Engines, More or Less

Most aircraft today have either one or two engines, but it certainly hasn’t always been that way. Historically, the number of engines has corresponded inversely with engine power: the less powerful the engines, the more needed. The B-36 bomber, the first intercontinental strategic bomber, had 10 engines, six huge Pratt & Whitney R-4360 Wasp Major engines, the most powerful and complicated reciprocating engines (recips) ever produced in any significant numbers, supplemented by four jet engines, two on each wing tip. As a young corporate pilot I used to fly with a captain, Richard Howland, who flew B-36’s in the Air Force. He said they almost never came back from a mission with all the engines still running, and often had more than one shut down. The B-52, the successor to the B-36, had eight turbojet engines. The eight engines were not there for redundancy, but because it needed that many to power it given the engines available at the time. .

Pilots love to argue the merits of different numbers of engines, but the fact is there is no right or wrong answer because the number depends on two variables, power required and reliability, and the two are interrelated. As a generalization, most larger general aviation aircraft have two engines because they need that many to be adequately powered. Whatever redundancy results from having more than one engine is secondary, and marginal—better than nothing (except at very slow speeds, where, unless the control response is precise, the remaining engine is often more liability than asset), but not much better. The Beech 58P Baron, a high performance, pressurized twin, is powered by two 325 hp engines, for a total of 650 hp. There are no reciprocating engines currently in production capable of producing that much combined power, hence two. (More powerful recips were made, like the Wright Cyclone shown in a previous post, Gone Wrong, and the R-4360 Wasp, but they have all been replaced by turbine engines.) The most powerful reciprocating engine still in production is the rare and exotic Continental GTSIO-520, a geared monster producing 375 hp. The more commonly available Lycoming 540, the engine that powers the pressurized single engine Piper Malibu Mirage, produces 350 hp. The Beech Baron needs 650 total hp because it weighs 4500 pounds empty and grosses out at 6200 pounds versus less than 3000 pounds empty and a gross of 4300 for the Malibu. There is no way the Baron could replace its two 325 hp engines with a single, existing reciprocating engine and perform adequately, and conversely, the only way the Malibu can carry any more weight would be to replace its single engine with twin recips (or with a turbine engine, which, when done at the factory makes it a Meridian, and when done by retrofit makes it a turbine Malibu). The Baron has two engines because it needs two engines, and the single reciprocating engine Malibu can’t carry any more than it does because it doesn’t have the power to do so.

But, here’s the catch: As engines are tuned to extract more power, they also tend to become less reliable. Common ways to extract more power are to reduce weight and to increase temperatures, often in the form of turbocharging. Turbochargers are nothing more than exhaust driven turbines, operating on the same principle as turbine engines do, except instead of using the turbine to drive a fan or prop, the turbine drives a compressor, increasing the air available for combustion. (Turbine engines also drive compressors, an essential component of their operation.) A byproduct of compression is heat. Weight reduction and heat both lead to less reliability. So to a certain extent you add a second engine not for redundancy, but to reduce the demand for power from those engines and, hopefully, make them be more reliable. So, paradoxically, part of the reason for the second engine is to ensure you never need it.

Early jets, the Boeing 707 for instance, had four engines because the early jet engines, straight turbine engines without a fan, did a better job of converting jet fuel to noise and smoke than they did power. They needed four engines to be powered adequately. But they also were tremendously more reliable than the recips they replaced. So while the four engines provided multiple redundancies, it was seldom needed. The four engines were there mainly for power, not failures.

Adding fans to jet engines not only made them more powerful, but also made them quieter and more efficient. In fact, engine designers quickly figured out that the fan was the component that had the greatest potential for increasing power, and each new generation featured bigger fans with higher bypass ratios—the ratio of air going around the hot section to the air through the hot section. A beneficial byproduct was sound reduction: the cold bypass air muffled the scream of the hot air coming from the compressors and turbines. And with increased efficiency came cleaner burns, which meant less smoke.

Among the first aircraft to benefit from fan jets were the Boeing 727, 737, and the Douglas DC-9. The 707 was also retrofitted with fan engines. The smaller aircraft, the 727, 737 and the DC-9, didn’t need four engines to be adequately powered, and started what would become a trend in airliner design continuing today: fewer and fewer engines. The 727 was, as far as I know, the first three engine aircraft since the Ford Trimotor. Larger, more powerful engines with big fans, the Rolls Royce RB211, for instance, made it possible to design large, wide body aircraft with just three engines, aircraft like the DC-10 and L-1011. Then Airbus, with the A-300, introduced a wide body aircraft with only two engines, and, while controversial at the time, established the trend for virtually all airliners thereafter.

Two things made all of this possible: power and reliability. You have to have both if you want to reduce the number of engines, and the modern, high bypass turbofan engine does that. The very largest aircraft, the 747, the C-5, and now the Airbus 380, still have four engines, for the usual reasons, but as far as I know there is no reason even larger and more powerful engines can’t be developed allowing even super sized aircraft like these to be powered with just two engines someday.

