Sept. 29: A Pilot Looks at the Skies

By Joe Sharkey

More than a dozen pilots from major airlines sought me out in the months after the Sept. 29 mid-air collision that killed 154 over the Amazon last year. Some spoke on the record. Several asked to remain anonymous.

One, the pilot interviewed below, asked that I send him the transcript of our long discussion before he would decide whether to allow his name to be used. This interview occurred several months after the crash, and I have not published it before.

The reason: Unfortunately, I had a computer meltdown not long after this interview, and all of my e mail correspondence with various pilots, including this pilot, was lost. (If he reads this and agrees to the use of his name, I will of course add it).

I’ve decided to publish this as the anniversary of the Sept. 29 disaster approaches, because this pilot’s comments – on why you don’t see another plane coming at you at these speeds, on technology that’s sometimes too precise for its own good, on Brazil’s ill-trained air-traffic control work force – are important to a general understanding of what happened.

JS: Why would a pilot talk about this?

I sincerely believe in aviation and aviation safety and I also think I understand why, through my reading of the press, your collision happened.

What’s very sad, of course, is that the military controls air-traffic control in Brazil, and there’s that added layer of self-protection that, during self-examination, doesn’t work.

It is very complex, and I don’t know if you’re aware of this, but many years ago, a SwissAir DC-8 landed at Athens, went off the end killed about 14 – 28 people, and the pilots were arrested. It was at the beginning of my flying career. It was back in the 70’s. I flew for USAir.

JS: You flew international?

No, I made a conscious decision after watching that fiasco in Greece, that I did not want to fly anywhere but the United States and maybe Canada. Or maybe the U.K. But once you start going across the Atlantic, you start going to places, well …

You recall, I think in 1989, a USAir plane went off the end of the runway at La Guardia. And the two pilots were thought to be – it was considered by the District Attorney that these pilots might be held on criminal charges. In America! I knew the one guy, from California. And that has always scared me as a pilot, because you don’t know when an accident will happen.

A suggestion on transponders

But getting back to the concept of midair collisions. Everything I’ve read indicates the transponder (on the Legacy 600) for some reason wasn’t working. Even over the North Atlantic, where there is no radar, if the transponder is working, the TCAS, the collision avoidance system, would work …

And if I may suggest: That a device that automatically monitors the output of the transponder be hooked up to an auxiliary transponder and when the main one fails to send out a signal, a secondary one automatically comes on with battery power that cannot be turned off from the cockpit. And it squeals an emergency code that cannot be changed. And this does two things: It still will trigger the TCAS in another airplane. And if another 9/11-like event happens and you turn off the transponder, it will keep some sort of transponder on so that traffic control can track you better. This was one of the things that the 9/11 commission specifically called for, a transponder that could not be turned off. And it hasn’t been done. It’s five years later.

Seeing an oncoming plane

And the other thing, again from your bully pulpit. When an airplane is flying along in the flight level, up around 30,000 feet or whatever, typically, they turn off most of their lights. There are a few flashing lights on the wing tips, maybe, on the roof of the plane, on the belly. But there are recognition lighting systems that make a huge difference in ability to just look out the window and see another airplane coming at you, even in broad daylight.

And see every little bit helps.

JS: Why didn’t the Legacy pilots or any of the passengers see the oncoming 737 on Sept. 29? One pilot and one passenger reported only seeing a fleeting shadow. The other pilot saw nothing.

I was in the jump seat watching other pilots fly one time in Southern California, and there was a child’s toy balloon that had managed to get up to about 29,000 feet. And I spotted it, and I yelled ‘Traffic 11 o’clock, it looks like a toy balloon!” And the two pilots, who were actually flying the plane, when they saw it, they almost jumped out of their skin, because the closure rate is nothing short of amazing. You see a flyspeck on your window and by the time you blink you’re colliding. You’re going 500 miles an hour; they’re going 500 miles an hour – that’s 1,000 miles-an-hour closing rate.

JS: So you could see nothing but a shadow even when you’re looking out the cockpit window?

