Wednesday, September 26, 2007
Fish Story
Copilots typically have to do all the drudge work prior to departure, partially because, as they say in the military, “Rank has its privileges,” and partially for a better reason, which is that the captain is always distracted with a million little questions and problems prior to departure and most of the drudge work just wouldn’t get done otherwise. But it often leaves a very busy and sometimes frustrated copilot trying to get all the “boxes” loaded (the navigation units and performance computers), the takeoff data computed, and, the bane of all copilots, the weight-and-balance form completed.
The weight-and-balance computation is critical to flight safety because an overweight aircraft won’t perform as expected, and an out of balance aircraft can become uncontrollable. Think of a canoe: The more weight you put in it the harder it is to paddle and may eventually sink. Put all the weight in the front or back and it will take on water and become uncontrollable. Verifying that the aircraft is loaded within its weight limits and that that weight is properly distributed is part of the copilot’s job prior to every flight: The copilot takes the data that the ground agent has collected, data on passengers, bags and cargo, adds the weight of the fuel, looks up a bunch of numbers on various charts, adds them all up, checks them against the limits, signs off on the form and hands it to the captain who checks it and signs off on it, taking responsibility for its correctness. What makes all of this really hard for the copilot is that a lot of the data—the final passenger count and the final bag count—doesn’t get to him (or her) until the last minute. Then, with everybody on board and everything loaded and ready to go, everyone waits and watches while the copilot looks up, enters, adds, checks, and finally finishes the form so the captain can check and sign it.
On this particular day I knew we were going to be close to full with a heavy fuel load to get us non-stop to Paris, and even without the final numbers I knew we were going to be close to the maximum allowable takeoff weight, which I had computed separately based on runway length, elevation, temperature, atmospheric pressure, and wind. So when one of the Air Afrique ground handlers came on board and said he wanted to put 10,000 pounds of fish in C1, the forward cargo bin, I looked at my partially completed form and said, “No, we can’t do it, not in C1, not in C2, not in C3, we’re already close to max gross and there’s no way we can take 10,000 pounds of fish.” (He actually gave me the weight in kilos, an additional complication, but a kilo is 2.2 pounds, so that was pretty easy to convert.) He said, “But this is very special fish, all on ice, it’s going to some of the best restaurants in Paris,” which I could believe because Dakar is right on the ocean and I had seen some incredible displays of fish in the outdoor markets and had eaten some fabulous fish dishes in the hotel’s restaurant. I said, “I’m sorry but there’s just no way we can take 10,000 pounds of cargo.” “He said, “But then the fish will all go bad. We will have to throw it all out.” I said, “Well then, you’ll have to take 10,000 pounds of something else off, and it can’t be fuel, it will have to be bags or people.” He said, “We can’t do that. We will have to leave the fish.” I said, “I’m sorry, that’s the way it is,” and he left.
I got my final numbers, lots of people, lots of bags, lots of fuel, no fish. I finished the form, gave it to the captain, he quickly checked it, signed it, and all the paperwork “went out the door”—we were on our way. It was my turn to fly, the captain having flown it down from Paris the night before. The takeoff roll seemed normal until I went to rotate. Rotation begins by bringing the control column back, which forces the nose to go up, the tail to go down, the wing to start developing lift, and the airplane to takeoff.
An important part of the weight-and-balance computation is the takeoff trim setting, which varies depending on how the airplane is loaded. The takeoff trim setting, a number, is set on the trim wheel by rotating it until that number is aligned with an index pointer. Then it is checked prior to takeoff as a part of the pre-takeoff checklist. The purpose of the takeoff trim setting is to provide a given degree of back pressure on the control column when rotating. Consistency of pressure results in consistency of rotation which is important to achieve the takeoff performance expected: Too slow a rotation results in a longer takeoff roll and less clearance over obstacles, and too rapid a rotation increases drag, delaying lift off, and in a worst case scenario, can cause a tail strike, damaging the aircraft. In this case when I tried to ease the control column back, the force required was much higher than I expected—I almost needed two hands, and the rotation was slow. The captain looked over at me, he knew I was a brand new copilot, and I’m sure he was thinking, “What have I got here? Can’t even do a decent takeoff,” but I got it to rotate and fly, retrimmed to relieve the back pressure, and sat there trying to figure out what was going on. I must have figured the trim setting wrong, I thought. I would have five or six hours to try to find my mistake once we got to cruise altitude, so I tried to put it out of my mind until then.
That’s when the second strange thing happened: We couldn’t get to our assigned altitude, barely climbing 100 feet per minute the last 1000 feet, versus 300 to 400 feet per minute that would be normal for our weight and outside temperature. We knew we were heavy, and weights are always estimated in scheduled service, never exact, since they are based on average passenger and bag weights, and it is not at all unusual to be heavier than estimated, but not usually by so much that you have trouble making your flight planned altitude. We got to altitude, eventually, and with nothing much to do for several hours I got my copy of the weight-and-balance form out and went over it, trying to find my error. I checked all the weights, all the moments—the balance part—checked the arithmetic, with a calculator, twice, and couldn’t find any errors. The captain checked it, the flight engineer checked it, and they couldn’t find any errors either. So I checked the flight log, and after a couple of waypoints points it was clear that we were burning fuel faster than expected. Not seriously so, just a little bit more each segment, but at the current trend we were going to land with several thousands pounds less in reserve fuel than we had planned for. The weather was good in Paris, traffic was expected to be light, Charles De Gaulle has several very long runways with precision approaches in both directions, so all the things you want extra fuel for—vectors for traffic, holding, missed approaches—weren’t factors, but still, we were burning more fuel than we should have. Extra weight equals extra fuel burned: It takes energy to lift and carry that weight—nothing is free in aviation. Everything was beginning to look like we had not just a little more weight than estimated, but a lot more, and that it was toward the front, probably in C1.
