Original subtitles

Scoops. Pilots from Airborne Express are testing a plane that's been recently

modified. Kick the tires, light the fires, and make sure she's exactly as

advertised.

Next thing is our stall series.

These pilots were trying to test the stall warning system.

Feeling some buffet here.

That's a stall right there.

But something goes terribly wrong.

That's number two engine.

Pull up.

Really? Really?

A rain. A rain.

Flight 827 crashes into a mountain in Virginia.

Everyone on board is killed.

It was just complete devastation. While the flight data confirms the plane

entered a stall... And then here, airspeed falls off a cliff.

It doesn't explain why the pilots weren't able to recover from the very

condition they were testing.

It's a sobering thought and a significant one.

It's three days before Christmas at Piedmont Triad International Airport in

Greensboro, North Carolina.

The crew of Airborne Express Flight 827 and three technicians in the cabin have

been waiting more than four hours for maintenance to be completed.

Think we're getting out of here tonight?

At this rate, I think we're all going to be spending Christmas together in this

cockpit. What do you think, Terry?

I think you can explain that one to my wife and kids.

Roger 827. Happy holidays, dispatch.

Finally.

Just after 5 p.m., Flight 827 is given clearance to fly.

The flight engineer is 52-year-old military veteran Terry Welty.

Taxi in pre-takeoff checklist.

Brakes.

Flying tonight is 37-year-old Captain Keith Lemming.

Checked. Checked.

Cabin and field compressors are off.

Anti-skid.

Armed. While not flying, 48-year-old Captain Garth Avery is the designated

pilot in command.

They were very experienced as pilots. The flight engineer was incredibly

experienced, and they all knew the DC-8 had many, many hours in this airplane.

Taxi and pre-takeoff checklist complete.

Tower ABX-827 Heavy is ready to go.

The flight crew is operating a modified McDonnell Douglas DC-8.

The DC-8 is an incredibly reliable machine, and it always has been.

The DC-8 is a long-range, narrow-body jetliner introduced in 1959.

As long as you use it exactly the way it was designed, they're just as tough as

nails.

G1, rotate.

At 5.40 p.m., flight 827 takes off.

Positive rate.

Gear up.

Positive rate.

1,000 feet.

Roger. Although it's a lesser-known airline, Airborne Express, or ABX, is

one of the largest cargo carriers in the United States.

after both Federal Express and UPS.

In a bid to expand its fleet, ABX purchased the aircraft that would become

Flight 827 and modified it before putting it into service.

ABX 827 is 2500 for 5000.

MCT set.

Ignition off.

But before this plane can be added to the fleet, it has to be taken for a test

flight.

This is known as a functional evaluation flight, or FEF.

A functional evaluation flight is basically to see that everything is as

it's supposed to be. Kick the tires, light the fires, go out and fly, and

make sure she's exactly as advertised.

After departing Greensboro, Flight 827 is scheduled to spend about two hours

flying the FEF before returning to the airport.

ABX 827, for our maintenance check, can we get a block of 10 to 12,000?

The flight crew needs a dedicated block of airspace to perform the tests.

Could you take a block of 13 to 15?

We can do that.

ABX 827, roger. Climb and maintain block altitude of 13,000 through 15,000.

Having reached a quiet block of airspace, the crew works its way through

the FEF checklist.

Scoops.

On.

The DC-8's scoops are part of its pressurization system.

Didn't get the left one. Try again.

It's really common on test flights to have issues crop up. This airplane had

undergone so many changes that it needed a thorough test flight.

All right, scoops off.

Left one ain't working.

With multiple redundancies built into the pressurization system, the pilots

are safe to continue their flight.

ABX 827, turn right, heading of 360, vectors for traffic.

Air traffic control monitors the airspace, redirecting flight 827 when

necessary. ABX 827 right 360.

And then we're going to do our manual drop.

You ready for hydraulics to come off, Keith?

Yeah, go right ahead.

Okay, go ahead, Terry.

The manual drop tests the plane's landing gear without the use of

hydraulics.

Okay, gear freefall 1.5, VS zero.

Okay, you ready?

Ready, gear down.

Gear... Down.

Three green.

Got it.

We're getting a little bit of ice here.

The biggest hazard from icing is that it changes the shape of the wing.

And the airplane doesn't fly nearly as well as it did.

The crew comes up with a plan to avoid performance issues from ice buildup.

