you speak of superior skills and that...this would also speak to the idea that spanair didn't allocate the resources in their training and evaluation department to create superior pilots.

I recall a TWA L1011 that crashed on takeoff when a stall warning was activated and the copilot gave up flying the plane and the captain put it back on the runway...hard.

at least that one was survivable. imagine if spanair had cut throttles and crashed upon the runway.

oh well, I hope future pilots will remember spanair and detroit in the same breath.

[Phalconphixer]

7SR;
The reason I suggest this is because of the timeline.
From Vr to impact took 15seconds.
From Vr to stick shaker took 6 seconds.
From the initial stick shaker action to impact took 9 seconds.

By my reckoning given that during that 9 seconds the aircraft was rolling left and right and the necessary actions were being taken by the crew to respond to these undemanded rolls, and the fact that the EPR was recording 1.95 anyway (pretty well firewalled anyway, I assume for this sort of reading) I would guess that the crew just ran out of options.

REMEMBER that a HUGE SIGN SAYING STALL and a VERY LOUD warning sound would get your attention.

9 seconds to convert what has effectively become a 130 ton brick back into a flying machine…

9 seconds to diagnose the problem, realise that the slats were not deployed, operate the flap handle, wait for the flap/slat mechanisms to respond and deploy, and become active; they just ran out of time.

The whole point being that in the Sim it doesn’t matter; you have laboratory conditions and no-one is going to get hurt. You can always go back and do it again and the only damage is to your ego…In real life there would be the realisation that this is for real and even the coolest crew would be subject to an element of panic as the need for self preservation kicked in.

So sorry, I’m sticking to that line.

NoD:
Thanks for that explanation about ground effect; that’s the bit I couldn’t get my head around!

[agusaleale]

http://www.faa.gov/other_visit/aviation_industry/airline_operators/airline_safety/safo
A SAFO contains important safety information and may include recommended action. SAFO content should be especially valuable to air carriers in meeting their statutory duty to provide service with the highest possible degree of safety in the public interest. Besides the specific action recommended in a SAFO, an alternative action may be as effective in addressing the safety issue named in the SAFO.

Subject: Importance of Standard Operating Procedures (SOP) as Evidenced by a Take-off Configuration Hazard in Boeing DC-9 series, MD-80 series, MD-90, and B-717 Airplanes.
Purpose: To emphasize the overall importance of SOP and specifically the need for SOP to ensure proper operation of the Take-off Warning System (TOWS) for DC-9 series, MD-80 series, MD-90 and B-717 airplanes.

Background: A recent loss of an MD-82 aircraft during takeoff and a subsequent Airworthiness Directive (AD) by the European Aviation Safety Agency (EASA) serve to underline the criticality of correct take-off configuration. The investigation of this accident is still ongoing and the probable causes have not yet been identified, however, preliminary information released by the investigating authority indicates the airplane’s flaps and slats were not configured for take-off.
A review of accidents and incidents involving civil transport category aircraft shows that, worldwide, take-off configuration errors have figured in 49 accidents and incidents since 1968. These events have resulted in 392 fatalities. It should be noted that the FAA has already taken actions in response to these accidents and incidents such as revising airworthiness standards and issuing ADs. The hazard of mis-configuration of the flaps and slats at take-off can be mitigated in two distinct ways:

1) warning systems, and
2) standard operating procedures.

The recent MD-82 loss underlines the need for the industry to consider its SOP, as well as warning systems when mitigating take-off configuration hazards.

Discussion: DC-9 series, MD-80 series, MD-90 and B-717 airplanes are specifically equipped with a TOWS intended to prevent mis-configuration during take-off. Likewise Original Equipment Manufacturer (OEM) -recommended and air carrier-approved SOP have been designed to prevent a mis-configuration take-off. A warning system and SOP can only be effective mitigations if the system is properly maintained and the SOP is properly designed and followed.
The AD issued by EASA addresses SOP for DC-9 series, MD-80 series, MD-90 and B-717 airplanes. This AD revises Airplane Flight Manual SOP to require the crew to check the TOWS before engine start prior to every flight. This was previously recommended by McDonnell Douglas following a 1987 accident. In the AD, EASA states concern that “some operator’s procedures no longer reflect the initial intent of the [McDonnell Douglas] recommendation…as the check is performed less frequently.” Readers may review the entire AD at the following website: http://ad.easa.europa.eu/ad/2008-0197
SOP are universally recognized as basic to safe aviation operations, as evidenced by the MD-82 example. In 2003, the FAA issued an advisory circular (AC) on SOP, AC 120-71A, “Standard Operating Procedures for Flight Deck Crewmembers”. In that AC, the FAA noted the following key features of SOP:

“KEY FEATURES OF EFFECTIVE SOP.

a. Many experts agree that implementation of any procedure as an SOP is most effective if:

(1) The procedure is appropriate to the situation.
(2) The procedure is practical to use.
(3) Crewmembers understand the reasons for the procedure.
(4) Pilot Flying (PF), Pilot Not Flying (PNF) / Pilot Monitoring (PM), and Flight Engineer duties are clearly delineated.
(5) Effective training is conducted.
(6) The attitudes shown by instructors, check airmen, and managers all reinforce the need for the procedure.”

In order to be most effective, operators should review OEM recommended procedures, define SOP, explain the reason behind the SOP, and effectively train SOP. Each operator should avoid a “double standard” between SOP as trained and as operated in routine practice. To do otherwise is to eliminate one of the most simple and effective hazard mitigations in flight operations. Readers may review the entire AC at the following website:
http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/0/b173ba8a295764f086256cde006a44ad/$FILE/AC120-71A.pdf

Recommended Action: Directors of Operations, Directors of Maintenance, Directors of Safety and Directors of Training should review their procedures to ensure that maintenance procedures and flight crew SOP are effective for ensuring proper operation of a TOWS. Operators of DC-9 series, MD-80 series, MD-90, and B-717 operators may refer to the OEM-recommended procedures for the TOWS. Operators of other airplanes should review their maintenance and flight crew SOP to determine if the procedures achieve a similar assurance of configuration warnings.
Directors of Operations, Directors of Maintenance, Directors of Safety and Directors of Training should ensure that their operations and maintenance personnel are effectively trained in and follow approved standard procedures for their aircraft.

