Re: Canadian 757 out of fuel incident?

From:         jmk@cbvox1.cb.att.com (Joe Knapp)
Organization: AT&T
Date:         30 Jun 94 00:34:18 
References:   1
View raw article
  or MIME structure

Here is the scoop:

Wall Street Journal, December 12, 1984

   Crisis Aloft: 
   Electronics and All, Airliners Still Run Out Of Fuel in Midair

   -----

   A Chain of Errors Caused Air Canada Jet to Make Spine-Tingling Landing

   -----

   Silent Crying, But no Panic

By William M. Carley
Staff Reporter of The Wall Street Journal

     Robert Pearson, the pilot of an Air Canada Boeing 767 jet,
felt all was going well as his plane flew westward 41,000
feet over Red Lake, Ontario. "I'm going to sit here and watch
the trout swimming in the lake," he jokingly told his
co-pilot.
     But suddenly warning lights blinked on in the cockpit,
then alarm gongs sounded, and the jet's two engines both
flamed out. The plane, carrying 69 passengers and crew, had
run out of fuel in midair.
     What followed on that flight on July 23 of last year
marked one of the most harrowing half-hours in aviation
history. The Boeing 767 can glide, and the captain, a former
glider pilot, decided to try for Winnipeg airport, 65 miles
away. Alerted by radio, air controllers cleared the skies
around Winnipeg. Fire trucks were called to the runways. But
the Air Canada jet was dropping too fast -- it couldn't make
Winnipeg.
     Capt. Pearson made a sharp right turn toward Gimli, a
former air-force base nearby. There were people and vehicles
on one end of the Gimli runway, which was being used an an
automobile race track. As the plane touched down, its nose
hit the runway. Amid a shower of sparks and grinding metal,
the jet slid nose-down to a safe stop, just short of the
throng of people and vehicles.
     Even though the world's airlines have entered the age of
electronic fuel gauges and have long had redundant fuel
warning systems and spare tanks, airplanes are still running
out of fuel. In the last few years, in fact, there has been a
rash of incidents involving planes that ran out, or nearly
ran out, of fuel supplies.

Some examples:

     -- Last year, a Republic Airlines DC-9 ran short of fuel
     en route from Fresno, Calif., to Phoenix, Ariz. After the jet
     was forced to land at Luke Air Force Base, Ariz., 25 miles
     from Phoenix's commercial airport, safety investigators found
     that just five gallons of usable fuel remained in the tanks,
     enough for a few minutes of flying.

     -- A 747 jumbo jet flown by Pan American World Airways
     barely made it into Newark, N.J., airport a few years ago. As
     the big plane touched down on the runway, two of its four
     engines flamed out, and on a taxiway a third died, all for
     lack of fuel.

     -- While circling near Portland, Ore., a United Airlines
     DC-8 ran out of fuel and crashed in 1978. Among the 186
     aboard, two crew members and eight passengers were killed.
     Although such accidents are relatively rare, eliminating
     them is difficult. That is because the causes often are
     complex, sometimes involving mechanical failures, sometimes
     human faults, and often a combination of both, safety experts
     say.