Yes. Let me give you an idea about the human eye. The human eye, unless it is focusing on something in particular, is really only good to about 20 feet. When you’re just staring out the window, you’re not focused on anything. But if you say, “Oh look, on the ground over there, there’s a train moving.” Then your eye focuses on it and you can make out some pretty good detail. But when your eye is relaxed, and the eye goes from scanning outside, which is distance, to scanning inside at the instrument panel, which is two or three feet from your face – it’s a hard accommodation for the human eye. And there are blind spots in every human eye. It is said of Chuck Yeager, the famous pilot, that he just had exceptional vision, which made him a better World War II dogfighter. So there are many concepts there.

The Crowded Sky

But one thing I would like to talk about the concept of midair collision and avoidance. Over 40 years ago, there was a novel called ”The Crowded Sky.” What happened was, and this is what happened to you guys, I’m afraid, air-traffic control screwed up, there was a mechanical error compounding it, and two planes were head-on. The author of this book, more than 40 years ago, said that the way we work, the airway structure, is completely stupid. And it is stupid. In a freeway, you have cars going north on the right side on both side of the line. And one side goes north and one side goes south, but they’re never head-on.

In a vertical airway, all planes are head-on. If one is going north between, let’s say, New York and Boston, and one is coming south from Boston to New York, they want to go on the same, most direct route possible. So instead of separating them, left and right, they separate them by altitude. That’s the stupidest thing in the world when you could just say, if you’re going to go from Boston to New York you’re going to have to go 20 miles out of your way, over here, at one altitude, and the other direction is over here. And we used to do that in America. They were called “alternate airways.” But what this man wrote about more than 40 years ago was called the “airway alleys.” And it’s just the altitude separation. There are conceivable scenarios where, let’s say, you lose pressurization. You’ve heard of this: rapid decompression. And the plane must go down quickly to allow people to breathe. If you’re on an airway and you go down, you could conceivably run into another airplane going in the other direction. And be head-on.

You see, if you’re all going in the same direction it would be, at worst, more like a glancing blow and there would be this reduced closure rate. Instead of a 1,000-m.p.h. closure rate, it would be a relative closure rate of something like 20 or 30 m.p.h., which is much more handle-able.

There are things that can be done right now, technologically speaking, that would prevent this from happening. It has been a great debate among many pilots. I’ll tell you right now what happened to you, in my humble opinion. It was an air-traffic control error, a complete misunderstanding of what the pilots were going to do, because of the following combination of poor English skills by the Brazilian air-traffic control and the concept of the difference between a clearance and a flight plan.

A clearance – that’s what we agree on right now and off you go. A flight plan is something you file to get the clearance. I have filed a flight plan that says I wanted to go 20,000 feet and I wanted to go from Boston to Newark and then to Baltimore, and I’ll get on the radio, and I’ll take off, and they’ll say, just go direct Baltimore and don’t go over Newark at all. And you know what? As soon as those words are out of the mouths of the air-traffic controllers, and I respond by saying, “Roger, direct Baltimore,” you throw the flight plan away.

JS: Legally?

Virtually. And this is what happens. And the same thing (on altitude).

“Well, you’re filed for flight level three-five-zero.”

“Well, we’re requesting two-niner-zero.”

“Okay, climb and maintain two-niner-zero as a final.”

And now, you’ve just thrown the flight plan away. Happens routinely. Who cares?

Amazingly, when you file a flight plan (and the airlines do it for the pilots, but when it’s corporate or a private pilot or something you have to do it yourself), and you put down that you have an alternate airport if you go to your first airport and it’s fogged in and somehow you lose radio communication, legally you should proceed to your alternate – there is a huge problem, since they’ve thrown the flight plan away when you got your clearance, they have to go look up which alternate you’re going to. So in my opinion, the poor guys – your two pilots – were probably doing everything right and there was a misunderstanding by Brazilian air-traffic control what the pilots were expecting to do. And this was a very bizarre flight plan according to what I have read, that they were going to fly at one altitude and then change altitudes after crossing a certain space. It’s just something you don’t do normally – at least in the United States.

You file a flight plan. Off you go, and if you want to change your altitude, you say, Mother May I change my altitude? And Air traffic control says it’s okay, and they you’re changed. Honestly at first, we were reading rumors of changing altitude 5,000 feet – and there are guys who horse around out there, and I thought, well maybe these guys were just showing how neat the plane flew.

