Upset Training Done in Flight

 - March 8, 2016, 4:28 AM
Flight Research conducts its training flights in military airspace and near Mt. Whitney. The Impala shown here is one of the company's 45-aircraft strong fleet. (Photo: Flight Research)

There we were, flying along at 15,000 feet and fully stalling a Sabreliner.

This isn’t a war story told over a pint of beer at the local watering hole; it’s just one of many maneuvers that I flew during Flight Research’s Upset Recognition and Recovery Training (URRT) course, not only in the company’s Sabreliner 60 but also even more radical maneuvers in an Aermacchi MB-326 Impala single-turbojet jet trainer.

I attended the two-day URRT course in January, held at Flight Research’s Mojave Air and Space Port headquarters in the southern California desert not far from Edwards Air Force Base. The course was not only an excellent introduction and refresher on URRT but also a chance, albeit in a small way, to feel like I was part of the Mojave test flying quasi-military culture for a few days. Flight Research offers URRT in two-, three- and four-day courses, and it recommends initial students attend the four-day course, then return for recurrent training every two years.

My classmates were mostly current corporate pilots, attending either for the first time or for refresher training. One attendee was a flight attendant who asked to join her pilot colleague for the class, hoping that she would learn more about what goes on in the cockpit, and how an upset might affect her safety duties in the cabin.

There are a variety of upset-training opportunities out there, from courses in light aerobatic airplanes to military jets such as the L-39 and simulator-based training in larger business and regional airline-type jets. Now that FlightSafety International has cracked the nut on FAA certification of simulators aerodynamically programmed to replicate stalls outside the normal flight envelope, high-speed upsets and other maneuvers, loss-of-control training will likely become part of every pilot’s training regimen. And whether it’s done in a properly programmed simulator, an aerobatic light airplane or Flight Research’s Sabreliner and Impala, training to recognize and successfully recover from an upset and prevent loss of control is essential, and the more opportunities for training, the better.

Loss of control remains the top cause of fatal accidents across the board, from the largest airliners to the smallest light piston singles. That the aviation industry needs to work on this problem is abundantly clear. “We’re concerned about the growth in loss-of-control accidents,” said Flight Research chairman and CEO William Korner. “We feel this is a growing trend. If we continue to train pilots solely focused on automation, loss-of-control accidents will continue to occur.”

Korner and the Flight Research test pilots and instructors believe that in-aircraft training is superior for helping pilots learn how to recognize and recover from upsets and prevent loss of control. “The simulator cannot replicate the emotions of loss of control in flight,” he said. “The real airplane handles differently, and we use aircraft that have control feel and loadings similar to the airplanes you fly.”

Welcome to Mojave

The class began bright and early on the first day, with a 7 a.m. muster at Flight Research’s main building, where the URRT class is conducted, as well as classes for the National Test Pilot School, the world’s only civilian test pilot training organization, which Flight Research owns.

During breakfast, Flight Research chief Korner introduced himself and our instructors and then we split into different groups. Our group headed to the classroom for a morning of loss-of-control aerodynamics, and the other group donned their flight suits and boots to learn how to operate the Impala’s Martin-Baker ejection seat. Some of the group would fly first, then attend the classroom session; I was scheduled to fly the Sabreliner on the first afternoon, then the Impala the following day.

The logistics of scheduling everyone to fly two different airplanes and attend two different classroom sessions are challenging, so not all of the events happened in what one might regard to be the correct sequence, but that is the nature of this kind of training. It would be far more costly to run the training one student at a time, and it wouldn’t make sense to put everyone through the classroom at the same time, then sit around and wait to fly.

Our aerodynamics class was taught by Doug Dodson, a retired Air Force test pilot and test pilot school instructor who holds a master’s degree in aerospace engineering. An active pilot, Dodson built a Glasair II and flies that and a Twin Comanche in his spare time.

Dodson explained that this class was originally developed for test pilots, and this was evident when looking through the fat binder in front of each student. Thankfully, we didn’t have to go through every chart and diagram or dig too deeply into the accompanying mathematical formulas. He said this portion used to be a much longer session, but Flight Research has cut it shorter to match the needs of students and eliminate unneeded information.

 The goal for the training is not to turn us into envelope-busting test pilots but to help us learn how to return the airplane to a safe part of the normal flight envelope if we get close to the edge. Another important part of the training is to help alleviate the startle factor, which comes into play during an upset. “We can alleviate that a little by giving you some exposure,” Dodson said. 

