Hi and welcome to Failurology; a podcast about engineering failures. I’m your host, Nicole.
And I’m Brian.
Our podcast journey has been a wild ride and we are immensely thankful for all of your support throughout the course of our show. Especially to our Patreon supporters.
But we need a bit of break this summer.
And we didn't want to stop giving you those engineering failures you know and love.
So today’s episode is one of the mini failures we featured on our Patreon page in the fall; the Gimli Glider.
Brian is definitely the plane expert on our show. And I remember the first plane failure I covered before Brian joined us, Air France flight 447. It was like learning a brand new language and it was a challenge just to wrap my head around what happened. But I think Brian has helped me to appreciate the plane failures much more, and now they’re some of my favourites.
Yes, the Gimli glider is a well known almost catastrophic accident that was caused by unit conversion errors.
Units are hard y’all.
So without further adieu, here is our mini failure on the Gimli glider.
Hi and welcome to Failurology; a podcast about engineering failures. I’m your host, Nicole
And I’m Brian. And we’re both from Calgary, AB.
Welcome to our seventh mini-failure episode.
We’re bringing you engineering failures in bite-size pieces.
Make no mistake, these are still significant failures, but they either have pretty straightforward causes or not enough information available for a full episode.
Essentially, we have a list of failures we want to tell you about but haven’t been able to dig up enough info to talk about them for 45 minutes.
These episodes are also just the failure, no news and no ads (for now at least).
It’s like Failurology-lite.
This week’s mini failure is about the Gimli Glider, an Air Canada 767 aircraft that was going from Montreal to Edmonton that didn’t make it all the way.
July 23, 1983
shortly after 8pm central time the fuel pressure alarm sounded in the cockpit for the left side engine
pilots assumed the fuel pump failed and turned off the alarm since the engines were gravity fed in level flight
then the fuel pressure alarm sounded for the right side engine and they diverted course for Winnipeg
almost right away the left engine failed and the pilots planned for a single engine landing, while simultaneously communicating with Winnipeg and trying to restart the left engine
then the right engine also failed – since a scenario with no engines was never expected to occur, they hadn’t covered it in training; this seems like an oversight, its not like you can pull over to the side of the road.
In preparation to land in Winnipeg, they were at about 10,700m or FL 35 at this point.
When they lost both engines, they also lost all of their instruments except for a few basic ones powered by battery. 767 was one of the first jets with an electronic flight instrument system powered by its engines. When both engines stopped working, almost everything went dark. 767 has a ram air turbine (RAT) that swings out from a compartment and converts air flowing past the plane to rotational movement and powers the hydraulic systems and emergency systems (does not power VSI)- which is really good!
The Captain, Bob Pearson, a 15,0000+ hour pilot was an experienced glider pilot,,which is kind of the perfect coincidence here. He knew to fly the plane at optimal glide speed to have maximum range and therefore the largest choice of landing sites. He guessed a speed of 220 knots (410 kph)
Every 10 nautical miles or 19 km they lost 1,500m, giving a glide ratio of 12:1. A dedicated glider can do from 50:1 to 70:1. Gliders that are used by the Air Cadet Regional Gliding School (Northwest) for familiarization flights and glider training have a glide ratio of 22:1.
The First Officer Maurice Quintal had served in the RCAF and was once stationed at Gimli. What he didn’t know was the Gimli had been converted to a race track complex with a road race course, go kart track and dragstrip. A race was underway at the time of the incident.
Without power, they used a gravity drop to lower than landing gear, the main gear locked but the nose did not
As it slowed down to land, the ram air turbine they were using to control the plane also had reduced power, making the landing more difficult
They were coming in a bit too high and too fast and tried to “cross the controls” by applying rudder in one direction and ailerons in the other direction (forward slip). This further impacted the ram air turbine
With both engines out, the plane made very little noise to alert those on the ground
The nose wheel not locking added friction when they landed and helped them stop
There was also a guardrail in place to facilitate the drag track, they applied extra right brake, which caused the main landing gear to straddle the guardrail
Two boys were riding their bikes about 300m away from the landing site. Pearson said they were close enough to see the terror on their face. Could you imagine?!?
Fuel exhaustion due to refueling error
Ran out of fuel at 12,500m or Flight Level 410
Mistaking pounds for kilograms, aircraft only had 45% of its required fuel load
Plane has a fuel-quantity indication system (FQIS) with two redundant channels, but a design flaw caused it to fail when one channel failed. I think they missed the point of redundancy
Following a flight the day before the incident, an engineer in Edmonton ran a service check on C-GAUN’s FQIS, according to a bulletin issued by Boeing. The system failed, which made the fuel gauges go blank as a result. Drawing on experience from a similar incident with the same aircraft a month prior, the engineer, in lieu of spare parts, fixed the problem by disabling the second channel and tagging the circuit breaker.
He informed the pilot flying out of Edmonton the next day that the fuel would need to be measured with a floatstick. However, there was a misunderstanding, and the information made it to Montreal and the change of crews in a highly muddled state.
To complicate matters more, while the plane was on the ground in Montreal, a technician came into the cockpit and reengaged the second channel of the FQIS. Meanwhile, he was distracted by the fuel tank outside and never removed the tag from the circuit breaker. This causes the fuel gauges to remain completely blank.
Another miscommunication led to the crew using only a dripstick measurement of the fuel tanks. Flight computer required the fuel quantity to be entered in kg of fuel, calculation was done in pounds/litre and that value was entered.
Once the plane got to Edmonton, there was a crew ready to install a working FQIS they had borrowed from another airline
The Board of Inquiry found Air Canada at fault for procedures, training and manuals.
The plane had already flown from Edmonton to Toronto to Montreal with a failed FQIS without incident
The misreading occurred once in Montreal and then again in Ottawa, by the same captain who was not used to flying metric. (This was the first aircraft in the Air Canada fleet to use kilograms on the fuel gauges but the measurements needed to be entered in kg/L, fueler who checked floatstick reported the density in pounds/litre as was procedure for other AC aircraft)
Air Canada management was found responsible for “corporate and equipment deficiencies”
The flight and cabin crew were praised for their “Professionalism and skill”
This incident took place around the time flights were switching from three person crews, with a flight engineer, to two person crews. The task of checking fuel load was not properly reassigned from the flight engineer to someone else
Air Canada also needed to keep more spare parts in its maintenance inventory
Captain Pearson was demoted for 6 months and First Officer Quintal was suspended for 2 weeks. Three maintenance workers were also suspended.
Two years later, the pilots were awarded the first-ever Fédération Aéronautique Internationale Diploma for Outstanding Airmanship
There were several attempts by other crews, who were given similar circumstances in a flight simulator in Vancouver, but they all resulted in crashes.
The plane was temporarily repaired in Gimli and flew to a maintenance base in Winnipeg two days later for a full repair. It was then returned to full service with Air Canada. You could have flown on this plane!
The Gimli Glider flew its last flight on Jan 24, 2008, from Montreal to Tucson before retiring in the Mojave Desert in California
It was scrapped for parts in early 2014. That’s sad.
So there you have it, the Gimli Glider. Units are hard, but important. What was thought to be a full refuel was only 45% of the required fuel load, leading to an emergency landing at former RCAF Station Gimli, that had been converted into a drag strip. This could have been a tragic accident, but the pilot's glider experience saved the day and the lives of everyone on board.
Thanks for listening to this mini-failure episode. For our regular episodes, check out Failurology wherever you get your podcasts.
If you want to chat with us, our Twitter handle is @failurology, you can email us at firstname.lastname@example.org, or you can connect with us on Linked In. there are links to all of these in the show notes.
Bye everyone, talk soon!