Autonomy versus automation, a perennial discussion. But the nature of human intervention in flight is much more subtle than simply the mechanics of flying. The autopilot was selected as we climbed out of Heathrow. The blue sky beckoned, the south coast almost unnaturally stark on the clearest of summer days. A heavy Airbus A321 needs a little bit of attention on a turning departure so we had planned a pause in the flap retraction sequence until the heading change was complete. It afforded us a better turn radius. A bit more margin over the lower end of the speed range. A bit more resilience should we bounce through a bit of wake. Continue reading
This has become a bit of a meander through the Airbus vs Boeing subject. We have been diverted. Part 3 will be all about the navigation systems and concepts in both types. I promise. But right now we are sitting on the tarmac at Safety Stats International Airport. I have to admit to finding this unscheduled stop over more fun than I was expecting.
When my buddy Iain (flyer, thinker and expert-level counter-argument-maker) tweeted me with some Airbus stats showing that Airbus fly-by-wire was safer than other airliners I wanted to know more. Safer? Really?
If that is the case then these stats could form the basis of an argument against human pilots. A potential counter-narrative to skill.
The A320 was designed to solve a problem which may have only partially existed: airliners were too difficult to fly. The design should be easier to fly and therefore be safer.
Looking back over 20 plus years of Airbus operation the question is really whether the Airbus approach is safer and has made any real difference? Airbus will tell you so. Of course they will, they believe in their product. Iain’s stats are testament to that. The Airbus plan was to remove subtlety and technique at the fringes of the flight envelope and to replace it with accuracy and control without specialist skill.
To really work a wing hard takes practice, capability and skill. Any wing performs at its best really close to the stall; the maximum amount of lift is generated in this regime. When the need to manoeuvre quickly arises the required skill level rises exponentially. Getting it wrong by pulling too hard or not hard enough diminishes performance.
You can spend time looking up lift curves and learning about the bernoulli principle but the fact of the matter is that there is a fine and ragged edge upon which a pilot must dance if maximum manoeuvring performance is required. The dancing is all the more difficult if your partner is big and lumbering like an airliner.
In my old Pitts I could feel everything that the wing was doing. Through the stick, the ailerons, the elevators. Even the airflow separating and reattaching to the canopy told you something. In a bigger aircraft, that feel just is not there. There is too much metal around you, too much distance.
The Airbus flight control system is designed as a manoeuvre demand device. You want ‘up’? Then just pull; full in any direction means that the aircraft will give you all of the available performance in that direction without breaking the machine. All of it according to the laws of physics, not just what Airbus thinks is the maximum.
It’s a compelling selling point. The dance is danced for you. Beneath the sidestick. On a motherboard with millions of electrons. There’s a party going on down there. When you see it in action, when you watch the result, it is a marvel: the jet spearing up through the atmosphere on the fringes of what is possible. But the key question remains: does every last ounce of performance really count in the real world?
Imagine if you suddenly found yourself very close to a mountain, or perhaps facing a monstrous downward blast of air from a storm cell; the Airbus system might seem very attractive. In fact, it’s pretty certain that pilot-for-pilot, manoeuvre-for-manoeuvre the Airbus wins every time.
A side note here: When we are talking about performance what we are really thinking about is the wing describing the smallest possible radius in the sky and then using all of the available lift to climb at the steepest angle. We are interested in changing direction as rapidly as possible, then using that energy. The physics of the wing limits this energy, the way in which it generates lift and the mass and inertia of the aircraft.
Speed, range, load-lifting and other traditional “performance” measures are not immediately important to us because when things go bad in flight changing direction really promptly is what is needed.
The A320 was thought up in the early 80s. The density of flying back in the 60s and 70s was nothing like as common as we know today. The decades before had seen a terrible toll exacted upon the industry and travelling public. That rate had to be improved if air travel was to become more accessible for all.
The accident rate in the 60s was over 300 fatalities per 1,000,000 departures, today it is around 50 despite the growth in aircraft size. The very worst accidents involved a handful of common factors: flying into mountainous terrain (CFIT), windshear accidents, midair or ground collision.
The energy of the industry was focused on the ‘most wanted’. It was obvious that no one silver bullet would solve the issues. But having an aircraft that could maximise manoeuvring performance whilst minimising the skill required might be a huge stride in the right direction.
Adrift in the sea of data in this report Boeing has provided the closest thing to a scientific control that we could ask for: the 737.