We eventually got to Paris without further incident, although with quite a bit less fuel remaining than planned, and as I went about packing up my stuff and cleaning up the cockpit, the captain went back to say goodbye to the passengers and the flight engineer went outside to do a post flight walk around. When he came back in he had a funny smile on his face. He said, “You won’t believe what I saw in C1. Must be 10,000 pounds of fish, on ice.”
Made me wonder what happened to the DC-10.
Monday, September 24, 2007
Position Plots
Position Plots
One of the things I’ve always liked about flying at ATA is that we don’t have much tolerance for Captains who want to do things their own way. Non-standardization does happen, of course, as every crewmember who has ever served as a copilot or flight engineer knows. The reasons usually given are, “That’s the way I was taught to do it in flight school,” or “That’s the way we did it at Brand X,” or “That’s just the way I’ve always done it,” none of which are good enough reasons, but fortunately, we don’t have a lot of it. We may not always agree with or like the ATA way of doing it, but by and large we realize that there has to be one way for everybody, and if we really don’t like it we try to work within the system to change it by calling a fleet manager or writing him a letter or talking to a check airman or sim instructor and running our suggested change through one of them. I think that is one of our real strengths, and has helped to make us as good as we are.
What we do have, though, is a lot of very bright pilots and engineers who are always working on better ways to do things, who in the process sometimes introduce elements of non-standardized individualism masquerading as techniques. These individuals feel that what they are suggesting is within the limits of variation that we commonly describe as technique, and that in so doing they are improving things. In fact, in many cases there may well be a reason they haven’t thought of for why their change is not a good idea, and in most cases their variation goes beyond technique and is a change in our procedures. While trying to improve on our procedures is always a good idea, changing them on an individual basis is not.
There are obviously some gray areas here, and the point of this article is not to categorically draw the line between technique and standard procedures. That would be very difficult to do. (One airline, who shall remain nameless but whose first initial is D, as in Delta, defines technique as anything not in the book; that assumes a pretty good book, to me.) An example might help, though, and one that I have noticed is that pilots often have different ways of plotting their position when doing MNPS [Minimum Navigational Performance Standards—applies to the North Atlantic] crossings. There really is only one correct way to do it, and anything else is not technique but is non-standard. But the reason people are doing this, as I have discovered in talking to them about it, is not that they’re just stubborn and want to do it their way, but that they think they have come up with a better way to do it that is not so different as to be non-standard. And almost invariably this thinking, while good in itself, results from a misconception about how and why we do position plots. So let’s take it from the beginning.
We are required to prepare a plotting chart showing our cleared oceanic track anytime we operate in MNPS, NOPAC [Northern Pacific], or CEPAC [Central Pacific] airspace. We are also required to plot our position on that chart 10 minutes after passing each waypoint along the track. The purpose of that position plot is to verify that we are indeed operating along the cleared track. Very simple, really, and possibly because of that, subject to several misconceptions.
The first and most common misconception is that the purpose of the position plot is to verify that the nav system in use is tracking properly—following the cleared, loaded, and plotted track. In other words, that it is working properly. This is the first example of how something fairly simple gets fuzzy and complicated if we let it.
All we do when we plot our position on the plotting chart is verify that the position plotted—our location—is on the desired track. It doesn’t tell us anything about how well the nav system in use is working. It could, in fact, be in error by a considerable amount, but if it thinks it is working properly (there are no warning or error messages), then it will also think it is where it is supposed to be. An analogy might be to tell someone to drive 20 miles due east and stop. If that person’s compass were off 90 degrees but he or she didn’t know it—if, for instance, when it read due east it was actually indicating due south—that person would drive off due south for 20 miles and report in. “Here I am. I made it. According to my compass and odometer, I am 20 miles due east, just like you wanted.” He or she could even plot that point on his AAA highlighted road map to make sure he went to the right place, and sure enough, 90 degrees east for 20 miles would fall right on the map where it should. Remember, all we do when we plot our position is verify that we are where we’re supposed to be—on the desired track. If the nav system in use is not accurate, but does have proper data entered, it will still think it is where it is supposed to be, even if it isn’t. There are other ways to check on the accuracy of the system, but this isn’t one of them.
Another misconception that comes from this is that if the nav system in use always thinks it is where it’s supposed to be, whether right or wrong, what’s the point of plotting it? It will always be “on the line.” And, again, if the point of position plotting were to identify errors in accuracy or in the operation of the nav system in general, there would be no point. But that isn’t the point. The point is to verify that the aircraft is actually navigating along the desired track—the cleared route.