Garth, if we go up 15, I think we can get out of this stuff.

Uh, yeah, you can go up there, we've got that block.

As the flight climbs out of the weather, the crew continues the tests.

Next thing is our stall series.

For the stall series, the pilots will intentionally slow the airplane until

it's just about to stall.

to determine when the stick shaker activates.

Most commercial planes use a stick shaker system.

It's a vibrating warning on the control column that alerts the pilots to take

action before the plane actually stalls.

We should stall at 122.

I'm going to set that in my interior bug.

The pilots set their target speed for when the plane is expected to stall.

Mind set.

Shaker 128.

You just call all your numbers. I'll record them.

They have to write down the airspeed at which the stick shaker triggers.

And then they have to write down the airspeed at which the stall begins.

Captain Lemming slows the plane at a rate of one knot per second to reach the

stall speed.

Feeling some buffet here.

The pilots detect a change in the aircraft.

Yeah, that's pretty early.

As an airplane approaches stall, it starts talking to the pilots.

The airplane.

Starts vibrating or buffeting they quickly take action to avoid a stall set

max power 133

Then one

of the planes left engines begins to search

That's

number two engine

But that's not normal, and you know it's not normal, and so something like that's

going to get your attention.

Pull it back.

You got it.

The plane now banks to the left and begins to drop.

The pilots attempt to level the wings and slow the plane's descent.

All right.

Okay? Got it.

Realizing that the plane is descending below his jurisdiction, the controller

assigns the pilots to a new radio frequency.

827, change to Indianapolis 128.4.

But the controller's calls to change frequencies go unanswered. The flight

crew is busy trying to recover the plane.

ABX 827, Indianapolis 128.4.

ABX 827, going to stay on this frequency a minute, descending through 8000, call

you right back.

ABX 827, you're in an emergency descent?

Yes, sir.

Airborne Express Flight 827 is out of control and plummeting to the ground.

Rudder. Yeah, got it.

Seeing that Flight 827 is in trouble, the controller checks air traffic in the

area to find a safe altitude for the plane.

Okay, um, can you hold 7000?

But there's no answer.

Flight A27 is less than 5,000 feet from the ground and descending fast.

Now bring it back.

Left rudder.

Okay, good? Easy?

They could tell they were in an emergency. They could tell they'd lost

control. They were probably frightened to death.

The pilots continue their efforts to recover the aircraft, but are now too

low. Pull up.

Terrain, terrain.

Pull up.

Just half an hour into a standard test flight.

The DC-8 crashes into the East River Mountain in Narrows, Virginia.

ABX-827, do you copy?

ABX-827, do you copy?

No one has survived.

The mountains lit up and there was an explosion, black smoke and

red and blue flames coming up off it.

News of the crash quickly spreads throughout the community.

Brief counsellors arrived today at Airborne's headquarters to help family

and friends cope with their loss.

Obviously, it's just a terrible time for everyone here, particularly at this time

of year.

Bob McIntosh from the NTSB, the National Transportation Safety Board, leads the

investigation. We know this was a maintenance examination flight.

Obviously, it was not successful, and that's about all I can tell you right

now.

The wreckage site certainly presented some challenges.

We were going to have to deal with some mountainous terrain.

We had to get to work.

Benjamin Berman is one of the first team members to arrive at the crash site.

I got to the top of the mountain and I looked out at the scene and it was just

complete devastation.

I really thought, you know, what is it that could have brought that plane down?

That's our job as NTSB investigators.

Investigators begin their painstaking efforts to transport pieces of the

demolished plane to a hangar for examination.

We found the four corners.

Looks like the plane was intact when it hit the mountain.

So if you can see all four corners, which are the nose, the right wing tip,

the tail, and the left wing tip, in close proximity probably means the

airplane was intact until it hit the ground.

This particular wreckage scene told us that we most probably had a loss of

control and that we were going to have to find the reason for that loss of

control. We recovered the CVR and FDR.

Let's get them off to Washington.

The cockpit voice recorder was going to tell us not only what they said, but how

they said it.

While investigators wait for the CVR download... Got the checklist for the

FEF. ...they review the checklist the flight crew was using during the

functional evaluation flight.

to get a better understanding of what they were doing prior to the crash.

Looks like standard FEF stuff.

They're testing the flaps, rudder, landing gear, stall system.

