It is not often in life when you are presented with an impending disaster that threatens many people, and you believe you are the only one who can do anything about it. On those rare occasions, even if you are the indecisive type, there is simply not an option to pass the buck to somebody else. By all indications you are the only one who sees it unfolding, or the only one in a position to do anything about it. Those circumstances are even less frequent in a cockpit with the redundancy of the automation, plus a radar controller, and other crew members. But when it does happen there is a stark confrontation with reality; you are it, the only chance. This is a story about that.
This story happened in a 767-200ER, such as N334AA, which was the first jet to hit the World Trade Center on September 11, 2001. I flew that jet many times, but do not recall which one was involved in this earlier incident.
I was pretty convinced that this was my situation as it played out many years ago. Sometimes while driving a car we have to swerve, stomp on the brakes, or even thread through a maze like Mario Andretti, but those reflexive heroics come from a different place because you merely react and hope for
the best. This story played out over a longer period of time, enough time to envision the CNN report, and think about the 200 people behind me, and who knows how many in the aircraft we were about to hit? In some ways time was warped that night. Every seemingly endless nano-second came with the awareness that time was of the essence to convene a meeting, over the radio and between several players, because one of us was definitely reading from the wrong play-book. The aircraft we were on course to hit was on a different radio frequency and the mess quickly became worse than a game of telephone, where I thought the accuracy and timing of every word out of my mouth could be the difference between dominating tomorrow’s news
posthumously, or continuing routinely to our destination. These were the ingredients of an experience that became etched upon my gray matter as we disappeared into the Chicago mist on one winter evening around 1995.
My challenge now is to tell the story so you can appreciate the dilemma from my point of view. Let me start by referencing a mid-air collision between a B727 and a civilian airplane over San Diego, California a few years prior. See:
https://en.wikipedia.org/wiki/Pacific_Southwest_Airlines_Flight_182
Looking back, I was surprised by the date cited in that link (1978) because it meant the industry took over a decade to adopt a mandate to put TCAS (Traffic Collision Avoidance System) into the cockpit. But when you consider that the device went from concept, as a direct result of that accident, to a government mandated installation in that amount of time, maybe it was not so long. The development of TCAS was an admission that the skies had become too crowded, airplanes too fast, and cockpit duties too demanding for the existing separation tools. Until then the separation provided by ground based controllers was augmented by the old see-and-avoid concept, but clearly an additional redundancy was needed. It was not really an invention because it was developed relatively easily from existing technology. They tapped into the standard transponder, which had been used for decades by
controllers to see airplanes better on radar. With TCAS, airplanes could now see each other, or at least a blip on a monitor, regardless of the amount of darkness or cloudiness. Furthermore, the programming within this device
produced climb or descent instructions to avoid a collision. My story occurs at a time when it was new equipment for all of us in the industry. Design flaws in first generation software were readily apparent, and we were told that
most or all of the fleet had been upgraded to the second generation. Were there still flaws? TCAS was first received skeptically as another device to restrict a crew, or take control away from the controllers, but it was not
long before most of us had a story where TCAS had saved the day. Here is mine. I may as well add the obvious; the TCAS is not much help until the switch is turned on.
I was the copilot on this dreary evening. We were heading to Dusseldorf, Germany in a Boeing 767-200ER. The “ER” part of this designator stood for “Extended Range”, which basically meant our fuel tanks were large enough to go long distances; however the “200” part meant they did not make the engines any bigger to accommodate that extra weight and only a slow climb could be expected during our departure. The captain was at the controls and I was in the right seat as copilot (AKA first officer or FO). Another copilot (AKA FB) was in the jump seat. The purpose of an FB is primarily to allow rest breaks on long trips, but he or she is always in the cockpit for takeoff and landing, when on board, just for the extra vigilance during these busy phases of flight. We were about to demonstrate the wisdom of that practice.
As I hinted earlier, the weather was a definite factor, with low overcast and drizzle. The visibility below the overcast was only 1 mile, and the base of the overcast was published at a low 300 feet. Soon after takeoff the captain would be flying solely by reference to the flight instruments. This was all in a day’s work, so far.