So then, why wouldn’t it ever not be on the plotted course? There are three possible reasons why the position plotted might not fall exactly on the desired course; two of these are fairly inconsequential and one is very consequential and is the reason we do it.
The first and most common reason the plotted position might not fall on the line is because the line is wrong. Those little tick marks can be hard to see and hard to count in the dim light of the cockpit and it is also easy to mix up latitude and longitude. You might plot 50 North and 51 West, for instance, instead of 51 North and 50 West. (You would never be cleared over a longitudinal line in MNPS airspace that wasn’t an interval of five or 10—20 West, 30 West, etc—but it is still easy to mix coordinates up.) In almost every case where the plot doesn’t fall on the line, it is because the line was drawn wrong. If this happens, have somebody else in the cockpit verify that that is indeed the problem, redraw the line, and the line should fall on the plot.
The next most common mistake is that the plot was done wrong, either written down wrong, or simply not plotted in the correct spot. Very easy to do, and again, if you think that is the problem, have somebody else verify it and correct it.
But the third reason the plot might not fall almost exactly on the plotted track is that you aren’t where you are supposed to be—the track is right, the plot is right, but the aircraft is not navigating along the desired track—uncorrected, you are headed toward a track bust. There are really only a couple of ways this can happen, and plotting an accurate position is the only way to catch them.
The most obvious cause for a navigational error is that the nav system in use does not have the correct coordinates for the next waypoint. (The waypoint behind has to be right, or you would have had a deviation off track on the last 10 minute plot.) If this occurs, immediately check the next loaded waypoint against the oceanic clearance: it should be readily clear to everyone in the cockpit whether it is correct or not. Reenter the correct next waypoint coordinates in the nav system in use, and reintercept the correct track as expeditiously as possible. Then, of course, replot it to be sure everything is back to normal.
Other causes for nav errors would be that the autopilot is in heading instead of nav, or in radio nav instead of long range nav, or has no lateral mode at all, or isn’t even coupled—the latter unlikely but possible. (But being left in heading mode instead of nav is not at all unlikely, and was the reason for one of our very rare track busts.) It is also possible that the autopilot is “cross-coupled” (the captain using the B autopilot, the first officer the A autopilot) and the other side is set up for something else, FMS nav, for instance, with only an abbreviated or previous route entered, or Route 2 on the opposite GPS. Another cause might be that a reroute was entered into the non-flying side only, or didn’t transfer over correctly. Finally, in a triple INS aircraft, the INS in use might be in manual waypoint change instead of auto. The only way you’re going to catch these kinds of oversights—mistakes that have not been caught already as part of the normal check and double check process—is with a position plot.
At the risk of seeming to be a bit obsessive, let me add that there is also a right way and a wrong way to plot the position. The right way is to locate the latitude and longitude along the grid marks to either side of the track, mark the top and bottom and left and right side on those grid marks with short tick marks, then put the plotter first along one set of marks and draw a line across the track, then do the other set, also using the ticked grid marks. The two lines should fall on the track. In other words, plot both the latitude part of the position fix and the longitude part independently using ticked grid lines. Anything else sets you up to see what you want to see. If, for instance, you just lay the plotter along the grid marks for the longitude and mark that point on the track, and then put the plotter along the latitudinal grid marks and mark that on the track, the temptation to put the plotter where it should be to fall on the track, rather than on the actual noted position, is very, very strong. This may seem like nit picking, but it is not. A careful, accurate position plot is your last chance to catch a gross error in navigation. This is something you do for yourself, to stay out of trouble, not something you do for the company because they said so (although they do). It’s in your own interest to do it right. Otherwise this last good chance to catch a major error—flying off track or along the wrong track—is going to be missed.
Finally, which nav system position is the best one to plot? The nav manual says to use the nav system that is coupled to the autopilot, and it says that for a reason. What I see is a lot of guys using the third system, which would be either the Litton INS in a GPS/FMS/INS aircraft or the third INS in a triple INS aircraft, for their plots. The reason given is always, “The nav system in use always thinks it is where it is supposed to be, so what’s the point of plotting that one? The third system is independent. That’s a good cross check on the one in use.” And they are right in that respect: comparing the calculated position of one nav system to another, particularly two completely different nav system types, is an excellent way to check on the functioning of those systems. If everything is working properly they should agree on where they are, within the limits of system accuracy. But just because they agree on where they are doesn’t guarantee that the nav system in use is going where it should. Only an accurate position plot of the system in use—the one that is doing the driving—can do that. Plotting the third system, one that is not coupled, simply tells us its position, which is nice to know, but really doesn’t mean anything because it’s not doing anything. (It may be supplying position information to the FMS and to the other two INS’s in a triple mix configuration, but it isn’t driving the airplane.) There is absolutely nothing wrong with comparing the position, track deviation, and track angle error of one system to another—in fact, that is an excellent idea and one that is provided for in the way we set up our nav systems using the appropriate MNPS checklist—but doing that won’t tell you, not for sure anyway, that you are actually navigating along the cleared track. An accurate position plot taken off the nav system in use—the one coupled to the autopilot—will. Every time.