There's nothing unusual on the list of tests the pilots were performing.

Maybe there was a problem with the plane before the FEF.

I'll get Berman to talk to the maintenance team that modified the

plane.

It was an old airplane, but it had just come out of a major retrofit.

Can you take me through exactly what you guys did to the airplane?

Well, it was a fairly major overall.

Investigators learned that extensive maintenance and modifications were

carried out on the aircraft over the course of six months.

Yeah, looks like you guys pretty well took the whole thing apart and put it

back together, huh?

As we looked at the maintenance records, we could find that there was a great

deal of work that had been done on the aircraft.

Perhaps something had fallen through the cracks.

Could technicians have overlooked something during the modification of the

airplane, causing the pilots of Flight 827 to lose control?

Pull up.

Terrain, terrain. Really? Pull up. Really?

Got the dispatch logs for the plane right here.

Investigators review Airborne Express Flight 827's dispatch logs to see if any

malfunctions were reported after the plane was released from maintenance.

Looks like they finished the modifications and delivered the plane on

December 15th, one week before the crash.

Wait a second.

This wasn't the pilot's first attempt at this flight.

The fact that they tried to...

to the flight earlier, that was really relevant.

Look here. The pilots conducted a partial FEF the day before the crash.

Scrapped it midway through, low hydraulic pressure.

That can mean getting the landing gear down and in place or not, or being able

to power the flight controls or not. If the hydraulic pressure is not right,

that's very serious.

Okay.

And what do you think caused the low hydraulic pressure?

Did a hydraulic issue cause a loss of control?

Got it.

Maintenance crews say that they suspected that trapped air in the lines

led to low hydraulic quantity indication.

Did they fix it?

Yeah, they replaced nose gear actuator the day of the accident.

But was that the end of it?

Or could our pilots end up having issues with the hydraulics again?

We needed to figure out what had been going on. Could something have gone

wrong in that maintenance?

Investigators are able to recover the plane's rudder components from the crash

site. They examine them for any signs of damage to the hydraulic system.

No pre-impact hydraulic leaks.

Hydraulic system was working fine.

The hydraulic system wasn't a probable cause for the mishap and loss of

control. If it wasn't the hydraulic system, what was it?

We should call the witness list. Maybe somebody saw something.

You were outside your home at the time?

And what did you hear? The team begins to interview witnesses to see if anyone

heard or saw something that could point to the cause of the crash.

The local newspaper people were there asking people what they had heard and

seen that night.

All you could see was just parts.

It was just a plane had cleared out a big path.

That area was full of information.

Witnesses describe seeing the plane descend out of the clouds at a steep

angle and hit the mountain at high speed.

Thank you so much for your time.

I think we've got something.

Three witnesses confirm they could hear the plane making a banging sound

followed by a slapping sound.

Kind of odd noise.

It sounded kind of low and it started coming up real loud.

The witnesses told us that there was slapping and banging that they heard.

If airflow entering the engine is disrupted, it can cause the engines to

backfire.

If airflow over a wing is disrupted, it can make a slapping sound.

Both can be signs of a stall.

If the witness reports were true, it suggested that possibly there was an

engine issue that may have been part of the sequence of events.

However, it would require further examination of the flight data recorder.

NTSB investigators turned to the plane's flight data recorder to confirm whether

Flight 827 was in a stall before it crashed.

Isolate the airspeed.

There's a gradual reduction in airspeed here. One knot per second. Looks like

they were setting up for the stall series.

And then here, airspeed falls off a cliff.

That's a stall.

Let's see the engine parameters.

They search for further confirmation that Flight 827 stalled.

That looks like a compressor surge from the number two engine.

That tracks with what the witnesses said they were hearing.

Lack of airspeed and a compressor surge are clear signs that the plane stalled.

Give me the roll angle.

They were banking left, right, left, right, left.

They were in a roll reversal.

A roll reversal is a phenomenon that occurs when a plane banks steeply to the

left and right.

The airplane started to roll.

radically from side to side. If you look at the flight recorder traces, you can

see an 80-degree bank one way, 115-degree bank the other way, which is

more than knife-edge flight.

And now the airplane does exactly the opposite of what you were asking it to

do.

Yeah, that's four roll reversals before the crash.

How did the crew of Flight 827 allow their plane to get so out of control?