So, how safe is all of this? I have to admit that after almost 15 years of flying nothing but three engine aircraft, the Boeing 727 and the Lockheed 1011, I was skeptical about flying two engine airplanes over long stretches of water—what the FAA calls ETOPS, for Extended Twin Engine Operations. (ETOPS is also jokingly said to stand for “Engines Turn Or Passengers Swim.” It’s funny the first time.) Even though turbine engines almost never quit, I loved knowing I could lose one and still have two left, and I also loved knowing I would still have at least two generators, two engine drive hydraulic pumps, two sources of air and so on. I say, “at least” because the center engine on the 1011 had two engine driven hydraulic pumps, meaning I had the same hydraulic redundancy as a four engine aircraft as long as it was a wing engine that failed. And if the center failed (which it did once for me, see previous blog Gone Wrong) I would be down to two engine driven pumps, but that was what two engine aircraft started with. And I loved knowing that if one of those engines did fail over water it would be a huge non-event: just descend to a lower cruising altitude and continue or turn back, depending on the ETP (see previous article Equal Time Point Considerations). A little bit scarier was thinking about losing another one after that, because then you would be down to your last engine and your last set of fully functional engine driven systems, and fuel remaining could be critical if you had to cruise for any length of time on a single engine, but it could be done, and it was a whole lot less scary than thinking about losing a second engine on a two engine aircraft.

So it was with less than full enthusiasm that I transitioned from the 1011 to the Boeing 757/767. (The 757 and 767 cockpits are identical, except that you step down into the 757 and step up into the 767; the type rating allows you to fly either). But it didn’t take me long to get over my misgivings. The 757/767 has multiple system redundancies, even with just two engines: power transfer units, ram air turbines, hydraulically driven generators, an auxiliary power unit with an electric generator unit identical to that on each of the engines, and so on. But what made it most easy to accept was that not only did the engines always work, but the systems themselves always worked. The 1011 was (I should say “is,” but there are only a few left) an incredible airplane, but it was complex and fussy: little things were always going out on it, usually nothing serious because it was so well designed with so many backups, but many nuisance failures nonetheless. The trip without a “write up”—something for maintenance to fix—or the log book without a deferral—something minor that had been deferred for a short period of time, usually long enough to get it back to a maintenance hub—was rare. The Boeing, on the other hand, just didn’t have failures. It was a much simpler aircraft, much less complex, but rugged: “If it’s a Boeing, it’s going.” The 1011, on the other hand, loved the gate, as they say. Once you could get all those little problems squared away and get it in the air, it was your castle: nothing could touch you. But it would often be sitting at the gate, or on the ramp, long after all the 757s had departed.

So I came to trust, and even love, maybe not in the same way as I did the 1011, but still respect and admire, the 757. It proved itself to me by never letting me down for many hours over many miles of ocean. And the other two engine aircraft flying today, the Boeing 737 and 777 and the entire Airbus family except for the 340 and the 380, do the same thing every day. Because of that record of reliability the FAA now allows two engine aircraft to be flown for as many as 180 minutes away from a suitable airport at single engine cruising speed (and for as many as 207 minutes in the Pacific, three hours and 27 minutes), meaning they can fly virtually any route they want in the world without having to deviate to stay within range of suitable airports. (Specific ETOPs procedures apply, but mostly they just mean everything has to be working—generators, hydraulic pumps, Flight Management Systems, etc.—to be dispatched at the 180 or 207 minute limit.) The FAA is saying, in effect, we don’t expect these aircraft to lose an engine very often at all, and when one does fail we don’t expect the remaining engine to fail in the three hours or so it takes to reach land, ever. (The actual requirement is to demonstrate a shut down rate better than one in every 50,000 hours of flight.) If you can expect an engine to fail less than once in 50,000 hours, it is pretty safe to say that that engine is probably not going to fail during the three hours it takes to fly on one engine to a suitable airport.

Which raises an interesting paradox about probabilities. The odds of pitching a coin and having it come up heads 100 times in a row are astronomical—virtually impossible. But, assuming you have already tossed the coin 99 times, and it has come up heads all 99 times, the odds of pitching heads again on the next pitch are still 50/50. The past record has no bearing on the next outcome. It seems impossible, but it’s true: the fact that it is so difficult to throw 100 heads in a row would seem to tell you that that the odds are getting worse with every toss that comes up heads, but the odds on each toss are still 50/50. The coin doesn’t know. And the odds of losing the second engine are the same as the odds of losing the first: very low, but the same. If the odds were one in 50,000 of losing an engine, either engine, then having lost that engine the odds remain one in 50,000 of losing the second one too.

But wait, you say, doesn’t having two engines give you two chances to lose one? Yes, of course, but each has the same one in 50,000 chance, there are just two of them. The remaining engine doesn’t know the other has failed, any more than the coin about to be tossed knows you just threw 99 heads before it. Flying along for three hours on one engine may sound scary, but the odds that it will fail during that three hours are the same as the odds that the first one would fail in the first place: 50,000 to one, in our example. Nonetheless, human nature being what it is, my guess is that despite having flown thousands of hours in a two engine aircraft before a single engine failed, it would still be a very long and anxious three hours flying on that remaining engine, no matter how many times you told yourself that the second one was no more likely to fail than the first one was. The one that just failed.