So to review, position plotting tells you nothing about the accuracy of your nav system; position plotting does tell you whether you are navigating along the cleared track or not. The only valid and meaningful position plot is one taken off the nav system in use; anything else is “nice to know” and might point to an error, but in and of itself is irrelevant since it isn’t doing the driving. The right way to plot position is to hold the position on the nav system in use 10 minutes after passing the waypoint, mark the corresponding latitude and longitude on the grid marks to either side of the track, connect them with a plotter, and see where the intersecting lines fall. Anything other than exactly on the line (within the limits of plotting accuracy), means that the line was drawn wrong, the position was plotted wrong, or you are off course.
Think of your position plot as a waypoint triple check. The first check is done on the ground after all waypoints have been loaded (then a circle is drawn next to each on the flight plan). The double check is a recheck of an approaching waypoint’s coordinates, the distance between the next two waypoints and the course between the next two waypoints (then a line is drawn through the circle). The triple check is the position plot—your actual position relative to the cleared track. This is your last chance to catch a deviation while it is still small enough to correct without adverse consequences, meaning before enough time has passed that you could have wandered over to an adjacent track. Remember, this is also air traffic control—strategic navigation, or separation. We could all probably find Shannon with much simpler procedures than this if that was all there was to it. But this is much more important than just simple navigation: this is what insures that we stay inside the little bit of protected airspace that has been reserved for us over a very big ocean honeycombed with many other little bits of protected airspace, each with a relatively tiny aluminum airplane inside it, full of people. It’s not brain surgery (for one thing, brain surgeons only have one life in their hands), but it isn’t tiddly winks either. For all kinds of reasons, it’s worth the effort to do it right.
Monday, September 17, 2007
Equal Time Point Considerations
The article that follows was something I wrote in 2001 as a check airman for my fellow ATA Airlines Lockheed 1011 pilots operating internationally and overwater. It is based on Part 121 and the ATA Airlines General Operating Manual (GOM), but I have added explanations in brackets as necessary where a reference might not be clear to a non-ATA pilot. It will be of interest mainly to transport pilots flying three and four engine aircraft, but I hope others will find it interesting and useful as well. September 17, 2007.
Equal Time Point Considerations
I’ve been taking check rides from Tom Hopp [a former Eastern Airlines captain and check airman, now an ATA simulator check airman and examiner] for almost my entire career at ATA. We both started at ATA within two weeks of each other, and he’s been lying in wait for me ever since. No matter how many times I have put myself at his mercy, he always seems to come up with something new to trip me up. The last time he had me in his sights though, I thought it would be different. I had a notebook full of surprises from previous check rides, and I was there for a simple SVT where he tells you in advance what he’s going to do, and virtually the order in which he will do it: a normal takeoff, an all engine non-precision approach, a rejected takeoff, an engine failure past V1 but before V2 followed by a two engine approach to a go-around, loss of the second engine on down wind with a single engine approach and landing. We had already done the non-precision approach and rejected takeoff, so their wasn’t much mystery about what was coming next: an engine failure just past V1. I pushed the throttles up, reminded myself that there was no law that said he couldn’t give me another rejected takeoff but heard and saw nothing coming up on V1, worked hard on tracking the centerline knowing full well what was about to happen, heard “V1, rotate”, waited for someone to say, “Engine failure,” but heard nothing. I glanced down at the engine gauges, but they were all normal, heard, “Positive rate,” said, “Gear up,” and then just sat there, trying figure out what was going on, feet spring loaded but with nothing to shoot at. Dead silence. We went through 300 or 400 hundred feet, and I finally decided he either wants to see another non-precision approach, or else maybe we’re going to do an all engine autoland or PAR, but what we’re clearly not doing is a “V1 cut.” Then it happened: “Engine failure.” Only now I had no center line. All I had was an attitude indicator which was slowing rolling, a flight director that was telling me to steer left, an airspeed indicator that was showing a slow deceleration, and an altimeter that was no longer climbing. My instinctive response was to bank into the good engine, which was wrong, didn’t work well at all, then I remembered my feet but forgot to get rid of the bank, wrong again, and Tom Hopp had done it to me again. I’m sure he thought it was real funny. The engine is supposed to fail at V1 while you can still see the centerline. We all know that, we train for it, we prepare for it, and just because Eastern had engines fail at points other than V1 doesn’t mean we have to train that way here.
But it’s the FAA that really got me into the mess. The FAA wants to see engine failures tested at the most critical point from a performance point of view, and that is directly after decision speed with the aircraft as low, slow, and dirty as it can get. The assumption is that if both the airplane and the pilot can handle the worst case situation, then anything after that doesn’t need to be tested. But while that may be the most critical point from a performance point of view, it isn’t necessarily the most critical point from a handling point view, as I demonstrated. That point comes with an engine failure low and slow, to be sure, but also without any visual references and by surprise. And by analogy, the same holds true for ETPs—Equal Time Points—which is why I have indulged in this rather lengthy introduction. We compute and plot an ETP for the most critical performance situation enroute—the simultaneous failure of two engines—knowing full well that that is almost certainly never going to happen, while somewhat overlooking situations which can and do happen where the ETP is a critical factor in the decision making process. One is a regulatory, theoretical, worse-case scenario, carefully computed and plotted and prepared for but seldom having any real world value, and the others are all the things that actually happen that are not very carefully anticipated or prepared for. So let’s start with what the regs say and require, the worst case scenario, and build from there.
