The Pilot’s Lounge #80: The ILS — That Last 400 Feet

The discussions here in the Pilot's Lounge at the virtual airport range from prosaic to fairly intense. Last Sunday evening some of the regulars returned from their annual mid-October color tour on a particularly windy day and got into it over speeds to fly on the approach. The group had "ooh"-ed and "aahh"-ed at the colors, then landed at Mackinac Island, an airport that is renowned for attention-grabbing gusts, swirling winds and shear when landing to the west. Moments later, they had watched a Cessna 210 approach with partial flaps as the wind gusted to 30 knots. They heard the pilot chop power over the threshold, but then juke and jive and float 2/3 of the way down the runway, with the clearance between the airplane and the ground varying impressively, before he finally bled off enough of the massive excess of airspeed he was carrying to touch down. They then held their collective breath as he swerved to the right at the very end of the runway and stopped at the edge of the pavement. Their observations of the hapless gyrations of a pilot who hadn't figured out that excess airspeed in the flare is not a terribly good idea on a gusty day, sparked a conversation that somehow evolved into the speed to fly on the ILS and how to make the transition to visual references and land.Look, I never claimed that discussions in The Pilot's Lounge follow logical sequences, I just report them.A couple of the instrument instructors spoke up and said that they were getting increasingly concerned about what seemed to be a common thread among pilots with whom they'd flown. In the last 400 feet or so above the runway on the ILS, unpleasant things were happening. A pilot would have done a decent job of keeping the needles near the center as he or she neared decision altitude; airspeed, descent rate and heading would be appropriate, but when the pilot spotted the runway it was Katy bar the door. There would be a whirlwind of activity in the left seat as the power was slammed back, flap deflection increased and the dive for the threshold of the runway commenced. It was as if there were some sort of prize to be obtained for landing short. It made no sense to the CFIs; the airplane had been nicely established on the localizer and glideslope and, just because the pilot thought he or she could see the runway, it was suddenly time to make all sorts of speed and configuration changes, despite that fact that they had yet to get as far as the approach lights.

Ben, one of the local charter pilots and a long-time instrument instructor, spoke up. He keeps up with publications about how people learn and said that the dive for the runway practice worried him because, from what he had seen in the literature on habit patterns, pilots who practice doing things one way will do them that way in the future, even when it's not appropriate. [Editor's Note: See Linda Pendleton's recent column about this.]In Ben's mind, that isn't good, because if there is any ice, fog or low scattered clouds around, the habit of making significant changes before getting to the runway threshold on an ILS could be the last screw-up such a pilot makes.Ben said he looks at NTSB reports during his wait times on charter trips and he'd noticed several examples of pilots who had been flying an ILS in poor weather, broke out, spotted the runway or the approach lights ahead and crashed short of the runway or touched down on a parallel taxiway.According to Ben, the pilots who survived and were interviewed were convinced that they had a good view of the runway after shifting from instrument references to visual references, even in seriously restricted visibility, very low ceilings and/or scattered clouds below the airplane. The pilots then proceeded to disregard the very instruments that had been providing excellent guidance, and made a power and/or configuration change prior to reaching the runway threshold and came to grief. Those who landed on a parallel taxiway tended to do it during daytime, when there was snow on the ground, often under "flat light" conditions, in a crosswind, after disregarding the localizer needle. Some said they broke out, saw what looked like a runway directly ahead, and not remembering they were crabbed due to a crosswind, they landed on that snow-covered surface before realizing it was a taxiway.As Ben spoke, I recalled reading of an accident involving a Lockheed Constellation at Shannon, Ireland, some 40 years ago. The crew shot the ILS in heavy rain. Upon spotting the approach lights both the captain and first officer were convinced they were high. Ignoring the glide slope needle, the flying pilot closed the throttles and stuffed the nose down. They landed the airplane in the approach light array. Refraction of the approach lights through the rain-covered windshield was blamed for the mishap. In recent years I have seen publications that assert there is a potential for such a visual illusion and others that pooh-pooh it. Whether the illusion exists doesn't matter, because the power setting and descent rate that has been working all the way down the ILS will probably continue to work at least until crossing the threshold, so why get excited and make changes just because one has spied the runway?