Although Flight Research is all about using real airplanes for upset training, there is a new push to program full flight simulators to replicate extreme corners of the flight envelope such as full stalls, high-speed upsets and other bad neighborhoods.  “Things are changing,” he said, “and they are developing models for simulators. But you can’t do 1.4-g turns for three minutes in a simulator [and feel the acceleration].” The simulator also “is not startling,” he added.

We spent the morning reviewing the flight envelope and V-N (velocity versus load factor) diagram, something that is worth doing regularly, and stall characteristics, high-speed aerodynamics, stability and (the meat of the entire course) Flight Research’s recovery technique.

The stall discussion was especially important because the FAA finally changed its stall recovery practical test standards (PTS) in 2012. The previous standard, in place since 2006, required pilots to initiate recovery at the first indication of a stall by adding maximum power and recovering with a minimal loss of altitude and airspeed.

The new PTS focused on the proper way to recover from a stall, by unloading the wings by reducing the angle of attack. “Today and tomorrow you’re going to do real stalls,” Dodson said. “We’re going to stay in the stall and look at it.”

Another important part of the training was a discussion of the use of rudder in bank angle excursions, and how it is critical not to use the rudder to try to generate a roll rate to control bank angle, especially in transport-category aircraft. As we would learn and practice, lowering the angle of attack quickly improves roll authority and responsiveness. And reducing the g load has the beneficial effect of helping prevent stalls, because at zero g it is impossible to stall. There was much more to this segment of the training, but like most pilots, I was eager to get going on the flying portion of the training.


An upset is defined as more than 25 degrees pitch up or 10 degrees down or a bank angle greater than 45 degrees or an inappropriate airspeed. The startle factor, Dodson explained, is sort of like the five stages of grief: denial, anger, bargaining, depression and acceptance. But pilots must quickly move through that into the correct recovery.

Flight Research’s upset recognition and recovery technique centers on two key processes.

The first is what any pilot needs to do in any emergency: RCA, for Recognize the situation; Confirm what is happening; and turn off Automation, particularly the autopilot. This can happen fairly quickly, especially when the airplane is beyond upside down, but it is important not to rush the process and exacerbate the predicament.

The UTAP steps are fundamental for all upset recoveries:

  • Unload–reduce angle of attack
  • Throttles–as required
  • Ailerons–bank to wings level
  • Pitch–reestablish level flight

Dodson went over a number of ways that we could use UTAP for upset recoveries and showed us videos of Sabreliner and Impala upsets and how UTAP works in every case. To answer the frequent question of whether a pilot should continue rolling when in an inverted upset and whether a split-S maneuver makes sense, the Impala video showed what happens in each case.

In an upset where the Impala rolls inverted at 11,500 feet, the airplane loses 3,000 feet doing a split-S recovery, in which the pilot pulls aft on the stick to pull the nose down toward the ground then levels off without rolling the wings. Continuing the roll in the direction of the roll results in an altitude loss of 1,000 feet. Finally, performing the UTAP process and stopping the roll and reversing it results in losing just 500 feet. “You almost never continue rolling on over in our airplanes,” he said. “It’s better to reverse.”

Flying UTAP

While the classroom time is important, the flying serves to help cement what we had learned thus far. Because of the logistics of getting everyone their flights and imparting the classroom training, our group didn’t cover the module on upset-related accidents until after we did both the Sabreliner and Impala flights.

Flight Research’s commitment to running a safe operation was clearly evident in the preflight briefings conducted by our instructor pilots. Flight Research operations manager Scott Glaser was my instructor for the Sabreliner flight. A flight-test engineer and warbird pilot, Glaser has worked on F-22 flight control systems, the 747SP modified as the Stratospheric Observatory for Infrared Astronomy and Virgin Galactic’s SpaceShip Two program.

The briefing first covered details of the Sabreliner such as how the cabin door works, discrepancies, local airspace and airport operations, emergency procedures, fuel planning, positive transfer of controls, takeoff procedure, checklist usage and physiology. “You can’t train whenever you’re not feeling well,” he said. “At the earliest sign of discomfort, tell me. Not all items are required for graduation. I’m not turning you into a Sabre pilot,” he added. “This is a training flight.”