Some basics. ETP’s are a product of Part 121, and specifically that subsection that details the enroute limitations for transport category turbine powered aircraft certificated after August 29, 1959. That’s us. The first limitation deals with one engine inoperative, and basically says you have to be able to clear all terrain after the failure enroute of one engine. That’s normally not a problem for those of us lucky enough to be flying the three motor jobs, but is for the 757 guys Tom Hopp beats up on now. The second enroute limitation is for two engines inoperative. Note that in neither case does it matter how many engines you start with. The more the better, but even if you have six engines, you still have to meet the enroute limitations requirement for the failure of one and two engines (it’s just easier if you have more).
The reg provides two ways to deal with the loss of two engines. The first is simple: don’t ever get more than 90 minutes, with all engines operating at cruising power, from a suitable airport. (“Suitable airport” is defined separately but basically requires that it have a runway long enough to land and have a 40 per cent margin above the minimum landing distance.) If you do that, you don’t have to be concerned with any performance limitations following the lose of two engines. This allows two engine aircraft to fly at all under Part 121, and allows the rest of us to fly most of the time without having to worry about what we’re going to do if we lose two engines—we land at the nearest suitable airport which should be no more than 90 minutes away. (Single engine speeds are slower than all engine speeds, so it could end up being a little more than 90 minutes away, but not by much.) I don’t know what those guys with two engines do when they lose two engines, and I hope I don’t ever find out.
The second part provides a way for aircraft with more than two engines to get further than 90 minutes from a suitable airport, but at a price, and that price is the ability to show that after the simultaneous failure of two engines at the most critical point—the farthest point away in terms of time—that the flight can still continue to a suitable airport, clearing all terrain and obstructions by 2000 feet, and arrive at least 1500 feet directly over the airport with at least 15 minutes fuel remaining (4000 pounds total in our case). Fuel dumping is allowed.
The most critical point is what we call the ETP—the Equal Time Point—that point along the route of flight where it takes just as much time, and therefore just as much fuel, to proceed forward to a suitable airport as it would to turn back to one. If there never were any wind, that would always be the halfway point, allowing for the four minutes or so it takes to turn around, but wind moves the ETP backwards with a tailwind and forwards with a headwind in ways we all learned to compute at Mark Barnard’s or Gene Freemen’s knee [ATA navigation instructors]. Now the ETP computation is done for us by dispatch, we dutifully plot it on our charts, and that’s pretty much the end of it, because if we lose an engine it will usually be pretty obvious which is the better course, turn back or continue, and if it’s not obvious, if we lose one close to the ETP, then that’s what it’s for, to help decide (but even there, with two still running, there is some “wiggle” room).
With the lose of one engine we know we will have to descend to a lower altitude, maybe as low as FL 190, and we know that our fuel burn will increase because of the inefficiency of flying on two engines at a lower altitude, but unless fuel was critically low prior to the engine failure, we will still probably have adequate fuel to reach a suitable airport even with one out: surprisingly enough, a well trimmed airplane flown at long range cruise doesn’t burn a whole lot more fuel with an engine out than it does with all engines operating. (But long range cruise at FL 190 will be slow—you can’t fly at FL 190 on two engines at Mach .84. Well, actually you can, but only for a little while.) Normal cruise at FL 350, 400,000 pounds gross weight, for instance, is 485 knots with a total fuel flow of 17733. Long range cruise at FL 190 (stabilizing altitude at 400,000 pounds), is 416 knots with a total fuel flow of 18126—slower with a slightly higher burn, as expected, but not significantly so. If you lost the engine with, say, two hours to go at normal cruise, it would take about 2 hours and 25 minutes at the lower altitude, increasing the burn from 35466 to 42234 pounds, a difference of about 7000 pounds. (Actually, it wouldn’t even be that much because I figured that based on a constant burn over the entire two hours and 25 minutes—I don’t have a flight planning computer handy—whereas in reality either the burn would drop or the speed increase as fuel burned off, as always. So 7000 pounds is on the high side in this case.) That’s a big hit, but this is a worst case scenario, and as I said, unless fuel was critical before the failure, you should be able to live with that. Go arounds and diversions to alternates might be jeopardized, but remember, you just lost an engine over water: declare an emergency, call ahead and get priority with a secure runway, and then shoot a coupled autoland approach to Cat III minimums if necessary. We know the airplane can do it. But it can’t fly without fuel. Take the go around or diversion out of play.
So, you’ve lost an engine two hours out, but it’s under control and you have a plan that you are confident will get you to dry feet safely. Now, what if you lose another one. Oh boy. The first thing I would think of is that the odds of losing two engines within the space of two hours for unrelated reasons are about the same as for Ed McMahon meeting me on the ramp at Shannon. Therefore, if two can go, three can go, so at some point, in addition to briefing the senior flight attendant on the engine failures, I would instruct him or her to start preparing for ditching, as a precaution. It can’t hurt, it might end up being necessary, and the old military adage of a busy troop is a happy troop applies here. (“Happy” might be a stretch, but having something to focus on besides what just happened can only help.)