[sevenstrokeroll]

phalconpixer

have you flown the dc9 or md80?

I have. a well trained pilot could have recovered...if spanair doesn't take the time to train well, that's part of the problem.

I was able to speak: firewall power flaps 15...5 times in 9 seconds.

there is no diagnosis, it is a pavlovian bell/response if properly trained.

you think what you want...but if someone had to tell you about ground effect, you need alot more training to figure any of this out.

[bubbers44]

7SR.

Pilots who rotate to normal deck angles and don't get normal lift off performance should not just pull back more to get airborn. Something is wrong and if you have a long runway there is time to make a quick check of what is wrong. Why waste the last few thousand feet that would fix a flap setting error by over rotating? Hopefully grabbing the flap handle and setting them would get you off ok or even if that was too late the extra few thousand feet would get you off clean. Thinking outside the box if you do everything else wrong might save the day.

[justme69]

The judge has decided to cancel the status of indicted to one of the technicians that participated in the maintenance action (RAT heater disabling) done to the aircraft soon before the accident.

The judge estimates his participation as assistant to the other technician, carrying out actions not directly related from the outside of the aircraft and only briefly coming inside to label the heater as inoperative is not enough to warrant his responsability in any potential wrongdoing.

The main technician in charge of the repair and the chief of maintenance for Spanair (who also didn't participate directly nor was informed at the time of any actions on the aircraft) remain indicted.

[justme69]

Not much new.

The judiciary investigation continues. 4 survivors declared today in front of the judge for a couple of hours. 2 more will video-conference tomorrow. Also, the technician that looked into the RAT probe heater the night before the accident in Barcelona will tell the judge what he did about it a week from now.

The survivors, some of them declaring for the first time after being hospitalized (one even in induced coma) for a long time, all agreed that there was no "explosions" or "weird noises" or any "unusual events" other than the airplane feeling like it "lacked power" to elevate, rolled steeply side to side and fell.

The new technical commision that will work for the judge, independent of the CIAIAC one, has been formed today. The judge randomly selected 2 pilots with over 15 years of experience from a list of 30 qualified propossed by the official pilots school (federation). Similarly, 2 aviation engineers and 2 maintenance technicians were selected. All 6 members will be ordered by the judge to carry their own technical investigation and inform him of the findings.

The judge still considers indicted the main technician in charge of taking care of the fault discovered by the pilot soon before the accident (RAT probe heater turned on while on the ground) and the chief of maintenance for Spanair. The judge has decided not to consider the actions of the other technician acting as assistant in the "repair" as enough to warrant an indictment, and therefore he has been classified again as a witness.

[justme69]

Well, and there we have it.

It seems that the investigation pretty much points to the pilots forgetting to deploy the flaps and slats and no related mechanical malfunction other than the TOWS (as it was obvious already to anyone familiar with the situation).

It seems that right when the time to set the flaps/slats on the after start checklist came, the pilot can be heard on the CVR tell the copilot (running the actions) something like: "Take this chance now" meaning to take advantage of a gap in the tower radio busy communications to ask for taxiing clearance, so the checklist was interrupted at that point and permission asked over the radio.

It can be inferred that they later forgot this item and to properly deploy and check the wing configuration and that they only paid lip service (or nothing at all) to any posterior verifications (in the taxi checklist under briefing and in the inminent takeoff under Final Items).

The maintenance technician that checked the aircraft in Barcelona the night before the accident to attend to the problem of "excessive RAT temperature indications" reported by a pilot the day before, declared to have run the whole set of tests as indicated in the repair manual and found all systems working, indicating that perhaps the relay suspected to be faulty had actually an intermitent electrical problem (dirty/stuck contacts, loose or fractured cable connections or whatever).

This same relay, as we know, would've caused the erratic function of two devices: the RAT probe heater, which turned itself on while on the ground right before the accident, and the Take Off Configuration Warning System, which didn't sound in this accident in spite of the investigation pointing to a configuration error (flaps and slats not deployed).

At this time, the investigation officially hints to what we know: crew error due to "rushed" and not 100% "professional" operation of the checklists and procedures coupled with a recent hard-to-diagnose malfunction that wasn't properly noticed or corrected on time and made the configuration alarm inoperative for the takeoff.

The other circunstances surrounding the flight (heavy weight, tail wind, low air density, unfortunate random stall behaviour that deviated the aircraft from the runaway), completed the picture of the accident.

The judge is still trying to determine if the actions of the maintenance technician that disabled the RAT heater minutes before the accident constituted an act of neglicency.

And I guess this about wraps it up until the whole CVR transcript is made public and the official investigation ends a couple of years from now.

The judge is planning on closing the judiciary investigation much sooner, probably in six months or less.

May all the ones that lost their lifes in this unfortunate accident rest in peace. From everyone in pprune, our condolences to their families and friends and we sincerely hope steps can be taken to make as sure as humanly possible that it doesn't happen again.

[Alanwsg]

I've been following this thread with interest from the start and I'd just like to add my thanks to Justme69 for providing an unbiased and pretty comprehensive English commentary throughout.

Thanks.

[hetfield]

I fully agree.

justme69 did a very precise and accurate job here

Thx justme69

[vanHorck]

Justme69 should apply as moderator! Job well done!