With the arrival of Flight 827's Cockpit Voice Recorder, or CVR, investigators

zero in on the tests the pilots were conducting immediately before the plane

stalled.

Okay, you ready?

Ready.

Next thing is our stall series.

We should stall at 122.

I'm gonna set that in my interior, Buck.

Mindset.

Shigger 128.

If you just call all your numbers, I'll record them.

Investigators listen to the pilots working their way through the stall

series.

Feeling some buffet here.

Yeah, it's pretty early.

So the plane entered the stall sooner than they expected. Yeah.

When I was looking through the transcript before, there was something

that Captain Lemming said earlier in the flight.

Captain Lemming, we're getting a little bit of ice here.

Garth, if we go up 15, I think we can get out of this stuff.

Yeah, you can go up there. We've got that block.

So ice builds up, disrupts airflow over the wing, decreases lift, increases

drag, and the plane would have stalled at a higher airspeed than they

calculated.

Did ice accumulation on Flight 827 cause a premature stall?

It's happened before.

December the 12th, 1985.

Evidence suggested that shortly after taking off from Gander International

Airport in wintry conditions, a build-up of ice on Arrow Air Flight 1285 caused

the plane to stall at a higher airspeed, leading to a crash that killed all 256

people on board.

Investigators study meteorological reports to determine if Flight 827 was

flying through icing conditions.

There's a scattered area of light rain here along 827's flight path.

Cloud tops out at around 14,000 feet.

What was their altitude block again?

Uh, 13 to 15,000.

The type of weather they were in the night of the accident was really

conducive to icing.

They were in prime territory to pick up ice whenever they were in the clouds.

What speed did the pilots say they expected the stall?

Stall 122, Shaker 128.

Well, according to the FDR, the plane stalled at 126 knots.

It's only four knots earlier.

So even if there was ice, it was likely to have had little impact.

I mean, certainly not enough to make the plane unrecoverable.

Whether there was a great deal of contribution from ice, we could never

determine, but the consensus was that it was not a major element in the cause of

the accident.

You said stall 122 knots, shaker 128?

Well, when we were listening to the CVR, I don't remember hearing the stick

shaker.

The purpose of the stick shaker is to get your undivided attention, not only

by noise, but by shaking you.

Because there's only one thing that that's telling you, and that is, if you

slow anymore, I'm going to stall.

Let's have a listen.

Did the pilots receive proper warning that their plane was entering a stall?

Feeling some buffet.

Yeah, it's pretty early.

That's a stall right there. That ain't no shaker.

So they're slowing the plane down, waiting for the shaker to activate.

Then all of a sudden, they're in a stall. No warning. So the shaker failed.

With no stall warning there, the pilots had, in a sense, a right to believe they

weren't stalling.

Well, what's going on here? Well, it's not a stall because the stick shaker's

not vibrating.

Well, the question is, when did it fail? During the flight or before?

To determine if the stick shaker malfunctioned, investigators searched

the wreckage for stall components that might have survived the fiery crash.

Nothing, huh?

Yeah, that's what we thought. Stall system was completely destroyed.

We never were able to exactly identify what that failure point was in the

system.

Our team wanted to find out how much maintenance and inspection had taken

place on the system.

With no clues in the wreckage, NTSB investigators speak to the maintenance

technician about Flight 827's stall warning system.

Thanks for your time.

Did you test the stick shaker?

Yeah, we did.

On... On December 5th.

Two weeks before the accident. How extensive the test?

Quite extensive.

Have a look.

The shaker, the heating components.

The stick shaker system passed all those pre-flight maintenance checks.

What about the transducer?

Let me check.

The transducer is a small flap nestled in the recessed part of the wing.

It moves backwards when airflow is detected.

When there is insufficient air over the wings it moves forward and triggers the

stick shaker to activate No,

no, it doesn't look like we tested the transducer That

part is not covered by the pre-flight test It suggests that the failure of the

stick shaker to work was a bad transducer

But to me, why it failed wasn't as important as what the effect of it

failing was on the crew.

Feeling some buffet here.

Yeah, it's pretty early.

That's a stall right there. That ain't no shaker.

Even without the stick shaker activating, the flight crew knew they

were in a stall.

That's number two engine.

Bring it back.

Okay, good.

But if they knew...

Left rudder. Why couldn't they recover?

Easy. I bring it back.

Rudder.

Really?