     The Air Canada case is in many ways typical. There was a
long string of both mechanical and human problems that led up
to the spine-tingling landing at Gimli. Details of the case
have emerged in months of hearings before a board of inquiry
set up in Winnipeg by the Canadian government.
     The jet involved wouldn't seem a likely candidate for the
Gimli drama. It was Boeing Co.'s newly designed 767, a big,
wide-body plane with two powerful engines and the most
advanced electronic systems -- including fuel gauges -- in
the world.
     Months before the Air Canada accident, however, some
ominous signs began appearing. After United Airlines had
taken delivery and was ferrying one of its new 767s from
Boeing's production plant near Seattle, the pilots noticed a
problem. The fuel gauge showed an amount of fuel in the tanks
different from the amount the plane's flight computer
calculated should have been there.
     According to Boeing, later tests showed that "an
undetected fault could cause erroneous readings (in 767 fuel
gauges) of 1,000 to 3,400 pounds above the actual fuel
quantity." Because that posed the threat of running out of
fuel in midair, Boeing quickly told airlines around the world
to inspect 767 fuel-measuring systems for the fault.
Honeywell Inc., which supplies the fuel-measuring system to
Boeing, began redesigning it.
     Meanwhile, Conrad Yaremko, an Air Canada mechanic in
Edmonton, began working on aircraft No. 604 on July 5 of last
year. Because it was one of Air Canada's brand-new Boeing
767s, he checked for the fuel-measuring system fault.
Mysteriously, each of the three fuel gauges -- one for each
of the three fuel tanks -- went blank during the test. Later,
all seemed to work properly.
     The night of July 22-23 saw the same big jet back in
Edmonton. Mr. Yaremko was performing the same check, and
again the gauges mysteriously went blank. It was a warm
night, and Mr. Yaremko knew that heat is the enemy of
electronic circuits. So he removed the small box of
electronic gear that feeds data to the fuel gauges in the
cockpit.
     Then, he said in testimony at the board of inquiry, "We
put it {the electronics box} in the fridge for a little bit
to see if it would cool it down." The tactic didn't work.
Eventually the gauges did begin working again, and the
plane flew to Montreal -- where they went blank again. As
Capt. Pearson boarded the jet the afternoon of July 23 to fly
it from Montreal to Ottawa and then to Edmonton, he saw all
the fuel gauges were blank. This time mechanics couldn't get
them to work again.
     A hubbub in the cockpit area ensued. Capt. Pearson, his
co-pilot, Maurice Quintal, some mechanics and a stewardess
were there. According to testimony at the board of inquiry,
it was decided to use dipsticks to measure the amount of fuel
in the tanks.
     The dipsticks were calibrated in centimeters. That
measurement should have been converted to liters and then to
kilograms, because the 767 is Air Canada's first plane to use
the metric system and its fuel is measured in kilograms. The
Air Canada men, however, didn't do that. Instead of
multiplying liters by 0.8 to arrive at kilograms, they
mistakenly multiplied by 1.77 and arrived at pounds -- the
measure used for other Air Canada planes. Then, on the basis
of the erroneous calculation, they added fuel. Capt. Pearson
checked the figures, not realizing they were faulty, and
ordered a bit more fuel to balance the tanks in two wings.
Robert Desjardins, the chief flight attendant, was a bit
nervous about the unusual fueling procedure. "We better have
more than not enough," he remarked as mechanics were closing
the jet's door.
     "You've got more than enough. You can go all the way to
Vancouver," one mechanic replied.
     The flight from Montreal to Ottawa was uneventful. Then
the jet took off for Edmonton -- with only about half the
fuel needed to reach that destination.
     As the jet passed Red Lake (and Capt. Pearson joked about
watching the trout), co-pilot Quintal began making an
announcement to the passengers, according to the cockpit
voice recording.
     "Good evening, ladies and gentlemen, this is your first
officer. We're presently coming up over Red Lake. . . .,
cruising at 41,000 feet, the temperature in Edmonton is,
beautiful day. . . ."
     But 10 minutes later, at 8:10 p.m., four beeps warning of
low fuel pressure sounded.

     Capt. Pearson: Holy {expletive}!

     Co-pilot Quintal: Something's wrong with the fuel pump. .
     . .

     Capt. Pearson: Left-forward fuel pump, okay, what have we
     got here? I hope it's just the {expletive} pump failing, I'll
     tell you that. . . .

     BEEP, BEEP, BEEP, BEEP. . . .

     Capt. Pearson: Let's head for Winnipeg Now. . . . Hundred
     and twenty-eight miles (from Winnipeg), okay. . . .

     Capt. Pearson had his co-pilot radio air controllers to
tell them the jet had some problems and was diverting to
Winnipeg. He also had his co-pilot brief Mr. Desjardins, the
chief flight attendant, telling him to prepare passengers for
an emergency landing. And the captain began descending from
41,000.
     Then more beeps, indicating problems at more fuel pumps.
Co-pilot Quintal: {Expletive}, they're all going out. . .

     Capt. Pearson: All the {warning} lights are on!

     A chilling conviction settled over the cockpit. As Capt.
Pearson tells it later, "Maurice and I knew that one fuel
pump might fail, but to have all the pumps fail all at once,
and on a brand-new airplane, that was beyond probability. We
realized we had more than a pump problem, we had a fuel
problem."
     A few minutes later that was confirmed; a deep "bong"
sounded.

     Capt. Pearson: Okay, we've lost the left engine.

     Co-pilot Quintal: Okay, what will we do . . .?

     Capt. Pearson: Yeah, just run on one {engine}, let's just
     run on one {to save fuel}.

     Co-pilot Quintal radioed Winnipeg air-control center
warning the jet had lost one engine and requesting that fire
trucks be called out. Three minutes later the second engine
flamed out. With power gone, eight bright video tubes in the
cockpit displaying most of the plane's instruments went
blank. ("That cockpit became the darkest place in the world,"
Capt. Pearson said later.) The captain radioed Winnipeg
air-control center.