The first thing to do, obviously, is to go to max continuous power, IR [Increased Rating] if available, just like in the sim. It doesn’t matter if you are on downwind or at altitude, with one engine you need all the power it’s got. Trim for the yaw, get a drift down speed, trim for that, and declare an emergency if you haven’t already. Then start to think about your stabilized level off altitude. If this is a second failure following an earlier failure, then the ETP decision should already have been made: either you had already decided to turn back or you have been continuing on after the failure of the first engine. (While, in theory, ETP1, ETP for one engine remaining, is not exactly the same as either ETP2 or ETP3, because of differences in the winds aloft for the respective cruising altitudes for each, in practice they will be very close to each other.) So once you have made a decision either to continue or to turn back, you are committed to it even if another one fails.
You knew fuel was going to be more critical after the failure of the first engine. Now, having lost another one, it can only be worse. Remember that all you were guaranteed prior to takeoff was 15 minutes of fuel overhead the airport after the simultaneous failure of two engines at the ETP. So if you lost one, and have stabilized at some lower, two engine cruise altitude, and have been flying there for awhile after having either turned around or continued on, your fuel situation would actually be worse than if they had both failed at once. But then again, you are not exactly at the ETP either—you have already covered some ground either going back or continuing on, which is much more important, and the further you are from the ETP the better. So the seriousness of the situation really depends on how close to the ETP the second failure occurred.
In any case, fuel is an issue, and probably the last thing you want to do at this point is to dump fuel, but that may be what you have to do. Remember that the regs allow for dumping fuel to meet the ETP requirement. If you are on a very long range flight, say
How do you know whether you need to dump or not? You could refer to the Performance section of the Aircraft Operating Manual, but my guess is with everything else going on—multiple check lists, frightened flight attendants and passengers (of course the three of you are still so cool all you’re thinking about is whether this is going to get you out of your next pairing), calls to Stockholm Radio for phone patches, and trying to figure out what caused two engines to quit and hope it is not something common to all three—hauling out the AOM from under the desk may not be the first thing you think of. The quick answer is on the flight plan. If dumping is required, it will say so in the ETP section, second line, which will look something like this:
KBWI/ETP BURNOFF 74691 MAGW 395600 DUMP 0.
MAGW is the maximum allowable gross weight at the ETP, and the number following DUMP is the amount of fuel that must be dumped at the ETP to reduce to that gross weight. The important number is the first one, MAGW, because you could be much heavier than that if your engine failures occurred early on during the crossing. It may say “0” after dump, but that assumes you got to the ETP before losing your second engine, didn’t put any extra fuel on, got your flight planned altitude, and so on. If you weigh more than the MAGW figure, you’re going to have to dump. You can check that number out by going to the Performance section, page 150, the Single Engine Driftdown/Cruise Climb Procedures. That chart starts with a weight of 373000 for ISA up to +9, and there is a note, number 2, that says that if aircraft weight exceeds that shown on the chart, then fuel jettisoning should be considered. However if you go to the two engine inoperative long range cruise charts on the pages that follow (page 154-164), there is a chart for cruise at 1000 feet, 390000 pounds, ISA to ISA +20. So after allowing for the burn drifting down to 1000 feet, you can see where a number like 395600 for maximum allowable gross weight comes from.
There is another interesting part to that note, and that is that it says that the path down to 15000 feet is not significantly affected at higher gross weights, and that if fuel dumping is required it should be delayed until 15000 feet and accomplished by 11000 feet, and that the dump rate is approximately 4500 pounds per minute. Even at the heaviest weights, you would still have quite a bit of drift down time remaining at 15000 feet—it will still be coming down, but much more slowly as it approaches its stabilized driftdown altitude, so you should have plenty of time to dump beginning then, and, having burned some coming down, would know with more precision how much remained to dump at that point.
So the old adage that the only time you have too much fuel is when you’re on fire isn’t exactly true. Too much fuel is a very real possibility following a second engine failure, depending, of course, on how much you started with and when the failures occurred. You have to get that weight down to where the airplane can maintain altitude on one engine, and that may mean having to dump fuel. You don’t have any other choice: you can’t very well toss passengers out, and while you might like to be able to dump bags or cargo, there isn’t any real good way to do that either. You can’t even dump the lavs. You’re stuck with everything you took off with except the fuel. Which is why fuel planning and watching your payload—your zero fuel weight—is so important anytime dumping is required to meet ETP requirements.
In fact, the GOM says that no additional fuel or payload may be added without coordination with dispatch. In other words, the captain may not add up to 6000 pounds of discretionary fuel as normally allowed if dumping is required. And he or she would be foolish to add payload above that flight planned, unless he traded it for less contingency fuel—fuel over the minimum. (If you aren’t sure why, read the preceding paragraph over again.) I’m not sure why the GOM doesn’t let us add fuel, up to 6000 pounds, as always—it just means more to dump if it comes to that—but I assume dispatch is just being very conservative. But whatever the reason, the simple fact is this: however unlikely it may be that you will simultaneously lose two engines exactly at the ETP, if you want to make sure you are legal you better check with dispatch before adding any extra fuel or payload. In practice, a little extra fuel or payload probably isn’t going to hurt you because you are almost certainly not going to lose two engines at the ETP. But if you’re paperwork is selected by a Fed for audit, look out. That would be a really stupid way to pick up a violation. And if you were unlucky enough to lose two at the ETP with more than flight planned payload, that would be a really bad way to end a career. Could be a very long, lonely retirement, assuming you were lucky enough to successfully ditch the thing, get in the rafts, and then survive long enough to be rescued. And with the kind of luck that got you in that mess, I wouldn’t count on that either. (Anyone want to go see “The Perfect Storm” again?)