We then covered the test card for the flight, which shows each maneuver, configuration and thrust setting. The Sabreliner flight mostly involved stalls, roll handling quality checks and unusual attitudes. We wouldn’t be flipping the Sabreliner over; all flying was well within the jet’s flight envelope.

 We did fly as a crew, which makes sense because many customers are corporate pilots flying Part 25 jets. Glaser let me feel the Sabreliner’s lack of nosewheel steering, which means that brakes must be used to turn on the ground. “It’s the magic that is the Sabreliner,” he joked. During takeoff, I set and adjusted power and called airspeeds, then after gear and flaps retraction, I took over and climbed out to our working altitude above 15,000 feet while getting a feel for the airplane. I flew some warm-up maneuvers, banking up to 60 degrees and pitching to 30 degrees. “You’re not going to break her,” Glaser had said during the briefing.

We then did approach to stalls using both the old PTS method of recovering with power and the new, proper method of unloading the wings first. The Sabreliner wings are equipped with tufts, so I could look out the side window during the stall and see the airflow burbling near the wingtips. The jet also has an angle-of-attack indicator, but I found myself not looking at it that much as I was focusing more on flying the maneuvers.

In clean configuration and with the wings level, the Sabreliner began buffeting at about 115 knots, and with gear and flaps down at about 100 knots. Using the power recovery method from the old PTS, I found the controls felt very mushy, the recovery took a long time and the airplane lost more altitude. But when I used UTAP and first unloaded the wings, the controls felt much crisper and more responsive and we lost just 300 to 400 feet. We also did an autopilot-on stall, which required that I shut off the autopilot before recovering. We didn’t leave the autopilot on to the point where it exceeded limits and turned itself off, but I think that would be a good lesson to add if possible, because it’s a scenario that has occurred in a few airline and business aviation accidents.

I then flew with the stick shaker just beginning to vibrate, with wings level and with up to a 15-degree bank and while climbing at maximum available power.

The most unusual and valuable part of the Sabreliner flight was the full stall, and in this case Glaser told me to watch the tufts and to look at the AOA indicator, then to recover with pitch but not power in both clean and dirty configurations. The Sabreliner handled this easily, although one time the wing dropped and I did have to step on the rudder to bring the wing back to level. Normally I would have pushed forward on the stick to recover, then used the ailerons to level the wings, but we wanted to experience the full stall.

This was followed by a clean deep stall starting at 18,000 feet, where I held the elevator all the way aft and allowed the Sabreliner to stabilize while we observed the results. This would be like the Air France Flight 447 accident, where the pilot held the stick aft for nearly the entire upset until the Airbus crashed into the ocean. The Sabreliner stayed mostly in a wings-level attitude with a slight wing rock, rumbling a bit and at first descending at 2,000 fpm descent; but when the rate had doubled to 4,000 fpm I unloaded the wings and added power to recover. This is something that jet pilots can experience only in a controlled environment, and it definitely helped show how unloading the wing prompts a speedy recovery.

I then did some 45-degree banks where I gradually pulled the nose up to the stall to feel the ailerons while loaded up, then unloaded the wings and compared the much better feel of the ailerons to the loaded feel. Next was an accelerated stall where I pulled the yoke aft while banked until the shaker vibrated, then unloaded the wing and leveled off.

The next exercise was practicing unloading the wings by pitching up to 30 degrees and pushing forward with varying force to try to feel how much push it takes to get the g meter to show .75, .5 and .25 g. This bunt gives the student a better feel for how much force is needed for the wing unload during recovery. Then it was time for upset recoveries. Glaser would put the Sabreliner into an unusual attitude then tell me to recover by saying “fly the airplane.” It was my job to apply UTAP correctly, in nose-high banks and level maneuvers and nose-low attitudes.

This all went fine, except for one maneuver, where Glaser put the Sabreliner into a nose-high, low-speed, high-bank-angle attitude. In this case, the recovery technique is the same, but the pilot has to delay leveling the wings using the ailerons until the nose drops to the horizon. This is an important lesson, because if I had leveled the wings first, the airplane would still have been in a nose-high attitude and adding power would mush it out of the situation. Letting the nose drop while banked is a much more natural move and returns crisp control much faster. I could still unload the wings while nose high, but needed to delay leveling the wings until the nose dropped. I did this maneuver a few times because I didn’t allow the nose to drop to the horizon the first time.

I think I would have done better at this had I mentally gone over UTAP a few times before the flight, including how to respond to each situation.