Turning off a transponder

I’ve also known guys who have made mistakes on their altitude and they’ve turned their transponder off so air-traffic controllers couldn’t see their altitude. It happens. If you’re off your altitude, you could conceivably get in trouble if you’re caught. Of course, this is the mentality of a 12-year-old kid, not a professional airline pilot, but we have a few 12-year old kids still flying airplanes and they’re 50 years old. If you’re supposed to be at 37,000 and you’re at 37,500, the FAA says, “Hey, you were off your altitude. Let’s take your license away for a month, and the airline says, “You can’t fly for a month? You don’t get paid for a month.”

So, some of the guys, reach up and say, Oh shit. I blew my altitude. I’ll turn the transponder off and ease myself back down to my altitude and then I’ll turn the transponder back on. If anyone says anything, we’ll write up the transponder as “intermittent.”

Of course, a passenger wouldn’t know that.

JS: But that wouldn’t have been the case in the Amazon crash

No. Here you had technical issues with the transponder, including inherent design flaws in the unit itself. [Ed. Note: Among other things, the Honeywell transponder on the Legacy and other aircraft has been criticized by safety officials for its placement in a position where it is easily jostled. And after the crash, the FAA issues recommendations that new transponders be easier to monitor and issue an audible alert when they inadvertently switch off].

The reason they made that stupid design is even more ludicrous. You can be flying along, and air-traffic control says, “Change your transponder code to a different number.” Fine. There are certain different combinations of numbers. And if you go through one or two combinations, getting to your assigned number, for just one sweep of the radar scope if you are on the wrong number and that wrong number is one of two or three codes, it sets off a huge alarm bell at the air-traffic controllers. And they don’t want you to do that, because it’s distracting. So Honeywell says, okay, if you do this, move it fast enough within five seconds, we’ll just shut it down because we don’t want the controllers to be worried about a false alarm. As a professional pilot – and there are a lot of guys who are pretty mediocre out there – a professional pilot says, well I have to get to number 4213 but not go through 7700 or 0000, and you just manipulate the controls in the proper direction and avoid the bad numbers, the ones that would set off the alarms. It’s like a safecracker. But you’re up there getting six figures – you should know how to do it.

Better technology, harder to use

In the great race of modern avionics, they’re getting worse, not better. In the man-machine interface. The numbers have to be big. Pilots become far-sighted as they age. And you should have big numbers so there is no error. Who are we trying to impress by making the numbers so damn small?

It’s absolutely terrible the way modern avionics, computer controlled planes, fly by wire planes – they’re running away from the human being….

Back to the transponder…

This might be part of the equation. This is why I make the suggestion. If the transponder is not putting out something, the TCAS on the other plane won’t see it. In the planes I’ve flown, DC-9 and 737, when the transponder goes off, the TCAS will fail, and there will be a warning light of some sort saying, “Transpon failure. TCAS inop.” Do you know if the guys said anything like that?

JS: They said they had no idea the transponder was off. At impact, the transponder suddenly went back on again, and the Legacy pilots said they didn’t turn it on; it came on by itself with the jolt.

That is part of the equation, I’m sure. But I personally think what also happened is at the last microsecond, the pilots of the 737 saw something. And remember, in cruise flight, autopilot is usually controlling the plane and the pilots are usually a little bit relaxed. They usually tell jokes, and sometimes they’re looking at charts. Sometimes they are reading a newspaper. It happens. And if they saw something, they would have reached up, grabbed the control yoke and turned the auto-pilot off and most likely moved the control wheel to make a very hard right turn. Doing that, and cutting off a piece of the wing, and probably jamming the control somehow – altogether may have contributed to the crash.

If you were in a car and spilled a cup of coffee on your lap and looked up and you were head on with a school bus, only God himself could stop you from trying to swerve. And that’s what I’m saying. The plane may have been upset and lost a piece of the wing.

May I also point out that in 1989, near LA, a small Piper Cherokee hit a DC-9 and they both went down. And even earlier then that, in San Diego, a Cessna 172 also hit a 727 as it was making its landing approach to Lindbergh Airport there, and they both went down. So a small airplane can bring down a big plane. Which is also one of the reasons I talk about a having transponder that can never be turned off.

The Bush Administration wanted to do away with primary radar which would skim paint instead of use transponder at many sites because the transponder works so well most of the time.