So are there two or three simple lessons we can get out of this? I think so. One would be to remember that minimum fuel doesn’t just mean minimum fuel to reach your destination and alternate, it also means minimum fuel to reach a suitable airport after the loss of two engines at the ETP with 15 minutes of fuel remaining. The word “minimum” really takes on new meaning here.
Another is that if you do lose an engine, the farther you can get from the ETP before a second one fails, the better off you’re going to be.
And the third is don’t ever, ever put on more payload than you’re flight planned for unless you can reduce fuel by an equal amount, never of course reducing below minimum. Otherwise you may find yourself trying to figure out how to break into C1 and C2 [cargo bins] from the galley, open the doors from the inside, and shove everything out except yourself. I can’t wait to see what that logbook write-up might look like: “At 30 West simultaneously lost all power on engines 1 and 2. Unable to restart. On postflight found two cargo doors, all bags, and the Flight Engineer missing. No FIRM code. No other defects noted.” (And the info line for that week would begin again with, “Well, we had another rough week operationally…”). Just remember, if dumping is required, watch the payload. And dumping will be required anytime you lose two engines with a gross weight greater than that shown on the ETP section of the flight plan. That’s the long and the short of it.
Monday, September 10, 2007
Blue Angels
I have always felt that the real test of acting on principle was when it was not in your interest to do so, and that the best test of character was what you do when you don’t know anyone is looking.
The Blue Angels, the US Navy’s demonstration team, normally only does shows in the United States—it’s primary mission is recruitment, so that makes sense—but it does occasionally tour in other parts of the world, the last such tour in 1992. (This information, and much more, including a gallery of photos from which the one above was taken, can be found at the official website, http://www.blueangels.navy.mil/.) I was in Italy then, but I wasn’t aware of any other traffic one quiet morning as I sat the cockpit of an L-1011 holding short of the runway prior to takeoff at US Naval Air Station Sigonella, least of all the Blue Angels, so I was a bit surprised when the tower told us to, “Hold short, arriving flight of six. Blue Angels.”
Saturday, September 8, 2007
Last Flight
The picture attached to the profile section on the right (“About Me”) was taken by my First Officer on the night of September 24, 2006, just prior to my last flight as an ATA Airlines Boeing 757 captain. The flight was from McCord Air Base, Tacoma, Washington to Bangor, Maine, the first leg of a multi leg charter for the Air Mobility Command, a “Reach” call sign flight, one that would eventually take the passengers, US Army ordinance specialists, to Kuwait City International Airport. From there they would transfer to Air Force tactical aircraft for the final leg into
It takes several crews to complete a long troop movement like this, and my job was to take them as far as
My last approach and landing couldn't have turned out better. It was early morning by the time we arrived, the sun was just up, the air was perfectly clear and smooth. Bangor Approach cleared me for a visual landing, handed me over to the tower who cleared me to land, and the actual landing was one of my better ones. I taxied in, set the parking brake, called for the parking checklist, filled out the log book, and then reached over and shook my copilot's hand. I said, "That's it. It's over. Thanks for your help. It couldn’t have been a better last flight."
Saturday, September 1, 2007
Babies Cried
Part of the problem is that even experienced travelers often don’t have any idea of the various factors that go into landings, and therefore are not very good judges of what constitutes a good landing and what doesn’t. A somewhat rough landing in gusty winds to a short, icy runway is an excellent landing: get it on, get it on firmly, start stopping it right now. A pretty good landing in calm conditions on a long runway with water on it (which has a cushioning effect) is probably just mediocre: In those conditions, Ted Striker could probably pull off a great landing.
Some airplanes are easy to land and some have reputations as being difficult. The Boeing 727 was really hard to get a good landing out of, although really awful ones were rare too. It tended to bounce. Every pilot had his or her own theory of what caused it and how to prevent it, but none were very convincing or provable. It just wanted to bounce. And if you had lots of runway and tried holding it off until you were sure you were slow enough and close enough to the ground to keep it from bouncing, it would invariably drop right out from under you with a solid crunch. The late Don Lanham, ex-Braniff 727 captain and head of the ATA training department for awhile, also one of the funniest guys I ever knew, used to say, “Just fly it on down to within a inch of the runway and let it drop in from there.”
The L-1011 was altogether different. It didn’t bounce, but landings could vary from “Are we on the ground yet?” to “Is everyone okay back there?” A big part of the problem is that it is a large, wide body aircraft. At touchdown, the cockpit is still 35 to 50 feet in the air, and “the mains”—the main landing gear, arranged in two four wheel groups called trucks—are some 90 feet behind you. Being that high up in the air, with the mains so far behind, meant a certain amount of guess work and even luck was involved. There was just no way to have a good “seat of the pants” feel for where those mains were prior to touchdown.
