Flying on edge – getting down to the nitty-gritty

Margins are a basic in safer flying. Maybe that’s just another way of saying to always cut yourself a little slack, and what it means is to stay away from the edges of the envelope.

Where this often becomes critical is when the airplane is being asked to do something it either won’t do, or will just barely do. That is when precise flying is required and to use an old term, it often has to be done by the seat of your pants.

This is addressed in training with slow flight and short field takeoffs. The controls get mushy when slow and, on takeoff, the proper procedures have to be followed with the airplane in the proper configuration.

In the real world, when a pilot winds up on the feather edge of the envelope, it is usually because he has flown himself into a bad situation where the only salvation is found in a burst of brilliance plus fancy foot and hand work. At times like this, things that aren’t ordinarily considered can become very important and the time available to figure it out can be quite limited.

Pan Am 759 crash
Pan Am Flight 759 crashed on takeoff after encountering a thunderstorm.

A Pan American 727 crashed in takeoff at New Orleans in 1982. There were thunderstorms and downdrafts and wind shear present and the NTSB report delved deeply into airplane and pilot performance in such difficult situations. There is a lot there for pilots of all kinds and sizes of airplanes. It was a good example of on-edge flying that almost worked.

The Pan Am crew knew they would be challenging Mother Nature. One cell was right over the airport and another was just east of the airport, along the path they would be following after a departure on runway 10.

The flight was to Las Vegas with a full load and, as they were taxiing, the crew discussed the heavy takeoff. The first officer was the pilot flying. The flight engineer was told the takeoff would be with the packs (for air conditioning) off on engines one and three so more power would be available from them. Engine two didn’t have a pack. The FO was told to carry some extra speed after takeoff, which was standard if wind shear was expected.

Takeoff speeds were calculated as 138 knots for both V1 and Vr and 151 for V2.

There was much discussion of wind and weather on the ground control frequency as they taxied. Actual wind shear was reported and pilots had different opinions on which runway would be best to use because of the thunderstorms and the shifting gusty wind. As the crew listened to this, one of them remarked that a pilot who wanted to do something different must be looking at a windsock. For this crew in this airplane in these conditions, runway 10 was the only one available for takeoff.

While it was raining heavily at parts of the airport, there was apparently no thunder and lightning. This could have suggested that it was a tropical rain squall and not a thunderstorm. Also, the weather was stationary which also suggests a tropical condition. The tops, however, were up close to 50,000 feet so there was definitely some action there.

Cleared to go, they rolled, setting the power as high as allowable. The speed calls were made, the windshield wipers were flailing away but the rain was heavy at midfield and the visibility was limited.

Witnesses reported that the liftoff came about 7,000 feet down the 9,228-foot long runway and the airplane climbed to between 95 and 150 feet a.g.l. and then began to descend. It hit trees 2,376 feet beyond the departure end of the runway at an altitude of 50 feet a.g.l. It subsequently hit more trees and houses before finally crashing 4,610 feet beyond the departure end of the runway.

The NTSB looked carefully at the effect that heavy rain could have had on aircraft performance, but concluded that studies of this subject were incomplete. To my knowledge, they still are though there is bound to be some effect that could be difficult to quantify.

The simple fact was that the airplane had encountered a decreasing headwind and increasing tailwind as well as a downdraft and the result was a condition that equaled or exceeded the performance capability of the airplane. A 727 of this model in takeoff configuration will climb about 1,300 feet per minute if flown at Vss, the stick-shaker speed, in ideal conditions. That is slightly less than the value of the calculated downdraft. The indicated airspeed was 149 knots at first impact and Vss was 138 knots. The NTSB said that theoretically the airplane should have been able to maintain an altitude of 95 feet a.g.l. in those conditions. I guess that would assume an indicated airspeed of 138 knots.

Pan Am crash wind shear chart
The NTSB calculated the microburst conditions the Pan Am flight encountered.

At the time, flying at Vss in a wind shear condition was not a standard practice. NTSB speculated on this and gave a few reasons why it might not be a good idea. Later, flying at or near Vss became the standard for escaping a wind shear condition. The procedure I read for one jet called for maximum power, nose up until there is an intermittent onset of the stick shaker, then reduce the pitch attitude slightly and maintain it there.

The stick-shaker is a stall warning system, set to activate five knots or five-percent, whichever is greater, above the stalling speed so the wind shear escape does call for on-edge flying.