The 737 has been around for over 50 years. And yet the oldest and the newest 737 fly using the very same principles. It is a cable and pulley aircraft, with some added hydraulics which uses aerodynamics and some simple devices like a stick shaker to warn the pilots that they are approaching the limits of performance.
Sitting in a 737 isn’t wildly different to sitting in a B-17 from World War II. The control wheel, the thrust levers, the size of the windscreen. It is all fairly similar. The latest 737 Max looks pretty fancy. But the beating heart of the 737 is blue denim, stonewashed, rust belt America. It does the job. No fanfare, no nonsense, no gimmicks.
I quite enjoyed my short stint on the 737. I never did learn to land it properly. But it taught me lots and I did plenty of hand flying despite the excellent autopilot. It taught me to feel and fly a jet airliner. The 757 and 767 are a generation on and frankly are much more wonderful to fly. But again, the principles are very similar to the 737. Or to the Cessna 172 for that matter.
in the lifetime of the 737 it is reliability, engine tech and avionics that have changed. Not the mechanics of flying the wing.
The Boeing study includes a list of hull loss rate by type. The measure is important. It is the accident rate per 1,000,000 departures. The first number is hull loss with fatalities, the second is all hull losses.
The very oldest type of 737 had a fatal rate of 0.89 per million departures and a hull loss rate of 1.75 per million. The CFM-engine variant tumbled to 0.26 fatals per million departures with a hull loss of 0.54. The NG (the most numerous 737) tumbled still further to 0.11 fatal accidents per million departures and a loss rate of 0.27.
What leaps out is that the 737 has become generationally better despite the fact that it is essentially the same aircraft. That is quite strange. Or at least it would be quite strange if you were to take the Airbus logic position of the flight control system being responsible for this.
We need to keep things in perspective: both manufacturers produce astonishingly safe aircraft. The A320 family has a loss rate of 0.14 fatal accidents per million departures and a total loss rate of 0.24. These numbers are staggering. Find the ‘out of production’ bar that includes old, smokey, scary-but-cool jetliners. This is an order of magnitude different from today’s airliners.
So there is Airbus logic and normal logic. The fact that Airbus has even chosen to cut the data into ‘our jets’ and ‘theirs’ makes me a bit disappointed. It is an unscientific approach by a company which prides itself on technical knowhow. It seeks to misrepresent the data. Poor show.
The 737 has enjoyed a long and tough career of continuous development. Early 737s had very basic navigation capabilities, a very simple autoflight system and that was about it. If you have flown a fairly old light twin the 737-200 was just a rocket-powered version of that, scaled up a bit.
Navigation kit improved over the years, as did automation. The mid 80s meant a new engine breathing new life into the older 737. We would call it a Neo or something fancy now. They just called it the 300 back then. Simpler times.
Eventually EFIS arrived and with it a step change in navigation kit and capability. This brought with it some automation complexity and a bit of ‘what’s it doing now?’. More on that in Chapter 3. The A320 family arrived after the 737 became turbofan and EFIS.
All A320s, on the other hand, are delivered with a beautiful digital autopilot, Cat 3B capability and the benefit of lots of ergonomic and procedural design knowledge. It stood on the shoulders of the designs that came before it.
The arrival of the 737NG really was not a huge leap over and above the technology of the 737-300/400/500. So similar was it that I remember the CAA just adding it to my ATPL at the time. Sure, it is far better than its older siblings. Stronger, faster, longer, cooler.
But it is not a lot more technically capable. So I was surprised to see that the 300 to 500 has almost double the accident rate of the NG. I stared long and hard and then I had an epiphany. The NG was introduced at about the same time as EGPWS.
EGPWS is the fancy version of plain old GPWS which warns pilots that they are getting too close to the ground. Old GPWS just looked straight down. It relied on instrumentation to do this.
It saved loads of airframes and lives but sometimes the limitations of the system meant that the ‘pull up manoeuvre’ had to be flown immediately and flawlessly. Or perhaps that it just was not possible at all. Here is the Cali accident to illustrate the point.
EGPWS layers a database and navigation position on top of this system to give some ‘look ahead’ capability. This stretched out warning times from a number of seconds right up to a minute or so. The requirement to fly the manoeuvre when a warning goes off is exactly the same but an enormous increase in safety margin is possible just by looking into the distance a little.