The 1011 had a radio altimeter which showed the height of the wheels above the ground, but you couldn’t look at it and still look out the front window, so it was hard to use in those last 50 feet. The flight engineer, who sat behind the pilots, was supposed to watch the radio altimeter for you and make call outs every 10 feet from 50 feet above on: “50, 40, 30, 20, 10.” The call outs not only told you how high you were, but the cadence gave a sense of the rate you were approaching as well. It all helped, but not all engineers were equally skilled or careful in making the call outs.
I sort of learned this by accident. (Not that kind of accident—inadvertently.) I was driving to the airport, a two hour drive early in the morning with no traffic, and I was thinking about my crew and in particular why I always seemed to have good landings with some engineers and bad ones with others. I assumed it was just luck or coincidence, because I was doing the landings and what did the engineer have to do with that, and then it hit me: He has one very important thing to do with that, he makes the call outs. That particular day my engineer was one that I often had good landings with, so I made a point of sneaking a glance at the radio altimeter as he was making the call outs, and sure enough, his call outs were right on. And I kept doing that with other engineers, and the quality of the call outs varied all over the place, from pretty good to, “Somewhere around 50 feet I guess I’m supposed to say, ’50, 40, 30, 20, 10,’” without any real correlation to reality. So I learned two things, one obvious, the quality of call outs varied, and the other less obvious but more important, that that information was vital to good landings. If the call outs were bad you were better off ignoring them, taking your chances and trusting your instincts. If they were good, you almost couldn’t make a bad landing.
The Boeing 757 presented a completely different situation, because it was a two man airplane—no flight engineer—but its radio altimeter was equipped with an automated voice call out. So accuracy of the call outs was not an issue: they were always right on. In addition, while a long aircraft, the cockpit position on touchdown was much lower so it was easier to get a visual picture that corresponded to the wheels touching down. The rule of thumb with the 757 was, “Don’t do anything until you hear ’10 feet’.” That was easier said then done though, because the ground comes up pretty fast those last 50 feet at typical approach speeds of 130 to 140 knots and at a descent rates of 650 to 700 feet per minute. At 30 feet all of your instincts are telling you to start pulling back, but that was wrong. I proved the rule to myself over and over again by getting better landings at Maui, a short runway that required you to just put it on and get it stopped, than I did at San Francisco, where the super long runways often lulled me into starting to flare too soon and then use the extra runway length to try to squeak out a good one. It almost never worked, but like the golf lesson that only seems to work on the range and never on the course, I never seemed to be able to make myself do a Maui-type landing anywhere but Maui.
One San Francisco landing in particular was memorable. Conditions were perfectly benign, a nice evening at SFO, very little wind, the flight over to and back from Hawaii had gone well, I had a good crew all around, and a lot of the passengers were regulars. One in particular I had gotten to know well because he commuted back and forth between Maui and the mainland, which many of the regulars did, but his situation was a little different because he had broken his back years earlier in a surfing accident and was paralyzed from the waist down. He needed an “aisle chair”—a skinny wheel chair that can go up and down the aisles—to take him directly to his row where he had to be lifted out of the aisle chair and into his seat. And vice versa getting off. He was just a terrific guy with absolutely no sense of self-pity about him, and I knew he was back there, had talked to him before takeoff, and I probably I wanted to pull off a really good landing for him. I certainly wasn’t aware that I was thinking that, but probably I was. In any case, for whatever reason, I still don’t know what I did wrong, but the squeak I was looking for was more like a load of bricks being dropped. Airplanes can take a lot of abuse—in flight testing they deliberately drop them in hard from 50 feet to see if anything breaks—and this wasn’t anything that bad, but it wasn’t good. It was certainly the worst landing I had ever had in a 757, and this was after flying it for a couple of years.
When those things happen you always hope that maybe, just because you are way up in the front and forces tend to get exaggerated the farther you are from the center of gravity, that maybe it wasn’t so bad in the back. Maybe you can get away with, at worst, a “Whose landing was that?” from the flight attendants. (Pilots normally alternate legs, taking turns doing the actual flying, and flight attendants know that, but don’t necessarily know whose leg it was. When they ask, its always because something was either really good or really bad.) So after the check lists were done and the log book filled out, I opened the door, and Laura, the Senior Flight Attendant, was right outside the cockpit door, laughing. Almost uncontrollably. I finally got her to calm down enough to say, “So Laura, tell me. How bad was that, really?” She said, “That was so bad, babies cried.”
I wrote in Fly Like a Pro, (Tab Books, 1985), and I quote it here because more than 20 years later I still can’t say it any better: “Exercising good judgment is really what being a good pilot is all about. But that’s not what most people think. Most people think that the physical manipulation of the controls is what being a good pilot is all about. That’s why the passengers always pay so much attention to the landing. They think that the landing is the most difficult and critical part of the flight, and if the pilot does that well, he must also do everything else well. It just isn’t so. The most difficult part of the flight is not the landing; the most difficult part is making the proper judgments so that you arrive at a point where a safe landing can be made.”
I guess my landing was safe: no one got hurt and nothing broke. Still, babies cried.