     Capt. Pearson: Center, (Air Canada flight) 143. This is a
     Mayday, and we require a vector (direction) onto the closest
     available runway. We are (down to) 22,000 feet (altitude) on
     -- with both engines have failed due to, looks like fuel
     starvation, and we are on emergency instruments. . . . Now
     please give us a vector to the nearest runway."

     Back in the passenger compartment, stewardesses were
rushing to secure the drink trolleys. Then they briefed the
61 passengers, who had been told the plane had a fuel
problem, but no more than that, on how to brace themselves.
There were businessmen on board, an older couple, some
teenagers and some young mothers with infants. After the
briefing, all fell silent.
     "Obviously there was no engine noise . . ., it was very
quiet in the airplane except for a few people crying," Mr.
Desjardins, the flight attendant, told the board of inquiry.
"People were very afraid, and some were crying silently, but
there was no panic," Anne Swift, another flight attendant,
told the board.
     Among the few emergency instruments functioning was the
air-speed indicator. Capt. Pearson's first job was to select
the right speed. Too slow, and the jet would stall and crash;
too fast, and the plane would dive toward the ground. There
is nothing in pilot manuals on how to handle a gliding 767.
So, using the controls, the captain nosed the plane down just
enough to maintain 220 knots, the speed that he guessed would
provide the longest, safest glide.
     (The flight controls worked only because, when the last
engine had flamed out, a tiny air turbine had descended
automatically from the belly of the 767. Spun by the passing
air, the turbine provided power for the 767's "power
steering" -- hydraulics that enable the captain to operate
controls such as ailerons on the wings and the rudder on the
tail. All 767s have such turbines; but many other jets,
including 747s and 727s, do not have them.)
     Capt. Pearson tried to steer for Winnipeg, but his compass
heading had disappeared along with his video displays. He
still had a small emergency magnetic compass on the
dashboard, but it was swinging too much and was mounted too
far to his side for him to line up on.
     "So I steered by the clouds underneath us," the captain
said later in an interview. "I would ask Winnipeg center for
a heading, they would say 'left to 220 degrees,' and I would
turn left about that much, judging by the clouds, and then
I'd ask Winnipeg how my heading was. Using the clouds, I kept
eyeballing it."
     The 767 had also lost its vertical-speed indicator, a
measure of how fast the plane was dropping. And it was
dropping much faster than Capt. Pearson thought. Co-pilot
Quintal tried to calculate the rate of descent. The jet's
altitude was 14,500 feet when the Winnipeg controller said
his radar showed the plane 45 miles away. The jet had dropped
to 9,500 feet by the time the plane was 35 miles away.
     "We had dropped 5,000 feet in 10 miles," Co-pilot Quintal
says. "With only about 10,000 feet of altitude left, we could
glide only another 20 miles. But Winnipeg was 35 miles away
-- I told Bob (Pearson) that we'd never make it."
     Capt. Pearson radioed Winnipeg, and the Winnipeg
controller told him that Gimli was just 12 miles away, on his
right. Co-pilot Quintal, who had been based at Gimli while
serving in the Canadian air force, told the captain the
airport had long runways. The base, which had been obscured
by a cloud, was suddenly visible.
     "Bob turned right towards Gimli, like 'whoomph!'" co-pilot
Quintal says.
     The captain asked the Winnipeg controller for details on
Gimli. Use the right-hand runway, 6,800 feet long, the
controller replied.

     Capt. Pearson: There will be nobody on the runway when we
     get there, eh! Nothing?

     Winnipeg Controller: I don't know -- I can't tell you for
     sure. . . .

     Co-pilot Quintal saw the Gimli strip and pointed it out.

     Capt. Pearson: We're going to make Gimli okay.

     Winnipeg Controller: Great! We show you about six miles to
     touchdown.

     But all wasn't great. The jet was coming in too fast and
too high. The captain and his co-pilot discussed circling
once before landing but rejected the idea because they would
lose sight of the runway while circling and because they
might lose too much altitude.
     To cut the jet's speed and altitude, Capt. Pearson then
tried a rarely used technique, a side-slip, which he had
practiced as a glider pilot. With the rudder, he turned the
nose of the plane right, but with the ailerons he banked the
plane to the left. The net effect was to make the plane fly
toward the runway in a sideways attitude that would create
drag and rapidly slow the plane.
     The captain's maneuver, in which the plane seemed to
plummet to the left, frightened passengers and flight
attendants. "I thought he had lost control," Mr. Desjardins
said. But the plane began losing altitude quickly, and speed
began dropping from 210 knots to 170 knots, still much faster
than normal landing speed of 130 knots.

