LOOKING BACK AT TURNBACKS

PETER GARRISON – I once wrote an article about a high-time Bonanza pilot who tried to return to the runway after a loss of power after takeoff. As too commonly happens in such cases, he stalled and crashed.

MUCH OF MY DISCUSSION centred on a paper about “turnbacks,” as these often ill-fated attempts are called, written in 1982 by a midshipman at the US Naval Academy, the dashingly named Brent Jett.

Jett, who went on to become a captain and a multi-mission astronaut and retired from NASA in 2013, attempted, on the basis of calculations and simulator tests using pilots of various skill levels, to determine the best technique for turning back and to understand when and why a turnback is likely to succeed or fail. It was an unorthodox project, given that the FAA’s advice, repeated by many other authorities, is to land more or less straight ahead regardless of terrain. Any landing in control, however unfavourable the ground, is thought to be better than a stall-spin.

After my article was published, David F. Rogers, Ph.D., an instructor at the Naval Academy, sent a long letter with “corrections, comments and additions” to my article, which he said, in a separate online posting, had “rather muddied the waters.” A copy of this screed tumbled out of one of my file cabinets the other day and, since the topic of turnbacks is ever fresh, I thought I should give it its belated due.

‘the dashingly named Brent Jett’

The gist of Jett’s paper, which can still be found on line (search for “Jett turnback”), was that average pilots could turn back successfully most of the time, given sufficient altitude to start with. No surprise there, although, tellingly, “sufficient altitude” was not defined. More concretely, the best bank angle to use was 45 degrees and the best speed “just above the stall.”

I wondered what “just above the stall” meant. “Since most stall warning horns are set between 5 and 10 kts above stall,” Rogers replied, “practically this means with the stall warning horn blaring or the stall warning light bright.”

I had written that Jett’s simulator turns were conducted with full rudder “in recognition of the fact that a desperate pilot might try to hurry his … turn with extra rudder.” No, Rogers said, this was not the reason; it was to bring the nose around faster. This seemed to me to be a distinction without a difference, and in any case not a very good idea. Rogers acknowledged that this technique might result in a spin.

I said that minimum sink rate, not maximum L/D ratio, was the main consideration. That was incorrect, Rogers said; the real issue was energy management. I guess I should have said that the target speed was not the “best glide speed,” as you might suppose, but instead something closer to the “minimum sink speed” (but even lower).

I wondered whether an angle of attack indicator would be helpful. Rogers objected that angle of attack indicators may give faulty indications in uncoordinated flight. Be that as it may, I cannot see how the absence of an angle of attack indicator would be preferable to the presence of one.

‘The Possible Impossible turn’

I reported that Jett’s experimental subjects had a 75% success rate with 45-degree banked turns and a 95% success rate for 30-degree banked turns on the first try. “What Garrison failed to mention,” Rogers complained, is that when pilots were given two additional opportunities to try the 45-degree turn their success rate rose above 90%. You don’t get second chances in real life, but Rogers’ point, that if pilots practiced this manoeuvre their chances of successfully completing it would be improved, was well taken.

In 1995 Rogers published a paper of his own, entitled “The Possible ‘Impossible’ Turn”, in which he studied the turnback question from a mathematical standpoint. He modelled the turn under various conditions of wind, runway length and starting altitude, and plotted results as a “footprint” of points on the ground that could be reached by the gliding aeroplane.

Given a sufficiently high starting altitude – Rogers used 650 feet for his example aircraft, a Model 33 Bonanza – to permit either a teardrop-shaped turnback and downwind arrival or a 360-degree turn, it is the wind that determines whether the aeroplane can land back on the runway. In fact, it may undershoot or overshoot, depending on wind conditions, pilot technique, and the distance from the runway at which the power loss occurs.

The recommended technique is the same as the one in the Jett paper: a 45-degree bank and a speed, for the purposes of the simulation, five percent above the banked stalling speed. Since the stalling speed in a coordinated 45-degree bank is about 1.2 times the level stalling speed, the target speed in the turn is actually 1.2 x 1.05, or about 25 percent higher than the level-flight stalling speed. Rogers recommended a teardrop turn of 210 degrees in one direction followed by a 30-degree turn in the other. As a precaution, he would have you climb initially at the best angle, not best rate of climb, speed, in order to gain as much altitude as close to the airport as possible.

‘a steeply banked turn with the stall horn blaring’

Rogers compared his results with those obtained using a different technique –270/90 turns in a 35-degree bank at 1.3 times stalling speed – recommended by another writer, John Eckalbar, in a 1992 edition of the newsletter of the American Bonanza Society. Rogers concluded that, using his recommendation rather than Eckalbar’s, “the required runway length for successful completion is reduced by approximately a factor of eight.”

Rogers took Eckalbar to mean 1.3 times the banked stalling speed, which would be nearly 1.5 times the level stalling speed. Since speed is the principal determinant of success in Rogers’ calculus, that interpretation is very prejudicial to Eckalbar’s case. I take Eckalbar to have meant 1.3 times the level stalling speed, which translates to a fat speed cushion in a 35-degree bank.

All of this is obviously quite academic, compared with the likely performance of a pilot whose engine has gone startlingly silent just as he was climbing out. In practice, fatal outcomes are not usually due to a faulty choice of glide speed or bank angle. They are due to loss of control.

The problem is the execution, not the method. The transition from a climb to a steeply banked descent at minimum flying speed is difficult to execute – particularly in a low-power, draggy aeroplane, which will lose speed quickly, and doubly so when the pilot is taken by surprise. Rogers suggested, however, that we may underestimate the number of turnback attempts that end successfully because those don’t turn up in the accident statistics. A well-trained pilot, he said, is capable of performing the technically optimal turnback manoeuvre that his analysis identified.

“Testing in Bonanzas showed the ‘impossible turn’ is possible if practiced.“

A turnback in case of a tow-rope failure, he noted, is part of every glider pilot’s training; and what is a powered aeroplane but a very low-performance glider?

The important elements here – apart from the unpredictable variations in the time and nature of an engine failure and in the pilot’s delay in understanding and reacting to it – are training and practice. Even exploring the manoeuvre is not without hazard, however, and pilots wanting to do so should begin in the company of someone familiar with upset and spin recovery.

Having experimented myself in a Cherokee Arrow, I can testify that it is much easier to achieve precise bank angles, speeds and headings in a computer than in an aeroplane.

What is missing from both Jett’s and Rogers’ treatments is clarification of the influence of errors. How much worse is a 40-degree bank than a 45-degree one? What about speed? How is the altitude from which the turnback can be effected influenced by excess speed? After all, let’s get real; a steeply banked turn with the stall horn blaring is a recipe for disaster. Furthermore, there’s the problem of running out of runway if you make a teardrop turn and land downwind, especially if you’ve gotten a little faster than you intended. What about flaps? Throttle? Propeller? All of these are bound to affect the height at which a turnback becomes possible.

A pilot who has practiced the manoeuvre at a safe altitude, who has ascertained the altitude required to perform it in his particular aeroplane, and who thinks about it before each takeoff, clearly has a better chance of executing it successfully should the need ever arise. It probably never will, but chance favours the prepared.