Global positioning in the center of the airway

The other thing a lot of pilots are talking about – offsets on the airways. You know, 20 years ago, we used to fly using a system called VOR -- Very High Frequency Omnidirectional Range. Every 50 to 100 miles, the pilot would reach up, and you would change the frequency and you would find a new radio beacon -- but fly to the next beacon. Now, with all this global positioning system, you never have to do anything. And the GPS can put you on the middle of the airway, within inches (of your official position). The old VOR system, you were lucky to be two or three miles of the center of an airway. And that was acceptable.

And that’s why you guys collided. Everything was so precise. Both of your altimeters were exactly right at 37,000 feet. There’s an allowance of 75 feet. If yours had been 75 feet high and theirs had been 75 feet low, you would have missed. There used to be ‘slop’ built into the system. The same thing with determining the exact course you were on. You were both smack-dab in the middle of that alleyway, instead of being a little sloppy like the old system was. Some pilots have suggested that an automatic slop factor be added unknowingly to the pilots in the en route environment so you are not so tightly on that flyway.

I’ve been in two near midair collisions in my career. One at Philadelphia Airport. We were cleared to land on Runway 9 Right. And someone was cleared to take off on Runway 27 Left, which is 9 Right in the opposite direction. And we kept on saying, “You’re telling us someone’s taking off on 27 Left. Do you realize you’re heading somebody straight for us?”

“Oh, it’s okay,” they kept on saying.

Until we saw the DC-10, and we were in our DC-9 and had to duck underneath it at 800 feet. Now, it all worked out, but air-traffic control kept on saying, “No, there’s no problem,” and we kept saying, “Do you realize what you’re telling us?”

Three controllers were decertified on the spot because they didn’t realize the ramifications, even though we were questioning it over and over again. If you listen to the tape, you say, “What the hell happened here?”

And you see, in the United States, ever since Ronald Reagan fired the air traffic controllers, air traffic control has never been the same. My incident happened in ’89, the controllers were fired in ’81 or ’82.

There has been such a rush to embrace technology instead of a rush to embrace well-trained, sober, intelligent people and treating them well.

Brazil: Investing in technology but not in people

And the same thing in Brazil. They bought a brand-new air-traffic control system, and in some ways it’s more modern than what we have here. And they don’t invest in the people. It takes years to be a good air-traffic controller. But there is a rush all over the world to say, we’ll make a gadget and you push a few buttons and you won’t have to think. There are airlines now that are hiring copilots with between 200 and 800 flying hours and they’re putting them right into 747’s as copilots. Many of these so-called commuter airlines flying Embraers and Canadair regional jets are taking very low-time pilots. They call it Ab Initio training. They say, Hi, give us $70,000 and in a year you’ll be a co-pilot on an airplane.

I didn’t get hired at (a major carrier) until I had about 5,000 flying hours. And they’re hiring people at these regional airlines flying, these nice little 50-seat jets, with very little flying experience.

The communication problem in Brazil

One of thee things I’ve read is they noticed the communication problem. You see there’s a whole bunch of procedures to follow, and there are some differences between American regulations and ICAO regulations.

If you hang up right now and you say, “Maintain 37,000.” I’m going to maintain 37,000 feet until I’m practically on top of my destination, and then I have to land.

But if you said, “Maintain 37,000. Expect 36,000 in five minutes.” I’d start my watch.

And that’s the way they did it in America for the longest time, and they still do. You don’t give an altitude unless that’s a sacred thing. And (at the Legacy departrure airport in San Jose dos Campos, Brazil), the novice controller obviously said, “Oh, his flight plan says he’s going to go down to 36,000, I don’t have to talk to him anymore.” Which is bullshit.

Runway overruns

Fifteen years ago, I went to two representatives. A member of the House of Representatives and a Senator. And I talked about these overruns on runways made out of crushable concrete. And I said, please put them in. And of course last year, a plane slid off Chicago Midway and killed this little boy. And only now it’s starting to get done. And people for years have been talking about the tombstone mentality of the FAA. They wait for a bunch of tombstones to sprout up after a crash. Oh, we’ll fix it after the fact. Why don’t we fix it before anybody dies?

My note: this interview was conducted more than six months before the July crash at Congongas Airport in Sao Paulo, Brazil, that killed 199. The problem of runway overruns is notorious at Congonhas, especially when it rains. Brazilian authorities responded to the latest crash, the second deadly air disaster in Brazil in 10 months, by sharply reducing operations at Congongas, Latin America’s busiest airport, and by promising to build a new, safe airport at some unspecified time in the future. No real safety measures were undertaken.