For the most part, witnesses reported that the 727’s pitch attitude appeared normal, about 12 degrees nose up, during most of the observed flight though one witness did report seeing the pitch attitude reduced to five degrees nose up for just a moment before it was returned to 12.

From the report, it is obvious that the airplane was in a critical state and one thing I wondered about related to the takeoff roll. If V1 and Vr were both 138 knots and the airplane lifted off 7,000 feet down the runway, if the takeoff had been aborted at V1, which was the same value as Vr, would they have been able to stop in the 2,228 feet of runway that was remaining? That seems unlikely to me and it seems likely that there was at least some wind effect on the takeoff roll or the crew elected to roll for some extra speed because of the potential for shear.

It was raining so hard that the crew probably didn’t have enough visual reference to know how far down the runway they were at rotation and nothing in the report suggested that the acceleration to V1/Vr had been sluggish.

Before relating this on-edge flying to light airplanes, consider the NTSB’s finding of probable cause for this accident:

The National Transportation Safety Board determines that the probable cause of the accident was the airplane’s encounter during liftoff and initial climb phase of flight with a microburst-induced wind shear which imposed a downdraft and a decreasing headwind, the effects of which the pilot would have had difficulty recognizing and reacting to in time for the airplane’s descent to be arrested before its impact with trees.

 A contributing factor was said to be the limited capability of the wind shear detection technology.

That was a controversial probable cause because it exonerated the pilot-in-command, who elected to take off in those conditions. In effect, the NTSB was calling it an Act of God and the person who drafted the final report took offense when the finding came under fire.

Takeoff by Cessna
“The takeoff is where we often ask the most from our airplanes.”

The takeoff is where we often ask the most from our airplanes. Pilots who fly off long runways don’t think much about this but almost everyone has occasion to fly into less than long runways and there are ways to do this and minimize risk.

When I had a Cherokee Six, I used what I called the rule of 3,000 when it came to airfield performance. If the runway was 3,000 feet long or less, the density altitude 3,000 feet or higher, or the takeoff weight 3,000 pounds or greater I would do the takeoff calculations to make sure there was enough margin for comfort. That was important to me in the Six because it was an airplane that could become sluggish in a hurry and while I did fly it in and out of relatively short runways, I always kept the weight well down when doing so.

One of my favorite small airports was Gaston’s (3MO), right by the White River in the mountains of north Arkansas. At the times I flew there, the turf runway was 2,200 feet long; my late good friend Jim Gaston later lengthened it to 3,200 feet. Jim did that because of takeoff accidents on his runway 24. It was always recommended that landings should be on 24 and takeoffs on runway 6 but not all pilots paid attention to that. I would not have tried 24 for takeoff in anything short of a Super Cub. The view from the end of runway 6 was daunting. Big trees, close.

I flew my P210 in and out of Gaston’s several times. It was the only turf runway I ever used in that airplane. Before I would head there I always reviewed the short field procedures. I knew them but when your normal runway is long and wide, it’s best to think through any use of a short one to make sure you get it right.

The proper short field approach speed for the airplane is 72 knots. That is 20-percent above the stalling speed. For a balked landing, the drill is to use full power, flaps immediately from 30 to 20-degrees, and 70 knots airspeed until obstacles are cleared. If there were a recommended wind shear escape procedure on approach, that would probably cover it. I’d add that even with almost 9,000 hours in that very airplane, I would have felt like I was doing some far out and on the edge flying if I had to do that.

Gaston’s is in hilly country so the arrival path to runway 24 is dictated by terrain as is the departure path off runway 6.

Where wind shear encounters on takeoff can be avoided by waiting for the weather to pass or moderate, the normal departure from a short runway is what it is, even in fair weather. Pilots don’t often calculate required distances for takeoff and climb but there are times when it is the most important thing that you can do. That’s the only way that you can know you aren’t asking your airplane to do the impossible.

Be aware that when you calculate distance requirements for takeoff the numbers in the POH reflect the best the test pilot and the airplane could possibly do. You have to add in any margins. I always used the book figure for takeoff and climb to 50 feet as a minimum runway length.

Sugarbush airport
At airports like Sugarbush, you’d better know the performance charts for your airplane.

For my P210 the drill was flaps 10, brakes on, full power set, brakes released, rotate at 65, maintain 78 knots indicated until all obstacles are cleared and then transition to a normal climb. I almost never did this and when I did, it felt like I was close to the edge. For that reason, I would try it at home, at a spacious airport, before heading off to small airports like Sugarbush (OB7), or Minute Man (6B6). Those are both in New England.