Which brings us full circle to the question of the single engine Malibu versus the twin engine Baron—or any other single engine aircraft versus any other light twin. We said that the number of engines always revolves around two issues, power and reliability. The power issue can be solved fairly easily, as long as money is no object. We can increase the power available for the single by replacing the recip with a turbine; we don’t have to add another engine.

So power isn’t the question anymore, we just have to replace the recip with a turbine—exactly what the airline industry and the military did in the ‘60’s and ‘70’s, replacing all of their reciprocating engine aircraft with turbine powered aircraft. Which leaves reliability. And there again the answer is very simple if money is no object: replace the recips with turbines. I don’t have exact figures, but there isn’t any question that turbine engines are more reliable than recips—a lot more reliable. (They also cost a lot more, but that’s another issue that we will get to later.) The reasons are several, but the most obvious is that the turbine engine, even in its turboprop form, is many times simpler than a recip: all the pieces go around in circles, the fuel and air are dumped into a combustion chamber that has continuous, self sustaining combustion, there is no intricate valve or ignition timing involved, no carburetors or magnetos to adjust, and reciprocating motion doesn’t have to be converted to rotary motion. The one thing a turbine engine has to do that a reciprocating engine doesn’t is sustain very high temperatures. That turns out to be a pretty simple problem to solve, though, compared to those for a recip anyway, it just takes money. The metals necessary to withstand those temperatures are very, very expensive. Which is why we still have recips.

For an easy way to see the difference between recip reliability and turbine, compare times before overhaul (TBO) for each. Reciprocating engines typically have recommended TBOs between 1200 and 2000 hours, and require regular maintenance in between, usually on a 100 hour in service schedule, and still often don’t make it to TBO. Turbines typically have TBOs between 3000 and 4000 hours, with minimal maintenance in between, the only major service being a hot section inspection at the mid point for cracks. And they almost always make it to TBO. Turbines are also lighter (for a given amount of power) than a comparable recip, have less vibration and are often quieter. Finally, they burn jet fuel, which is not only cheaper than avgas, but much more readily available (and always will be, whereas avgas, while not an endangered species, is a very small part of refinery production and is no longer universally available). So for all kinds of reasons, turbines are the way to go, if you can just get past that initial cost. (Recips do have one advantage over turbines, and that is that they burn less fuel, but not much less, and that fuel still costs more, so the result is close to a wash.)

So we’d all love to have a turbine engine in our general aviation single or twin, but, of course, money is a factor, and replacing a recip with a turbine can be a prohibitive expense. Assuming we are staying with reciprocating engines, what about this single versus multi thing? To answer that question intelligently, I think we have to start with an acceptance of the relative lack of reliability of reciprocating engines: we may be able to show that turbine engine failures are so rare that we can assume we will never have two fail in a three hour period, but we can’t assume that same degree of reliability with recips. Recips will fail, with much greater regularity than turbines do, and we have to take that into account.

So does that mean that if we are going to fly with recips that we have to have two of them? What about the accident rate for multiengine aircraft, and their miserable performance on one engine? Aren’t we sometimes better off in a single, even if we do accept the fact that it could quit at any time and leave us with no alternative to an emergency landing?

And I think the answer is, “Yes,” sometimes we are better off in a single, or more specifically, some pilots are better off in a single, and some pilots are better off in a twin, depending on their training, experience, proficiency, and currency. A well trained, experienced, proficient, and, perhaps most importantly, current pilot—meaning he or she flies a lot and at regular intervals—will always be better off in a twin than in a single. The single engine performance may be very marginal, even close to nil shortly after takeoff, but it will still be something, and something is always better than nothing. But without good training, and a lot of flying experience in general, and without proficiency at keeping a twin engine aircraft upright on one engine and currency in flying that aircraft, he or she is better off taking his or her chances in the single. Those chances can be improved considerably by careful planning and good judgment: using airports with multiple emergency landing sites, using the longest runways at those airports, and, once airborne, keeping a continuous tally of suitable airports within range and adequate emergency landing sites when out of range, flying around mountainous areas and large bodies of water, flying as high as is practical to increase gliding range, avoiding areas with low ceilings, all are ways to increase your chances after an engine failure, meaning your chances could be pretty good. If you can’t handle a twin engine airplane on one engine really, really well, you’re better off taking your chances on your gliding skills than on your engine out skills. Because when you lose an engine on a single engine airplane, you have one very simple task ahead of you and that is to find some place to land. But when you lose an engine on a twin engine airplane, you have a beast with a mind of its own that will turn on you unless you do something to control it. Single engine safety is only as good as the planning and judgment that goes into it, and multiengine safety is only as good as the pilot flying it. And that’s the long and short of it, more or less.