Sunday, December 2, 2018

Response to AeroTime New Hub article

The following is my response to an AeorTime New Hub interview between two Air France pilots; a B777 captain and an A320 captain.

Interview: Would Air France F447 have happened with Boeing?


Of course, I covered this subject in Understanding Air France 447, (Could this have happened in a Boeing? - my conclusion was: it already has!)

Most of the comments  I agree with. Especially the emphasis on being able to fly the airplane, and not just rely on computers. Those pilot skills must be there and be able to be called upon at the most inopportune moments. Maintenance of pilot skills, like any skill, requires practice. That means hand flying the airplane when practical. Otherwise, it's like being called upon to give a surprise concert for a large audience and you haven't practiced for a year.

The comment section on the page could not contain my full remarks, which are below:

Air France 777 captain appears to misunderstand how the Airbus works. His statement "how Airbus auto trim works. As the co-pilot pulled back on the stick, the pitch trim helped him going backward to trim the aircraft because the computer acts as if Bonin* wanted the speed as low as possible... The pilots were not aware of the fact that the pitch trim was trimming all the way back, because on an Airbus you don’t have an artificial feeling, no stick shakers."

Autotrim only serves to align the elevators with the stabilizer. In Normal and Alternate laws it has NO EFFECT on how the airplane feels or behaves. In the flight control law where AF447 found itself (ALT2) and the sidestick demands g-load. As a result, it's basically a point and PROT) the stick changes from g-load demand to full back law the hard protection is replaced (due to various failures) with a "stability", which is the airplane's natural aerodynamic stability to pitch down when very slow.

However, the transition from g-load to stability mode is based on indicated airspeed. In the case of AF447, the airspeeds were rejected by the flight control computers because it had become unreliable due to the pitot tube issues. This left the airplane in full-time g-load demand.

When at least 2 of 3 airspeed sources were deemed unreliable, the autopilot disconnected. This dropped the job of hand flying the airplane into the laps of the unsuspecting and very-soon-to-by confused pilots, when they least expected it, and probably least prepared to handle it. "Startle factor" is the term used here that describes their diminished mental capacity to perform.

He said "They should have read the pitch trim indicator, which is something you forget to do on an Airbus, because it’s always in trim! They also should have had a visual warning telling them to use the pitch trim manually (Man pitch trim)."  This is not correct. At least not for the fist full minute with the stall warning going off.  All they had to do initially was to point the airplane down. It did not require any trim action. Now later on, when the angle of attack was extremely high (>45°) and the trim was full nose up, my own tests indicate that manually reducing the nose-up trim may have been required for a successful recovery from the 20,000 ft/min descent they put themselves in.

But instead of nose-down input, which would have worked, they instituted the only stall recovery action they apparently remembered: full power. Except at cruise altitude, they were pretty much already af full power and the extremely minor amount of power still left would produce no results.

Unfortunately, Bonin's initial actions on the stick (be then intentional or as a consequence of something else) were to pitch the airplane up. But the airplane's neutral stability allowed the nose to just stay there.
 It did not require trim to hold it there. That was its job. If he had pulled it back and then just let go completely, the result would have been much the same. The lack of the low-speed stability meant that as the airplane approached the stall AOA, there was nothing to pitch it down except the pilot - who was probably trying to follow the ever-reliable flight director which was telling him to continue the climb.

As the airplane began to stall and lose lift, the airplane started to sink, and thus the actual g-load was reduced. The airplane responded to this less-than-demanded g load with all it had at its disposal to increase the g load: Up Elevator. This obviously just made things worse and within a few seconds, the angle of attack increased to 45° and more (though the pitch attitude remained at 15° or less).
What this required from the start was for the pilot flying to FLY THE AIRPLANE. That is, command a pitch and power setting commensurate with level flight (2.5° & 83% N1). Any pitch and power setting remotely close would have kept the crew out of trouble for a long time. It only took about 40 seconds for all of the pitot tube issues to resolve themselves, and they could have been on their merry way like the dozens of other crews that employed that strategy in similar circumstances. However, by the time the pitot tubes were clear, the flight was in a far worse position, having climbed 3000 feet and lost over 100 knots of indicated airspeed. A recovery at this point required far greater skill than the minimal skill required to fly straight and level just 60 seconds earlier that they failed to exhibit.

All this and more, including what recovery steps were required at each step of the way are detailed in my book "Understanding Air France 447".

It is further interesting to note that Airbus seems to have taken this accident to heart in the design of the A350. It has extremely robust airspeed backups (even if all pitot tubes become unusable), and Alternate law maintains all of the same hard protections as Normal law (assuming the physical flight control surfaces are not damaged beyond carrying them out). Further, the autopilot remains engaged in an extended envelope (right up to the abnormal attitude parameters (120° bank, 50° pitch, etc) So, in this case, the AP/FD would not have disengaged at all.

A320/330/350 Captain/instructor/check airman, also B777 type rated

Friday, October 26, 2018

Soaring: Gold Distance Flight

The following article appears in the members-only section on the Soaring Society of America blog page. In Soaring badges are awarded for certain milestones. The Gold badge requires a 5 hour flight, an altitude gain of 3000 meters (9843 ft), and a 300 km (186 statute mile) cross-country flight. This is the story of my Gold badge distance flight.

The path to the Gold badge was a long time coming. The five-hour duration, common with the Silver badge was accomplished in 1994 in Minnesota. The altitude gain was accomplished at Minden Wave Camp in 2017 and also qualified for the Diamond altitude gain. It was the solo wave flight at the end of the camp and achieved over 27,000 feet with about a 19,000 foot altitude gain.

Photo #14947 | Wave flight at Minden Wave Camp 2017
The distance leg has been the most challenging. It wasn't until a few years ago that I even realized that the local pilots were regularly doing regular cross country flights at all. The 2017 Wave Camp experience really help to light my fire on this.
Back at Warner Springs started working on small local goals, extending the limits of local soaring and small tasks to practice with the flight recorder. A 2017 Silver badge attempt was foiled by a change in the rules the previous fall that measured the 50km required distance from release, instead of a designated start point. This put me short by only about 2 miles. It would be another year before conditions would be good enough to make the attempt again. In the meantime I purchased my own glider, a 1978 LS3, which provided better performance than the rental ships. I updated it with an S80 moving map vario, transponder, FLARM, and ADS-B in/out.

Preparing for the Gold distance, I spent quite a bit of time with the Badge and Record guide trying to determine just what would qualify for the 300 km gold distance. I’d had a few email exchanges with Rollin Hasness, the SSA Badge and Record guy who helped to clarify some task and declaration questions. However, sometimes you don’t know what you don’t know.
I had some gold tasks set up and made a few attempts over the summer months but the combination of the conditions and my chosen flight path prevented me from getting very far. I programmed a couple of routes that used a turnpoint twice and a 161nm triangle task to have options to pick from on the flight day. Along came a Saturday forecast with excellent conditions. This was to be the day! I shared my task file via email with my soaring buddy Dave, who would fly with me in his Discus.

We consulted that morning with Garret Willat, my official observer and trusted instructor. Should we fly to the North or South first and double back or the triangle. He suggested using the triangle but moving the southern turnpoint (Jacumba) 10 miles west due to the forecast location of the convergence line. I modified the task in my Oudie IGC, but then the whole task was also short by nearly that many miles. That meant we’d have to modify other turn points as well. Modifying the task on the recorder is not quite as easy as on my desktop computer, and we didn’t want to be last in line to launch. The gliders were lined up for takeoff and I decided to keep the original task after all and not risk messing it up. We decided a good approach might be to get as close as possible and make an out-and-back dash to Jacumba from as close as I could be within the lift line.

Photo #14942 | Overlooking the desert
Dave had trouble loading the task into his nav system so I lent him my Nano3 flight recorder with the task installed. I would act as his official observer, and off we went.

We soared to our start point ( or so I thought) over Mt Palomar. We climbed high then dived down through the start gate to have more leeway on the finish altitude.

As they often do, the forecast wasn't quite as wonderful as predicted. Once underway, I could hear some of the regular X-C pilots were having a little trouble, but I had been lucky and able to stay near the 11,000 foot cloud bases (about 8000 above the field), taking every opportunity to stay high. I wasn't going to win any speed records, but I mostly didn't want to land out. Dave took a different path and lost altitude after leaving Mt Palomar and headed back towards the airport. He as working hard to stay up, so I pressed on. So much for team flying.

I worked my way south, making a few retreats to the previous thermal to try a new path when one didn't work out. I joined up with a cross country veteran, who as it turned out would not be going as far south as I was. “How far south are you going?” he asked. “It depends on how scared I am” I answered. “Hoping for Jacumba.”

Finally, I could see Jacumba airport. It sits right on the Mexican border. As predicted, it was  out under the blue sky. I explored the edge of the lift and stepped out from under it but it didn’t look like I was going to be able to execute the out-and-back dash and arrive back in the lift as high as I wanted to be. “oh, well, maybe another day” I thought and started to head back North. Looking back I saw a cloud form to the East of the line, closer to Jacumba. This could work! I headed to the new cloud and climbed as high as I could, made the dash out to Jacumba and rounded the point, making sure my trace was inside the turn sector on the flight recorder’s path. I flew back under the cloud line and headed back north.

Shortly after another pilot reported good lift to the east of the cloud line near Mt Laguna, about 10 miles from my position. This was right along the line where the mountains descend into the desert below. In that desert was my alternate airport for this leg: Agua Caliente.Photo #14944 | Hanging out at cloud base
As I flew northward I was having marginal success with maintaining altitude which eventually dwindled down to about 8,000 feet (2000 feet above the peak of Mt Laguna). A thermal any time now would be great. I had Agua Caliente in sight and was looking at the path I would take around the intervening hills to reach it.  As I reached the peak, my luck improved and I worked a thermal back up to 11,000 feet and continued to work my way north.

Photo #14945 | Beautiful flat botto
In the meantime, and unbeknownst to me, Dave had made it to Mt Laguna and was just a little north of me. As I reached his postion, he was down at 6,000 feet below the edge of the mountains. I caught up to him and circled above, keeping an eye on him. He was within gliding distance of Agua Caliente, but it didn’t look like it from my vantage point above. I was hoping I wasn’t going to have to watch him land out. I circled above at the cloud base and eventually Dave was able to work himself back up (which qualifyied him for his Silver badge altitude gain). Stopping to thermal a few times and staying high, we flew towards the northern turn point of Mt. St. Jacinto, a 10,800’ peak just west of Palm Springs. There was another glider going in that direction who reported where he had found lift as well. We made it about 1500 feet above the peak. Turning southward again, Dave elected to take a detour to Mt Toro to the southeast, while I headed to my finish point on My Palomar.

I arrived at Mt Palomar with plenty of altitude and circled my finish point, just to be sure, then made a high speed run to home, with some 360’s south of the airport to loose energy. About five hours since takeoff I landed happy to have completed my Gold distance.  Another pilot informed me that that 300km triangle would also qualify for the Diamond Goal flight. That was an exciting prospect, a second diamond and gold badge in one!

Post flight, Garret and I finished the paperwork for the badge application.  While Dave had not made the Gold distance task as planned, his reaching Mt Laguna, along with his climb back from his low point plus more than five hour duration since release allowed him to complete all three legs of the Silver Badge in one flight!
Photo #14943 | View of Warner Springs from over Julian
Upon returning home and going over the flight again, I discovered that in constructing the 300km task (converted to 161 nautical miles) that the correct conversion was 161.9nm. The task was short by less than a mile. I let Garret know that we’ll chalk this one up to lessons learned and try again. The next day I realized that the Badge and Record guide had provided another option: the post-flight declared finish point. The task was a loss for the Diamond Goal but for the Gold Badge we were back in business and sent in the files.

Rollin informed us that the application was unsuccessful as I hadn’t crossed the start line and the distance was too short. I had programmed a turnpoint segment as a start geometry as  the Badge and Record Guide had allowed, but that part was out of date. I flew past the start point, but I was not within ½ km of the start waypoint to be within the required 1km start line. Ouch, another unexpected lesson.  But how could the distance be short, even with the post-flight declared finish? There was even extra distance now. Rollin discovered that when we transcribed the task to the application in order to reflect the post-flight finish point, we were off by exactly 1 degree of latitude on the Jacumba turn point (33N instead of 32N), which made the task look like it was about 120 miles too short! Garret confirmed that indeed the intended point was as stored in the task in the IGC file. While the declared start point wasn’t valid, I could use the release point along with the post-flight declared finish point and have the task flight meet the requirements. Good enough for the Gold distance, but that won’t work for the Diamond Goal. The task was accepted, and I’m the proud holder of Gold Badge number 2765.

Photo #14946 | Gold badge with an altitude diamond
I sent Rollin some suggested corrections for the Badge and Record guide from my lessons, and he appreciated the feedback. Now, I’ve got the Diamond Goal task set up. I wonder what additional lessons that will bring.