Investigators return to Flight 827's CVR to get a better sense of the actions the

pilots took when their aircraft went into a stall.

That's a stall right there. That ain't no shaker.

Set max power.

133.

So as soon as they recognize they're in a stall, they push power to the engines.

That's the right move.

That's number two engine.

There's the engine, Serge.

Pull it back.

So what do they do next?

All right.

Okay.

Start bringing the nose back up.

They pulled the nose up. But that would make the stall worse.

Pulling back on the control column in response to a stall is inconsistent with

a normal stall recovery and would not reduce angle of attack and break the

stall.

Well, that's what would have triggered the roll reversal.

How'd they deal with that?

Rudder? Got it. Rudder.

Got it.

Left rudder. Left rudder's buried.

Investigators discover that in addition to moving the control wheel right and

left, the pilots also used the rudder to stop the plane from rolling.

You have to be ahead of it, and that's a very difficult thing to do.

So they spend almost a minute trying to fight the roll reversal with the rudder.

No mention of a stall.

Investigators suspect that the pilots were so preoccupied dealing with the

roll reversal, that they failed to realize that the plane was in a stall.

Still doesn't explain why Captain Lemming would have pulled back on the

yoke to try to recover from the stall.

To recover from the stall, you can't pull back.

You need to point the nose of the airplane down even more in order to gain

speed, or you will not recover.

You will fly into the mountaintop in Virginia.

And we wanted to know why.

Why did he do that?

Especially at those most critical moments.

I'm looking to see if there's anything in ABX's operations manual on the DC-8.

Here's a copy.

Investigators searched through ABX's procedures to understand why Captain

Lemming pulled back on the control column to recover from Flight 827's

stall.

Wait.

Bob, did you see this?

It says that priority in a stall recovery is to maintain altitude by

increasing the pitch.

Pulling back on the control column increases the pitch of an aircraft.

Well, it might be okay for an approach to a stall, but not for an actual stall.

So the things that you do to recover from an approach to stall are totally

different from what you do to recover from an actual stall.

If you're right over the trees, you've got to try the approach to stall

recovery because you're going to hit the trees.

If you're at 36,000 feet and you stall, if you don't put the nose down 5 or 10

degrees, you're not going to recover.

That can't be the current procedure, can it?

I've got a joint memo with the FAA.

Investigators discover an update to Airborne Express's flight test

procedures. And what's it say?

The updated procedure called for a reduction of pitch to recover from a

stall.

The revised procedure called for pilots to lower the nose when entering a stall,

the opposite of what Captain Lemming did.

The change was made after a loss of control incident in May 1991.

What flight was that?

It was a DC-8 doing a post-modification FEF.

The same plane and flight as the 827.

On May the 16th, 1991, another airborne express, DC-8, lost control during an

FEF.

The pilots managed to recover from the stall, and the policy has been to reduce

pitch during a stall ever since.

So why didn't the crew of Flight 827 use the updated procedure?

Okay.

We learned that the Director of Flight Technical Programs was responsible for

the procedures.

the checklists, and the selection and training of pilots.

As a result, we wanted to talk to him to learn more.

The NTSB meets with the airline's Director of Flight Technical Programs to

determine why the crew of Flight 827 didn't follow the updated procedure.

Just have a few questions to ask, and then I'll let you get back to work. Of

course.

How do ABX pilots train to recover from a stall?

I tell them whatever you do, don't lose altitude. So if you're in a stall, you

set max power and you increase the pitch of the aircraft.

Well, that's different from the updated policy, which calls for a nose-down

pitch recovery.

Correct.

He just didn't agree that that was necessary.

If you increase the pitch, it will make the stall even worse.

If the engines are properly spooled as you're approaching a stall, you can

power your way out of it.

The director believed that if the engines were running properly...

Pilots could increase thrust and fly out of a stall.

He had his own opinions about what should be done, and they were not in

agreement with the proper procedure.

See, what I don't get is, old procedure, new procedure, there is no procedure

that says to pull back for the duration of the stall.

I keep going back to the way they were dealing with this roll reversal.

Rudder. Rudder. Got it.

And left rudder.

The crew were unable to recover the airplane in the approximately 90 seconds

that it took from when the lack of the stick shaker was identified to when it

impacted terrain, and our question was why.

Investigators assess ABX's DC-8 flight simulator to see exactly how the flight

crew experienced a stall during their training.