Ben's comments caused me to back up and look at the last portion of an ILS approach and landing. Leaving approach speed out of the equation, let's think about what's going on and why we're doing the ILS in the first place. Unless we are just practicing, we're shooting the ILS because there is some question regarding getting to the runway safely, usually because of foul weather. Unfortunately, we have spent hours and hours doing practice ILS approaches on good weather days, so when we looked up, we saw a nice, clean, dry runway, shimmering in the sunlight. We've seen that hundreds of times before, right? When we see a runway in good VFR weather, we know what to do, don't we? We immediately slow down to the speed we normally fly on final, and aim for the end of the runway, completely disregarding the aiming point markers on the runway and that happy, little, glideslope needle that has been our friend up until now. We land just past the numbers, while congratulating ourselves on a well-flown ILS.Egads.Because we do it that way again and again, odds are we'll do it that way when the weather is down around our shoelaces. When we shoot an ILS for real, that first view of the runway presents a much different picture than we have seen it on our carefree, VFR practice days. The rain is hammering against the windshield, or the ice has reduced vision to a small hole above the defroster. We pick out the approach lights and use them to line up as we start to see pairs of runway lights at the approach end of the runway. There is no horizon; the approach light array helps us tell if we are wings level. If we've got any ice on the airplane, the last things we want to do is make a power reduction or extend any flaps at all. But, what is our habit pattern? What is ingrained into our very souls to do at this moment? Chop the power and drop flaps, right? You betcha. We are going to make radical changes to that airplane right when we have lousy visibility, and even if we have shot many actual weather ILS approaches, our total time in that visual environment is measured in seconds or a few minutes. The transition to visual references is easy to do when the weather is good, but can be horribly difficult when visibility is down to a few thousand feet.The habit pattern that involves abruptly disregarding the gauges that got us so successfully to a point some 200-400 feet above the ground is a serious breach of etiquette, even more so than leaving the dance with someone other than with whom we came. Imagine, if you will, what happens if you chop the power and drop flaps before you are even over the approach lights. Your descent rate suddenly increases and you sink into those scattered clouds that formed due to the rain cooling the air near the ground. On glideslope, you had spotted the runway through the openings in those clouds, but you changed the variables with your precipitous action, and you've just gone IFR again. Only now you are slower, your descent rate is higher and you've dirtied up the airplane. What are you going to do, Popeye? Can you make a go-around successfully? Can you transition to instrument references, stop the descent rate, get a positive rate of climb and reduce the flap deflection to something that is consistent with an aggressive climb before things come to a stop, accompanied by a loud thud?

0)]Let's play the scenario once we have acquired some ice. Chances are, the POH makes it clear we should use no flaps when landing with ice on the airframe. We've seen the very frightening NASA videos on tailplane stalls, so we know, intellectually, that adding any flap deflection at all, or slowing down, could very well jam us into the planet. We intend to do it right: We will fly the ILS flaps up, leaving the power at the setting that has worked to nail the glideslope, until we actually roll the wheels on the runway. We've read the accident reports about pilots, carrying ice, who made a power reduction in the flare and stalled the airplane at about 10 feet above the runway. Nevertheless, we break out of the clouds, spy the runway and "bam." Habit patterns from hours and hours of practice ILS approaches grab us by the neck and we make the immediate power reduction, flap extension and a dive for the end of the runway. The last sight we have, as the tail stalls and the airplane pitches down, is a windshield full of snowy ground.Believe it or not, for an air carrier, getting below the glideslope (a "fly-up" indication) is a violation of the FARs. Those folks are required to stay on or above the glideslope all the way down to the flare. That's part of the reason those fixed distance markers are on ILS runways; if you shoot the ILS correctly, they are approximately where you will flare. Yes, they are not right at the threshold. Big deal. There are no ILS approaches to terribly short runways. Plus, and I've checked, there are no prizes for landing short after an ILS.Have you looked at your Jepps lately? They show a threshold crossing height, that is, how high you will be over the end of the runway if you are on the glideslope. It's ordinarily about 51 feet. In a Cessna 172 or Piper Warrior, were you to be slowed to 1.3 Vso at the threshold, 51 feet up, and then land per the book, you'd stop about 2,000 feet from the threshold. That leaves at least 3,000 feet of runway on virtually all ILS runways.As a result, why not at least wait until well over the approach lights -- or even until crossing the threshold -- before getting all excited and wiping out the power and tossing out the flaps? Waiting, being a little patient, before slowing the airplane, eliminates the nasty risk of crashing short of the runway. I've heard pilots express worry that they will run off the far end of the runway should they delay decelerating to landing speed until over the threshold. It won't happen. We'll talk more about speed in a moment, but in general aviation piston pounders, even in the highest performance twins, getting down and stopped on an ILS runway from a 50 foot threshold crossing height is simply not a problem. After all, the jets do it daily and they're going a lot faster than us bugsmashers, and they don't touch down until at or after those aiming marks on the runway. The risk we face is crashing short of the runway or losing control in ice, so I respectfully suggest that we need to amend our habit patterns and learn to wait a bit, to leave things as they are after breaking out of the clouds and keep flying that ILS until we cross the runway threshold; to not get excited on those nice practice days and dive under the glideslope when we spot the runway.