Chicago O’Hare circa 1992
O’Hare tower had cleared us for takeoff on runway 32-Right, with instructions to turn right to 060 and climb to 5000 feet altitude. Since the runway number reflects the compass heading during takeoff, this meant we would
be turning right from our initial heading of 320 degrees (northwest ) to the heading of 060 (northeast). Meanwhile, a different controller on a different frequency cleared an AA Fokker 100 for takeoff on runway 4-Left assigning them runway heading, which meant his initial heading would remain 040 degrees (northeast). The second frequency was used most of the time to reduce congestion on the crowded frequencies. I apologize if this seems like too much information. Suffice it to say that these were routine departure instructions that had been developed over the years in Chicago, instead of last minute creations by whimsical controllers. They were part of a master plan to maximize the volume of takeoffs, while keeping aircraft separated, until further instructions could be issued after takeoff. This incident highlighted a weak link in that plan. I believe the Fokker crew made two mistakes which, combined with all these other circumstances, resulted in this event. The industry refers to such an incident as a near-miss, even though it would be more appropriately named a near collision.
First they did not turn on the transponder, which for decades was one of the last items on the before-takeoff checklist. Since they missed this, and the weather precluded a visual verification, the tower did not know when they were airborne after issuing the takeoff clearance, and did not hand them off to the next frequency at the normal point.
Second they automatically changed frequencies. A quick comment was probably made in the Fokker cockpit about usually talking to departure control by this point, causing them to conclude that they were expected to change automatically to departure control. With the transponder off, the departure controller would have instructed them to “reset transponder” and, when they turned the switch, everybody began hearing the bells and whistles warning of an immanent collision. By my count, “everybody” would have included both tower controllers, the departure controller, and all the pilots
involved. That is at least 8 people on three different frequencies.
From my vantage point the takeoff was busy, but progressing normally before the alert. The workload of managing the airplane acceleration, retracting the gear and flaps, and hand-flying the climbing turn on instruments, motivated the captain to ask me to turn off the outside lighting to remove the disorienting effect of the scatter-back light reflecting off the clouds. This was the one time we were allowed to turn off the lights at a low altitude and the only time in my career that I remember actually being asked to turn them off. In hindsight it meant that the other airplane would have absolutely no chance of seeing us. This would have been cited as a minor contributor in the accident report, if the worst had come to be. The gear was up and we were in a climbing turn, slowly accelerating just above minimum flaps-up speed. We
were a newly hatched bird with limited abilities, hardly flying, and we needed time to increase our energy level. I saw a sudden and concerning target pop up on the TCAS screen. Our information showed that he was clearly
airborne, but at a lower altitude and climbing. How could ATC (generic air traffic control) allow another aircraft to be so close to us, and in the weather? I mentioned it to the other pilots as I reached for the microphone button. As I started my radio call, the target changed to the color “amber” and became a TA (traffic alert) with the accompanied announcement “Traffic Traffic”. I knew the radio call had to be clear because there was not going to be a lot of time to straighten this out.
“O’Hare tower, this is American 46, what do you know about this traffic alert we are receiving?”
He knew where we were since, in addition to seeing our transponder signal on his radar, he had cleared us for takeoff just a minute earlier. For similar reasons, he should know about all of the nearby traffic as well.
Tower replied “He just took off on 4 left, but we’re not talking to him, we think he might have switched to departure control.”
By now our alert had advanced to an RA (resolution advisory), an urgent instruction which appears in red with some sort of directive designed to coordinate separation like “Traffic-Climb, Traffic-Climb”(more about
possible directives later). I looked to the right to where the baneful bogie should be as indicated by TCAS, but the clouds were too thick to see anything.
I asked “Is there something we can do to avoid hitting him?”
By now the FB (jumpseater) was tapping my shoulder and pointing out lights emerging and converging. He had to get out of his seat to get this view. Remember, we had turned our lights off, reducing the chance that they might see us.