Now I'm going to slow this thing down into a stall.

There's the shaker.

There's the stall.

Now I'm going to pull back on the yoke like our pilots did.

So I...

continued to raise the nose up higher and higher, and I was just astounded at

what the simulator did, or mainly what it didn't do.

Well, we're in a stall, and there's no roll, no pitch down, nothing like you

would see in a real-life stall.

The simulator just kind of reached the end of its program.

It just sat there.

If this is how they were trained, they would have no idea how a stall truly

feels.

Their training was not only zero training, it was negative training.

It gave them a false impression of what they would experience.

They would have been startled at what the wings did with the roll. Okay, easy.

Don't, don't.

And... they would not have recognized it necessarily as a symptom of a stall.

Because in their simulator, it didn't have that symptom.

Maybe they didn't realize they were still in the stall, but surely there

were other indicators that they were falling out of the sky.

Why didn't the pilots of Flight 827 realize they were diving towards a

mountain?

Okay, put yourself in the pilot's shoes.

You realize you're in a stall.

Feeling some buffet here. Yeah, it's pretty early.

That's a stall right there. That ain't no shaker.

You push power to the engines, pull back on the yoke.

Start bringing the nose back up.

Your number two engine surges.

That's number two engine. Pull it back.

Now you're dealing with a roll.

Planes banking left and right.

You're applying rudder.

Rudder. Got it. Rudder.

Rudder. Got it.

You've got ATC in your ear telling you to switch frequencies.

827, change to Indianapolis 128.4.

You look outside. What do you see?

Nothing.

Right. Because it's pitch black out and you're in the clouds.

There's no visible horizon.

You're just flying on instruments.

They had no visual cues.

Exactly.

You're now removing the visual aspect of a horizon out there that can help anchor

you. You're just stacking the deck against yourself.

Now, the question is, what were they doing flying a stall series test at

night in the first place?

Investigators review the flight's dispatch logs to retrace the pilot's

steps on the day of the accident.

The FEF was scheduled for 1.20 p.m., but the maintenance delays pushed it past

sundown. Had they just pushed it to the next day, none of this would have

happened. So why didn't they?

Why go through with the FEF at night?

All right.

One of the key decisions was to conduct the check flight.

later at night than they'd planned in the darkness hours.

And so, you know, did anyone pressure him to doing that?

What is the ABX policy on when to fly an FEF?

Investigators return to ABX's Director of Flight Technical Programs to find out

why the pilots proceeded with an FEF at night.

No policy necessarily, just a preference to fly them during the day.

Do ABX pilots understand the risks of flying an FEF at night? Sure.

And I'd rather they conduct them during the day, but there's no prohibition

against doing them at night. What are the recommended weather conditions for

flying an FEF?

Preferences to fly in clear skies and not on instrument, but we have no

specific restrictions on that.

We discovered that there was no written policy prohibiting this flight from

being conducted at night.

The NTSB's investigation has uncovered numerous factors that led to the crash

of Flight 827.

Stick shaker malfunction, incorrect stall procedure, negative simulator

training, FEF at night.

Had just one of these factors not been present, the crew would likely have

survived.

I think this accident was caused by a failure to recognize that a functional

evaluation flight requires rigorous preparation and criteria and controls to

be placed on it to protect the pilots who are flying it.

Start bringing the nose back up.

This was not God reaching down and slapping an airplane out of the sky.

There were reasons for this happening, and this could have been avoided.

In the wake of the crash of Flight 827, the NTSB makes a series of

recommendations to make functional evaluation flights safer and to better

prepare pilots for stall conditions.

This accident brought a lot of significant changes to the industry.

We issued recommendations to try to improve airline oversight and FAA

oversight of functional evaluation programs.

Those efforts were successful.

Flight simulators used for airline training were made much more realistic.

Every airline pilot receives training in recovering the airplane from the high

altitude stall where they have to use this airborne express revised stall

recovery procedure to get the job done.

The report also recommends a revision to the DC-8 maintenance manual.

calling for regular calibration and testing of the complete stall warning

system. There might have been other airplanes out there, you know, flying

around in service where the stall warning system really wasn't working.

It's a sobering thought and a significant one.

The strength of aviation safety is a direct result of learning the most we

can learn from every single accident.

This wasn't an accident in vain. Yes, it could have been prevented, but now we

know how.