Some of the most demanding flying I ever did was in my Piper Super Cruiser (PA-12) as I prepared for and then took my flight instructor check ride from Al Meyer, a quintessential FAA inspector, on August 24, 1953.

The spot landings were a particular challenge because of what Al would and wouldn’t let you do. After the power was chopped, you could clear the engine one time, briefly. No slips were allowed. The pattern could be altered but there still had to be a turn from downwind to a base leg and then a turn on final.

What to do if you were a little high on final? The normal approach speed was probably 1.2 Vso, or maybe 1.3, and if you slowed below that, the drag as well as the rate of descent would increase. In the Super Cruiser, it felt to me like it had to be quite close to the stall before it started to descend more rapidly. If that were done, the pitch attitude had to be reduced at an altitude high enough to allow enough airspeed to build for a successful flare and landing. Edgy stuff, but with enough practice I got good enough at it to satisfy old Al.

If you were a little low on final and gliding at the correct speed, there was nothing that you could do to stretch the glide. The best plan was to be a tad high, with that slowdown procedure at the ready.

When thinking about the edges of the envelope I always get back to the 65-horsepower trainers that we all learned to fly in back in the good old days. The J-3 Cub was the standard. One year in the 1950s or early 60s, Mr. Piper (W. T. Piper, Sr.) was given an award by the NBAA, the heavy iron crowd, and the person doing the introduction asked that everyone in the banquet hall who had learned to fly in a J-3 stand up. Not all stood but way more than half did.

What in the world does that have to do with the edge? Well, in those old trainers the envelope was pretty small. A J-3 would cruise at maybe 65 knots and stall at 35. So there wasn’t much there but a pilot with a good feel for the airplane could play the J-3 like a fiddle and fly safely in and out of places you wouldn’t even think about today.

On a hot day with two on board, a 65-hp J-3 didn’t have much rate of climb. It rather had to be nursed up to pattern altitude. When reading of the daunting challenge that faced the 727 pilots at New Orleans I wondered if either had ever before flown an airplane that would just barely climb, like Cub pilots flew every hot day? The similarity ends there because there is a big difference in doing it at 1,220 pounds and 38 knots as opposed to 171,139 pounds and 138 knots.


  • I believe there is a great hole in today’s pilot training. While students are introduced to real slow flight, once past that initial instruction, young pilots simply don’t practice that close-to-the-edge flight as religiously as they really should. There is still a lot to be said for seat-of-the-pants aircraft control and operation.

    Thanks, Dick, for once again hitting the nail on the head.

  • You’re right Dick. Loved flying into Gaston’s Resort in the Malibu, but as you stated, taking off from 24 (also with high power lines) was clearly not the way to go. But with the proper approach speeds, it was relatively easy to roll to a stop with gentle braking. Highly recommended for those wanting a grass strip landing in their logbook. Not to mention a great brunch on Sunday…!

  • I have successfully done my share of short-field landings and takeoffs… and some times with crosswind. The thing I learned is to never under estimate the drag your landing gear has while still on the runway, particularly in winter. It slows getting up to full take-off speed. I always yanked it off as soon as she would do it, level out and build up speed with ground effect. As the end of the runway approached, it was possible to pull some flap and enjoy a quick climb up to clear the trees, etc. This maneuver saved my buns a number of times.

    Mort, what `cha think?

    • Here, with a 1700 ft runway and cliffs on both ends here at 55I, I apply takeoff the same way. Nose off early, stay in ground effect, hit the flaps at VX, and off I go. This is SOP during the warm months.

  • I personally don’t think in today’s way of doing business that anyone in their right mind would take off in those conditions as the Pan American 727 did in 1982. It probably had the old internal combustion engines, and with all that rain coming down on the runway displacing all the air needed to support lift on that fully loaded aircraft is ludicrous, to say the least. That was suicide, especially when said that the NTSB could not quantify that large amount of rain the crew was exposed to. Also, the crew using only the “allowed” power to take off, instead of maximum power to be used when needed, and God knows they needed all they could muster. It all, in my opinion, was a disaster going to happen !!! It saddens me to hear of all those souls lost when if studied and applying common sense in some instances would have prevented all that tragedy. Too much say so, and innuendoes and not much “facts”. thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *