Piloting airliner before flight computers and autopilots

I’ve been watching a bunch of videos about airliner curiosities, terminology, practices, and such lately. The number of issues* and factors** pilots have to consider is mind-boggling. For example, weight (both overall and, possibly, it’s distribution throughout the frame) changes as you burn through fuel; the effect of engine thrust varies with altitude and temperature; lift varies with altitude, temperature, and flight angle; landing and takeoff distances vary based on a ton of factors; you get the idea.

So, how did pilots manage before flight computers and autopilots? Did they just fly with much wider margins?

  • i.e. tasks to perform/contingincies to plan for
    ** i.e. stuff to think about when considering the aforementioned issues

Calling LSLGuy!

Well, there used to be more people up there.

The flight engineer and the navigator were, at one time, special roles in commercial flights, taking care of complex mechanical systems and navigation, respectively, before computers took over those jobs.

Flight engineer

Air navigation:

I’m reminded of an old joke: One day, the entire aircrew is going to be reduced to a pilot and a dog; the pilot will fly the plane, and the dog will bite the pilot if he so much as thinks of actually touching any of the controls.

Autopilots are actually a fairly old technology:


Behold he E6B flight computer, a.k.a. “whiz wheel”, a circular slide rule used by pilots from the 1940’s onward for flight calculations both during pre-flight planning and, when needed, while enroute.

I feel like I should be a good person to answer this stuff because the fleet I fly ranges from aircraft with autopilots that basically don’t work (BAe146), to aircraft with quite modern autopilots and systems (Avro RJ). We also use a range of old fashioned support systems on the ground and some more modern computer assisted systems. I’m not sure where to start though so I’ll tackle the examples in your post and see where it leads.

The specific issues you’ve listed aren’t as big a deal as you might think and the use of autopilot and flight management computers* don’t make a huge difference. A lot of data that is used by flight management computers was just presented as tables in books in the olden days (and still is for pilots like me who fly old tech). The autopilot is just another pilot really. It’s there to reduce your workload but it’s not doing anything that the meat pilots couldn’t easily be doing.

  1. Weight pt 1. Pre-flight, this is taken care of by load controllers and flight planning people. Load control tells flight planing what the expected payload is, flight planing plans the flight and tells load control what the fuel uplift and burn will be, and load control add the fuel figures to their data and produce a load and trim sheet that has passengers and cargo correctly positioned so that no weight limits are exceeded for the flight. Load control and flight planning may be using computers or paper or a combination.

The load/trim sheet can be automated to some extent or may be manual. I work for a contract company and on some contracts the pilots are still doing manual trim sheets, it takes a few minutes.

A manual load / trim sheet has a graph depicting the aircraft centre of gravity limits and you manually plot the position of the CofG for various stages of the flight and make sure the plot is inside the box.

  1. Weight pt 2. With the aircraft correctly loaded on the ground there is not much for the pilot to do in flight other than fly. Large aircraft with complicated fuel systems may need to have fuel moved around to keep weight within limits, but this is generally taken care of with simple one-size-fits-all procedures such as burning fuel from tanks in a particular order. LSLGuy can provide more detail on this as my 100 seat RJ is very simple and we don’t have to think about weight distribution in flight at all.

As far as handling goes, it’s a bit like driving your car with different loads and different amounts of fuel. The handling will be different with four adults and full fuel compared to just a driver and a nearly empty tank, but you don’t really concern yourself too much with it. You unconsciously adjust your driving to account for it. Flying is the same. The handling will change throughout the flight but you just do what you need to do to get the result you want.

  1. Changing engine and aerodynamic performance. A bit like changes in handling, you don’t really think about it, you just do what you need to do to get the desired result. At high altitude you need close to max thrust to maintain cruise speed, at low altitude you need far less. You don’t need to give it whole lot of thought though, you just set whatever thrust you need to get the speed you want while taking care not to put any engine indications out of limits. When you drive your car uphill you know that you will need more power to maintain speed but you don’t analyse it too much you simply adjust the accelerator pedal to get the speed you want. If your foot touches the floor then you change down a gear while being careful not to put the engine rpm into the red arc.

There are certain limits on flight and the flight management computer (FMC) and other modern systems can be helpful in displaying what those limits are but there are other methods to achieve the same thing. While an FMC can tell you if an altitude is achievable, a simple tabulated chart of weight vs temperature can do the same for a non FMC aircraft. A modern speed tape will have dynamic low speed and high speed limits shown while in an older aircraft you may just need to know not to fly slower than a certain indicated speed while above a certain height.

When flying the BAe146, which doesn’t have much in the way of technology, we have a card in the flight deck displaying our current weight to the nearest 1000 kg. On the card are a list of important speeds such as take off and approach speeds for various flap settings. One of those speeds is “Ver”, the minimum clean speed (i.e., flaps up). When we are above 27,000’ we just need to know not to fly slower than Ver plus 20 knots.

  1. Landing and take-off data. Landing data is simple enough and is presented as tables in a book or the electronic equivalent. Take-off data is more involved. For a long time this has been taken care of by performance specialists. You are right that it is complicated and pilots are not generally trained to be able to do it. This is not a case of technology vs no technology, it’s more that it is a very involved process that must be carried out for every runway that you could possibly be operating from and must account for every likely combination of weight, temperature, wind, and atmospheric pressure. It’s not something you can just do in a few minutes the way a load and trim sheet is. From a pilot’s perspective you are either presented with tables in a book or an electronic version of the same thing. The result is the same for us, find the ambient conditions and look them up in the relevant table or enter them into the EFB. From there you’ll be given a max take-off weight or you may wish to work out a minimum thrust setting to use, they are both closely related.

The presence of technology in the cockpit doesn’t make a big difference with take-off data. An EFB setup can streamline the process and reduce errors though it does have its own disadvantages.

So, to sum up, flight management computers and autopilots are useful bits of kit but they are not game changes for the specific problems you mention. On the other hand, FMCs and the way they work as a navigation computer have made a huge difference to navigation. Without good nav systems and associated flight management computers, the pilot’s job would be much more involved (except it wouldn’t, because that work would be done by the navigator or flight engineer).

  • I assume you mean flight management computers and not fly-by-wire which is far from universal.

I still carry one actually. A more compact version produced by Jeppesen. Never used it in anger but something makes me very reluctant to remove it from the bag.

Before autopilot systems were widely installed, pilots “hand flew” the airplane through all phases of flight. Ernest K. Gann’s Fate Is The Hunter does an excellent job (IMO) of illustrating what commercial & airline piloting was like in the early days, and has a passage that speaks specifically about hand-flying an aircraft (DC-3, IIRC) to exacting altitude & airspeed standards in spite of distraction, both natural and manmade.

Fun aviation trivia:
Lawrence Sperry, the aforementioned inventor of the autopilot, is also generally acknowledged as the founding member of the “Mile High Club”. Makes you wonder which one ol’ Larry thought of first - the autopilot, or the in-flight nookie…

Obviously the autopilot. You want something to keep it up when you’re getting it up.

As early as 1924 markers were set out to guide air mail pilots. They includedbeacons for nighttime and giant arrows on the ground for daytime.

As **Richard **said so well …

The administrivia gets detailed. But all that is done while the airplane is parked. In the Olden Dayes somebody sat down with what looked like a tax form and a pencil and totted everything up then looked up some limits in a big book of tables to derive the answers. Nowadays there’s a computer program at HQ that does all the totaling and table looking up. Which we get as a printout, or in the latest iteration, a download to our iPad or datalinked directly into the airplane. We still have the forms and procedures to do the whole shebang manually if there’s a big computer outage.

The tolerances haven’t changed a bit. Just how much human sweat goes into adding and subtracting.
As to flying, it’s (unsurprisingly) as **Richard **said. You do whatever it takes to make the jet go where you want it to. And we still use a combo of new and old methods when it comes to verifying navigational or limitational stuff.

E.g. about half-way up every one of our climbs after takeoff, we check the computer to ensure it thinks we are light enough to climb to the planned cruising altitude. As a separate matter we consult our ground paperwork to determine our zero fuel weight, add the weight of fuel shown at the moment, and consult a sheet of cardstock stored in the cockpit which lists the maximum allowable weight for various combinations of altitude, speed, and degree of turbulence. Both should agree that we’re capable of making the planned cruise altitude by X margin of weight. Knowing X’s value is actually the most useful outcome: that “slack weight” is a real good proxy for how much excess performance will be available and how quickly, or not, any encounter with turbulence or mountain wave would become critical.

Both the computer and the people recheck this stuff periodically as the fuel weight burns off. It’s frequently advantageous to climb as the flight wears on. But you want to retain a reasonable X margin.

etc., for another 20 factors. The computer continuously figures out lots of nice-to-know info. But very little of that is trully need to know all the time. We can estimate or compute the need to know stuff real quickly.

There may come a time when pilots have lost how to do this themselves and if HAL doesn’t know then nobody will know. IMO that won’t be progress.

The seat of their pants was as wide as necessary.

And you’ve wasted your time if you drift below a mile high, you’d have to do it all over again.

There’s the story I read when China was first modernizing and thinking of replacing their Russian airliners with some Boeing back in the 80’s. The engineers from Boeing were trying to understand how the airline currently calculated takeoff capability at, for example, Lhasa where the airport is about 11,000 ft high. After some back-and-forth they finally realized what the airline pilots were saying was: “we start the runway roll, and if we pass that marker there and are still not airborne, we jam on the brakes and go back and unload some passengers and luggage and try again.”

In other words, “Just fly the plane!”

Works for me, sort of, and hopefully will work a lot better as I get better at it. Whenever I ask my instructors for more details instruction on various techniques (like, how to steer down a runway in a crosswind), one of my instructors typically says “Just fly the plane!”

I’ve got a pilot friend that I go flying with now and then (in small GA airplanes or gliders). We recently talked about the prospect of having self-flying airplanes, in the same sense as the current notions of a self-driving car. There was just recently something in the news about Microsoft working on such a thing. (ETA: Added linky.)

One of the points was, that some people fly little GA planes because, y’know, it’s fun to do (the proverbial “$100 hamburger” being just a pretext to go flying). What will it be like when there’s a whole generation of “pilots” who only know how to push the “Start” button but have never flown stick and rudder? Where’s the fun in that?

All those millions of things the pilot has to be thinking about:

One instructor told me this little story once: (And note, this is just about flying a little GA plane, let alone an airliner):

One of his students made a list of everything a pilot needs to be thinking about while landing an airplane. (And note, this is just about flying a little GA plane at a GA airport, let alone an airliner.) He came up with a list of something like 130-some things the pilot has to keep in mind.

There’s an apparently well-known phenomenon among student pilots and their instructors: Vastly many beginning student pilots are phobic about learning to use the radio. It seems that there’s so much for the newbie pilot to think about, it fully occupies all their mental bandwidth, and talking on the radio (with its somewhat contrived and scripted artificial language) alone takes up a shitload of mental bandwidth (for a newbie). Taken all together, it’s just too much mental overload for a beginning pilot to fly the plane during the pattern and landing, and talk on the radio at the same time.

Thank you all for the information. It sounds like flying is/was a lot less precise without flight computers. I was thinking each control has to be set to a specific value based on conditions. But y’all are saying you calculate an estimated value then adjust it as needed to get the result you want. Also, a lot of the stuff is just efficiency-maximizing gravy.

BY “control” I’m referring to: throttle, offsets for the manual controls (so a neutral flight yoke position = level), weight adjustment unit*, and the like (I’m wanting to think some of the wing surfaces like flaps and slats are used beyond takeoff and landing but I know that’s not true). One example (I just thought of), how do you handle the situation on a dual engine plane if one engine is performing slightly better or worse? Do you manually adjust individual engine throttle (I’d think that would get tricky when you go to adjust overall throttle)?

*I made that one up.

I was hoping you’d come back to explain what you meant. Now we can give you something closer to what you meant.

In your car when you drive do you hold the wheel exactly straight? Or do you gently zig and zag slightly turning the steering wheel as necessary to keep the car in the lane? How do you maintain speed? By selecting a foot position or by looking at the speedo and gently adjusting foot position to hold, say, 60mph regardless of changing wind or slope or surface?

We do exactly the same thing. There’s a few things that are carefully pre-calculated and pre-set. But after that everything is adjusted as needed to maintain the target. Whether that target is a speed, an altitude, a course, or a point of impact with a runway.
Your example of two slightly different engines is well-chosen. That happens every day. Have you ever driven a car that pulls to the left or right? Pain in the ass, but you cope. You hold in a little left steering to keep your car going straight despite the car’s desire to turn right. You’re stuck; there’s nothing you can adjust while driving to correct that pulling tendency; you just need to fight it until you can get it fixed at the shop.

In airplanes there are 4 axes of control: roll (bank left/right), pitch (nose up/down), and yaw (swivel left/right). Plus total engine output power.

In small airplanes there is an adjustment, called “trim”, to fine tune control in the pitch axis so the airplane nose will neither move up nor down at the current combination of power, weight, speed, and altitude. Works good. For the other two axes, any imbalance in the airplane and the forces acting on it is just there. The airplane wants to pull left or right and your job is to sit there and hold the controls a little opposite as needed to keep everything going where you want.

In bigger airplanes there is “trim” on all three directional axes. And we’re continually tweaking one or another so the thing flies true. You start out by pushing the two throttles forward together. When you get close to the correct power setting (think like RPM in your car) you diddle each throttle individually until they both/all read the correct number. Then away you go. Almost inevitably one will still be putting out more power than the other. So you offset that with a smidgen of rudder input while things are dynamic and once you’ve settled down to a more steady state, adjust out the asymmetry with some combo of moving the throttles differentially, and adjusting the various trims.

And yes, everything you adjust affects everything else. A tweak of power triggers a need for rudder & pitch trim, and maybe even roll trim. Over time as the machinery gets cold at altitude and weight burns off it all needs adjusting again. And again. And again.

Each and every airplane is slightly out of perfect adjustment of all the moving parts. They’re even slightly bent here and there. Just like all cars’ wheels are slightly misaligned. The older they get, the crookeder they fly. So we tweak the adjustments to fly as straight as possible and write up maintenance discrepanacies if one is excessively crooked.

Aka: “Stick and rudder as necessary”

Still not sure what you mean by “flight computers”. Airliners have a flight management computer which is basically a navigator. It has the flight route loaded into it and using a number of different navigation sensors (radio beacons, GPS, IRS etc) it calculates where the aircraft is and sends inputs to the autopilot to tell the autopilot how to steer to keep on the nominated track. It also has performance data and fuel inputs that allow it to calculate things such as how much fuel you’ll have at the destination, and how high / fast you should fly to balance fuel savings against staying on schedule.

There are a raft of other small computers on board that look after various things such as thrust settings and pressurisation.

Some airliners have a fly by wire system that consists of a computer between the pilot’s flight controls and the control surfaces. It gives the illusion that the pilot is directly flying the aircraft but if the pilot tried to bank or pitch too much the FBW wouldn’t allow it. This is still relatively rare on Boeings, I think the B777 and B787 have it, but earlier types don’t. Airbus from the A320 onwards have FBW. Is this what you mean by flight computers?

A small amount of imbalance between engines doesn’t do much other than create an annoying dissonance in the combined engine sound. You get a “beat” similar to playing two slightly out of tune guitar strings.

That said, normally a thrust computer of some sort sits between the throttles and the engines and can smooth out variations in engine performance so that with thrust levers lined up, the engine thrust is the same on each engine. If this is turned off and the engines are mismatched you may get some thrust lever stagger if you manually match the engines.

While cruising you don’t normally have much reason to adjust the thrust so if you do have thrust computers off or not fitted, you may adjust the thrust levers with some stagger to try match the engines. Once you are flying the approach though, it is easiest to just move them together–the affect on handling is not noticeable.

That is so cool.

It seems to me from the above, that the biggest problem for early aviators was navigation; especially in the dark. With no satellites flying a compass bearing was pretty inaccurate. If there was a navigator (as in military planes) then they could use the stars to navigate by. I watched a film about an air strike during WW2 where the bombers flew from Australia to Japan and back; much of the 15-hour flight in darkness and bad weather. Record-breaking in its day.

I have also read about the brave people, many female, who ferried plane of all types of planes from Spitfires to Lancaster bombers from the factories to the bases where they were to be used. They flew alone, with no weapons to defend themselves and they were flying at night over a blacked out country.