After an hour, we returned to Mojave and Glaser landed the Sabreliner. My study partner then did his Sabreliner flight while I learned how the Impala’s Martin-Baker ejection seat works. Shane Martucci patiently explained the seat and went through the ejection process step by step and made sure I fully understood not only how to send it rocketing away from the Impala, even from zero airspeed and zero altitude, but also how to buckle and unbuckle the multiple harnesses and belts and leg restraints. The seat provides two ways to start the ejection sequence: a pull handle between the legs and one just above the pilot’s helmet. Martucci maintains the Impalas and their ejection seats, and he clearly knows his subject and made me feel confident about the quality of the Flight Research equipment.

After his flight, my study partner and I sat down with Glaser for a debriefing. This is an excellent part of the Flight Research training because it helped us cement the learning and also discuss any issues that arose. “This is where learning happens,” Glaser said, and he was right.

In discussing the upset recoveries, especially the nose-high banked upset where I had to allow the nose to drop before leveling the wings, Glaser pointed out, “You don’t need a heroic maneuver.” And that was evident in the Sabreliner lesson; these weren’t radical maneuvers requiring violent aerobatic recoveries.

My partner and I then put on our flight suits and helmets and spent some time in the rear seat of the Impala learning the basics about how it is set up and what controls we’d need to work with on the next day’s flight. This included fully strapping in and unstrapping, good practice for the next day where we wouldn’t want to be fumbling.

Our instructors hosted the entire group that night for a satisfying dinner, which for us was another way to feel like part of the Mojave flight-test atmosphere and learn more about all their flying adventures.

Flying the Impala

I had flown the Impala at Mojave for an upset training story two decades ago, but none of that had stuck and I definitely needed the refresher. Mostly what I remember from that flight was turning down the opportunity to experience an inverted flat spin. After an hour of rolling inverted and recovering, spinning, climbing, diving, looping and all sorts of maneuvers, my stomach was simply not up to any more gyrations. This time, I was hoping for no recurrence of those symptoms.

Our instructor was Bill Oefelein, a Navy F/A-18 pilot, graduate of the Top Gun Fighter Weapons School and Patuxent River Naval Test Pilot School and a test pilot instructor and Space Shuttle astronaut with 300 space hours logged during a trip to the Space Station.

During the preflight briefing, Oefelein outlined important items such as when I remove and reinsert the final pin that safeties the ejection seat and the words he would use if we need to eject. Because we would be pulling more g than in the Sabreliner, he explained the anti-g strain maneuver, tightening our gut and leg muscles to keep blood from pooling in our lower extremities.

Our Impala would be loaded to about 1,000 pounds below its mtow, which was just less than 10,000 pounds. The hard deck (lower limit for maneuvering) was 10,000 feet msl (7,000 to 8,000 feet agl). Weather was high overcast and relatively calm winds, although these were expected to become gusty in the afternoon. Minimum ejection altitude was 7,000 feet agl, he explained, “but only if there is no indication of recovery.”

Even though the Impala is a single-pilot jet, we would still work as a crew, with both of us going through the checklist. “This is situational awareness to help make sure I don’t miss anything,” Oefelein said.

The plan was to fly the stall series and unusual attitudes, similar to what we had done in the Sabreliner but with steeper banks. Roll upsets would be next, and these also would be much more radical. “Anything is fair game in the Impala,” he said. In the roll upsets, he said most corporate pilots tend not to be aggressive enough. “You guys have thousands of hours not being aggressive. We want to instill this muscle memory.”

After the roll upsets, Oefelein planned to demonstrate a tailslide, then we would be able to try one ourselves. “The rest is gravy,” he said, “aileron and barrel rolls, loops, Cuban 8s, if you’re up to it.” We would also do a spin.

Oefelein did the takeoff, then handed over the controls at 500 feet, and I climbed at 200 kias to 15,000 feet while banking and pitching to get a feel for the controls. The Impala is much lighter on the controls than the Sabreliner, and its military trainer heritage made me feel more like I was living out the lyrics to the Kenny Loggins signature song “Danger Zone” from the movie Top Gun. Perhaps what I was doing was not only getting to know my limitations but pushing them to new limits, as the song goes: “You’ll never say hello to you/Until you get it on the red line overload/You’ll never know what you can do/Until you get it up as high as you can go.”

OK, it wasn’t that dramatic, and Oefelein and his colleagues never did anything remotely encouraging us to cowboy around the sky and fool around. This training is serious business, although flying the Impala is a gas.

Once at 15,000 feet, we started with the stall series, and it was much easier to watch the tufts on the wing wriggle then collapse into confusion as the wing buffeted, compared to the limited view out of Sabreliner’s cockpit. This gave me more opportunities to put UTAP to use, and I felt more with the program during the flight because I had done some chair flying in between briefings earlier that morning, practicing the UTAP steps and imagining various scenarios and mentally applying the steps. That helped a lot.

Accelerated stalls were more aggressive in the Impala, banked to 60 degrees, but unloading the wing instantly stuck the airflow back onto the airfoils and set the tufts streaming straight.

My chair flying really paid off in the nose-high unusual attitudes, and I gently but firmly unloaded by pushing the stick forward then allowed the nose to drop while still banked, added power, rolled wings level, then pitched to the horizon. In the nose-low maneuver, Oefelein banked to 135 degrees before turning over the controls, and it was still important, even with the nose pointed down steeply and nearly inverted, to unload the wings first, then throttle back, level the wings and bring the nose up. “The unload is a little bit of a bunt to get better roll control,” he had reminded us during the preflight briefing.

Oefelein demonstrated a split S after rolling inverted, then it was my turn to see how this maneuver is not the best way to recover if flipped upside down. First I rolled, then pulled the stick back with the throttle at idle, the Mojave Desert ever present in the view outside the canopy and growing closer and closer as I pulled on the stick and the g load crept up toward the number four and the altimeter spun out what eventually became about 3,000 feet of altitude loss.

After climbing back up, it was time for roll upsets. We did some with flaps in the second position, and this is perfectly safe in the Impala because it has a blowback system that allows the flaps to retract automatically in case of an overspeed. The flaps-down roll upset showed how the roll rate is much lower in that configuration and how important it is to unload the wing first. We did these maneuvers at 110 knots dirty and 130 knots clean to replicate business jet speeds during final approach. “Typically this is where people stop feeling good,” Oefelein warned us in the briefing. So far, so good.

Next up was the tail slide. Oefelein demonstrated first and showed me the importance of gripping the stick with both hands and holding the rudders steady as we climbed vertically and ran out of airspeed. Not securing the controls could result in an inverted flat spin. At the top, the Impala calmly pitched over, pointing us straight down, then recovering as it picked up speed. When I tried the tail slide, I didn’t quite get to the vertical attitude, but it was fun pulling out at 150 knots and feeling three g pushing me into the seat. “This is a confidence-builder,” Oefelein said.

After he demonstrated a spin, I tried one, pulling the stick back into a normal one-g stall entry and stepping all the way on the left rudder. The Impala’s nose initially drops and starts turning, but it takes a little longer to get the airplane fully spinning. After three full turns with the desert landscape flashing by I pushed forward on the stick and stepped on the right rudder and recovered. So long to another 4,000 feet.

Oefelein said the lesson part of the flight was over and asked what I wanted to try. I opted for barrel rolls, having never had the opportunity to learn how to do them properly. He explained and demonstrated, then I tried a few, realizing that it is important to keep pulling back on the stick while rolling. For me, this was the maneuver that brought on a sudden bout of nausea, and after about four rolls I asked Oefelein to take over and fly straight and level while I took a break. After a few minutes I felt fine, and we headed back to Mojave for a simulated engine-out pattern and landing. It was bumpy coming in as the wind was gusting to 26 knots.


Our group finished in the classroom with Dodson, covering the module on loss-of-control accidents. The case studies are an important part of any upset-training program, but I would have preferred going through this module first, before any of the other classroom training or the flights. At all the other upset-training programs I’ve attended, no flying takes place until the applicable accidents are studied. I think it helps to carry that knowledge into the upset-training cockpit. But I do understand the logistical constraints.

Our training ended with a graduation ceremony and a wrap-up where we were asked to summarize our positive and negative feedback on the program. Flight Research takes feedback seriously and uses it to improve the URRT program continually. Everyone agreed that the training is valuable. Our flight attendant classmate said that she found it especially worthwhile and that she understood much more about what pilots go through and how she can help keep her passengers safe in the event of an upset.

Oefelein summarized our experiences: “We put you in no-kidding upsets. There’s a self-confidence that comes with that.”

“It makes the bag of tools more full,” said one of the other pilots in the course.