     Winnipeg Controller: Five miles to touchdown.

     Capt. Pearson: Roger, we have the field in sight. . . .

     Co-pilot Quintal flipped the landing-gear lever. But
without power, nothing happened. "There wasn't any noise
(from the gear going down), and that was a terrible feeling,"
the co-pilot says. He turned another switch allowing the gear
to fall by gravity. The gear on each wing dropped, but the
nose landing gear deployed only partway down.
     The plane bored in at 170 knots. As it hit the runway, the
captain's eyes were fixed on the runway threshold. The
co-pilot was frantically scanning a checklist on how to get
the nose gear all the way down.
     Then both looked up -- and ahead on the runway they saw
vehicles, people walking and youngsters bicycling. They
hadn't landed on the right-hand runway, which was still used
for aircraft, because its dark color had blended in with the
earth. They had landed on the lighter and more visible left
runway, which was being used as an autoracing strip. The last
race of the day had finished, and scores of people were
strolling and biking on the far end of the runway, while
others were finishing dinner in camping vehicles parked on
the runway.
     Without power, the captain couldn't use reverse thrust to
slow the plane. He hit the brakes. Because the nose landing
gear wasn't fully down, the plane tilted forward. Its nose
ground into the concrete, sending up a huge shower of sparks.
"I thought the underneath of the airplane was being torn
apart," Mr. Desjardins, the flight attendant, said.
     As the plane continued sliding, Jo-Ann Barry was washing
dinner dishes in the family camper parked on the runway.
     Suddenly a boy on a bike yelled that a jet was coming in. "I
opened the camper door and there was this huge plane coming
at us," she says. "It seemed very quiet, but it was coming
70-80 miles an hour." Her husband yelled at their
five-year-old son to get off the runway, then grabbed their
two-year-old son under his arm and sprinted off himself.
"Everybody was hollering, grabbing their kids and running,"
Mrs. Barry says.
     Fortunately, the jet's nose acted as an excellent brake.
The big plane, its tail high in the air and its nose on the
runway, shuddered to a stop just short of the throng of
people and vehicles. No one had been injured.
     What had gone wrong with the plane's fuel-measuring
system? Because of a loose wire in the Honeywell electronic
box, one data processor failed and couldn't provide
information to the fuel gauges. The system then should have
shifted automatically to a second, redundant data processor
in the electronic box. But because of a shortcoming in the
Honeywell system's logic, it failed to do so. Hence, all the
fuel gauges went totally blank.
     There were, of course, other problems, including the fuel
conversion to pounds instead of kilograms. Even if the fuel
had been calculated correctly, however, Capt. Pearson
shouldn't have taken off. According to the 767's Minimum
Equipment List, if at least two fuel gauges (for two of the
three fuel tanks) aren't functioning, the pilot isn't allowed
to depart. Capt. Pearson testified mechanics convinced him
that, even though the fuel gauges were all blank, the
dipstick procedure was sufficient to bypass the Minimum
Equipment List. Mechanics denied that.
     Although Capt. Pearson and co-pilot Quintal were praised
for landing the plane safely, Air Canada said it would demote
the captain to co-pilot for six months and suspend the
co-pilot for two weeks. Both are appealing the disciplinary
measures under union procedures, Air Canada says.
     Meanwhile, Air Canada has revised its fueling procedures
and has retrained its crews in fueling methods. And Boeing
767s in airline fleets around the world are getting a
redesigned Honeywell fuel-measuring system.
     As for the plane itself, Air Canada's aircraft No. 604 is
still in service. In aviation circles, the big jet has come
to be called "the Gimli Glider."

A Near Disaster

     8:10 p.m. -- First indication of low fuel pressure

     8:12 p.m. -- Jet turns toward Winnipeg

     8:14 p.m. -- Jet, at 41,000 feet, begins descending

     8:18 p.m. -- Left engine fails

     8:21 p.m. -- Right engine fails, jet at 26,000 feet begins gliding

     8:31 p.m. -- 35 miles to Winnipeg, "we'll never make it"

     8:33 p.m. -- Jet turns toward Gimli

     8:35 p.m. -- Six miles to Gimli

     8:38 p.m. -- Touchdown