“Stop your climb at 3000 feet” was tower’s astounding reply.
This was out of the question and to this day I do not know what he was thinking. We were climbing, slowly, through 3000 as he spoke and the target was converging from the right 700 feet below us and climbing much faster. If we had leveled at that moment, he would have hit us for sure.
“Negative, we cannot do that, what else can we do?”
I felt a little like a defiant child as I said no to his instructions, but his first idea was clearly not going to work.
“Roger, American 46, climb to 4000 feet and turn left to 360.”
Now that was something that we could do, and with relief I read back the instructions. I watched the airplane lights slide below us as the captain continued to give his full attention to the challenge of hand-flying a heavy 767 at the edge of its performance on instruments. We are supposed to disconnect the autopilot for a TCAS RA, though he had never engaged it. The TCAS showed the target passing 500 feet below us to the left and then back to the right again. As he continued diverging in heading and altitude, TCAS announced “Clear of Conflict”, and it was over almost as quickly as it began.
Aside from the reports mandated by the NTSB for such events, it was history and we continued routinely to our destination. In hindsight, leaving the transponder off until just prior takeoff, a decades-old procedure, suddenly seemed like an invitation for trouble. The switch was historically left off until the last minute so that parked and taxiing airplanes did not show up on the controller’s radar. But aircraft taking off straight into the overcast at night were simply not going to see each other; and the ability for a radar controller to see us was dependent upon a single person at a single moment turning a single switch. In this case, that person was the co-pilot of a Fokker 100, the
most junior bid-status in a fast growing airline. He certainly had accrued substantial experience somewhere, but he was likely a new-hire at American Airlines. Soon after this event, the procedure was changed to turn on the
transponder while holding short of the runway, a less busy time that allows pilots to back each other up better and adds more time to accomplish the item. Was this change coincidence, or direct result of our near miss? I do not know the answer to that, but as I write these memories it occurs to me that this might be the second change in procedures resulting from reports that I wrote, the first being the root cause of my near stall scenario (see Coffin Corner).
That was pretty much the end of the story, however I promised more detail about TCAS for your edification, a direction that is a little spooky for me. I mentioned that TCAS issues directives to accomplish separation during
an RA. The designers decided that these directives would be expressed only in rate of climb. Pilots are expected to continue what ever heading they were holding, or turn they were in, while simultaneously altering their climb or
descent to match an array of red or green lights in the vertical velocity indicator (VVI). If you were to ask me what command we were issued by TCAS that night, I would answer simply that I don’t know. It is not that I forgot
because I never knew. Surely it was screaming something, but I was too task-saturated to perceive it. Maybe the captain knew and maybe it was consistent with the solution I was engineering. The actions we had to take were so obvious to me that I dismissed whatever TCAS was directing. Truthfully, responding contrary to TCAS instructions can be a very bad move because the transponders talk to each other. They say something like “Hey brother TCAS, I see that you are heading my way so let’s coordinate our efforts, how about you keep climbing like you are and I will descend.” So the VVI in each airplane displays different instructions and if one of the pilots does not follow his correctly, it undermines the programing. In another twist, for obvious reasons, TCAS will not command a descent below 1500 above the terrain, could that have been a player? Also, it is possible to get an RA where airplanes cross altitudes. If TCAS thinks that is the quickest way to assure separation, it will add the word “crossing” to the aural instructions. I suppose my tunneled attention could have missed our TCAS saying “Traffic. Monitor vertical speed. Crossing”, while the Fokker crew heard “Traffic, Increase climb, Crossing”, to quickly fly through our altitude. They were light and only going the short distance to Cleveland and were probably out-climbing us significantly, so that might have been a viable solution. Maybe the controller got them to level off, after hearing my radio call. So, if we happened to follow the instructions or not, I do not know, but we found our way out of the mess, and I guess that is what counts. There were issues of trust and inexperience with TCAS, with plenty to go around for all involved, but now
that we had been rescued by the gadget, I think we all had a readiness to accept it as the lifesaver it was.
Captain’s Log 