Honeywell who manufacture the system have a simple stat: since its introduction not a single aircraft has been lost to terrain impact if the warning was followed. Impressive.
The NG is a post-EGPWS machine. The A320 family spans a couple of generations. From my slightly unscientific canter through the data it seems that the era, not the design is more significant.
The majority of A320s have been constructed post-EGPWS too. The big bulge in production starts alongside the opening up of Low Cost shorthaul, a mobilising Asia/Pacific region and a post 9/11 industry.
Most A320 pilots will never see the fly by wire ‘protections’ active outside of the simulator. We never reach those limits. They are not positioned in a way that means we are constantly bouncing off the edges of the flight envelope.
I think of the protections in the same way that I think of antilock brakes or an airbag in your car; in a crisis you will be glad of it. But I try to drive in a way that completely avoids the need to use them. These so-called protections are triage for when it has all gone wrong.
The accident statistics do not show a like-for-like difference between the baby Airbus and its competitor 737. The argument just is not made. Fly-by-wire Airbus types still have the same sorts of accidents as conventional types.
Most bafflingly I can only think of one type that has twice stalled into the surface of the earth from cruise flight recently: and that is the fly-by-wire ‘stall proof’ Airbus.
Why? My belief is that the Airbus marketing strategy has much to do with these terrible accidents. Remember that I mentioned that ‘pull means up’? That message is compelling. The training given to new Airbus pilots focuses heavily on the protection characteristics of the aircraft.
Traditional ‘knife and fork’ simplicity of understanding pitch and thrust attitudes is replaced with complex explanations of flight control system responses to failure modes. The love for the avant-garde wraps around the basic physics and science. It hides the fact that this is still just a jet airliner.
If you have spent all of your training being told how the system will help you out then you will have no reason to question it. The thousands of hours that follow will embed in you a sense of trust and expectation. Then, in your hour of most need, you are presented with not only a broken system but an aerodynamic and design concept conundrum to solve too. Confusing.
The unintended consequence of such an integrated and forward thinking system is to create something which stops being intuitive just at the moment when intuition is mostly what you have left.
I need to be clear though. Fly-by-wire and protected envelopes are a good thing. They have been around for years and long before Airbus claimed the idea as their own.
Stick nudgers, pushers, shakers, auto slat, low energy warning systems, rudder radio units, Q-feel systems: these are all 1950s tech and they all form part of an aircraft protection system. They are a sensible part of aircraft design today.
We would be surprised if we took a trip into the car showroom and listened to the salesperson explain that they had ditched seatbelts, ABS, power steering and airbags in favour of ‘traditional design goals’. With a knowing wink. The same is true of airliner design.
The Airbus flight control system was a bold move. It was complex and clever and cutting edge. It made space for a table and caused many to design their aircraft with sidesticks too. But the overall design is far from perfect and far from simple.
Airbus is now wed to the same basic type of flight control and flight deck design for the future. Commonality between types dictates that. It will forever be a legacy of a deep-thinking beating practicality. The challenge for Airbus now is to find the practicality underneath the cleverness.
That was how we ended up diverted from avionics back to flight controls: the lack of discernible difference in accident rate means that it is highly probable that the avionics, not the flight controls are responsible for the improving safety record.
Next instalment, navigation, automation and the digital jet.
The crackle of the envelope. A painfully long pause. The wrong name. The rest is news. Amidst the diamonds, the sparkle and the swish of the Oscars the red carpet gave us a peep behind the scenes at (PriceWaterhouseCooper) PwC this year. Let’s be clear; nobody died. The mix-up really doesn’t matter in the grand scheme of things but, like Hamlet, we have been entertained by a melodrama which has resonance for our flying lives. Continue reading
It’s breakfast. Switzerland. They’ve spelled my name wrong; I let it go uncorrected. It’s half eight in the morning somewhere else. My clothes don’t quite match my surroundings and my shoes don’t quite match my clothes. The February rain is likely falling as snow up on the hills but down here it’s the kind of drizzly rain that makes people walk fast and straight-legged with their shoulders drawn in. Tick-tock, the minutes slide by. Continue reading
Converting to a new aircraft is fun because it is a licence to be a learner for a while. In January I completed my training back to the Airbus A320 family of aircraft. The flying is fun but the baby Airbus is overly complicated by an obsession with unachievable precision and theory. Continue reading