On the ILS, fast is generally good, to a point; and, yes, we are going to slow down a lot and use all the flaps, before we touch down, because fast is a very bad thing on touchdown.Back in about 1974 a buddy of mine took his instrument checkride with an examiner who had become enamored of "stabilized" approaches. The examiner had heard or read something about how the airlines were shooting the ILS with airplane stabilized at an approach speed of something on the order of 1.3 Vso. Accordingly, the examiner required that my friend shoot the ILS in a Cessna 172 at 70 KIAS with full flaps. It was a nightmare. Can you imagine vibrating along at 70 knots indicated in a Cessna 172, full flaps hammering in the wind, as you carry almost full power to hold the three degree glideslope? Fortunately, most everyone in general aviation has figured out that stabilized approaches are a procedure for jets, not for airplanes with propellers, because of the differences in throttle response times. Early jets had a lag between shoving the power lever forward and the engine producing usable thrust that could be measured on a calendar (not nearly as noticeable with modern fanjets). Accordingly, approaches were flown in a "dirty" configuration, at a high power setting, so that a go-around was possible. The problem was that those who didn't understand that propeller-equipped airplanes didn't suffer from the power application lag as did the early jets, figured that if an airline procedure was in place, it had to be good for everyone else, no matter what type of airplane was involved.Fortunately, more thoughtful minds prevailed and, although one still finds pilots who insist on flying an ILS at 1.3 Vso in a propeller-equipped airplane, most have gone the way of the flat-earth crowd.

The speed to fly on an ILS is an amalgam of many variables. For pistons and turboprops, there is nothing wrong with scooting down the glideslope at 100-120 KIAS because it is so easy to get rid of that speed prior to touchdown. Yes, I have harped endlessly in this column on carrying too much speed on final, because the accident data I've seen indicates that landing accidents are most often the result of pilots carrying too much speed rather than too little. OK, let's break that down. The "too much speed versus too little" has to be measured at some definitive point to make sense. For the sake of establishing a datum, and to try and define the subject, the point where the speed for landing -- versus the instrument approach speed -- might best be evaluated just before the flare, something on the order of one wing span above the ground. For the sake of brevity, I'll assert that it is at that point the airplane should be at 1.3 Vso, more or less, with one-half the gust factor added on, but absolutely no more. Any faster and the bad things that happen to pilots come about as they float and drift and lose directional control and jam nosewheels onto the runway and generally do weird and painful things. Slower is not necessarily a bad thing, as it can be rectified with a dose of throttle to stop the sink rate. And, I agree, you've just heard me advocate going like clappers down final on the ILS while saying to come into the flare at 1.3 Vso, so what's a mother to do?Because this discussion applies to turboprops and pistons -- those airplanes with those propeller things on them -- we are flying air machines on which we can change our speed pretty darn quickly. So, if we whistle down the glideslope at 100 KIAS in our Cessna 172 or Piper Warrior or Diamond Star, we will be over the approach lights at a bit under 100 feet agl and over the threshold at about 51 feet agl. All we need to do is smoothly pull the power back to idle, beginning our deceleration to the 1.3 Vso target, while either continuing to descend, or holding altitude. It's OK; there is a bunch of runway in front of us and we aren't going to get hurt being on or a bit above the glideslope. Depending on the type of airplane, the airspeed will either be in the white arc, or will drop into it within a few seconds. At that point, full flaps are selected and the airplane is retrimmed. It's not a big deal. We're above the runway, even if the visibility is but a half mile; we can see perfectly well to make this final configuration change and we have passed the true area of exposure, crashing short of the runway. The speed bleeds off nicely, to 1.3 Vso, as we roll the trim wheel and let the airplane settle to flare height. The fixed-distance markers go under us, and we raise the nose and touch down smoothly, using maybe half of the runway, while keeping the risk of having something bad happen at a minimum. We haven't made any major changes before we got to the runway threshold; we waited until we had a good visual reference under us -- the runway lights, something we've looked at for years -- then the process of slowing the airplane took place under circumstances where we could see our landing area clearly. It provided a good reference as to whether the wings were level, and we had plenty of room land.

Flying slowly down the ILS has no up side. Let's say we fly the ILS at Vy, a speed faster than 1.3 Vso, yet for a Warrior, only about 80 KIAS or so. In most of the country, we will tie up traffic behind us very effectively because everyone else is going faster. Even assuming that we don't give a fig about anyone else in the system, an 80 knot ILS means that it's going to take a long time, adding to the work load. It's just not easy to fly an ILS when going slowly, and, should we have to go around, we have no energy at all for the process of initiating that go around and getting away from the cold, hard ground. Pulling up will not result in any significant climb, it just means that the airspeed immediately drops below Vy and we have to wrestle getting the power in and accelerating to Vy while trying to proceed upward. Our little airplanes climb poorly enough as it is. In some parts of the country they may or may not make the minimum climb gradient needed to avoid terrain or obstructions. Why make it hard on ourselves? Come down final with some margin above Vy. For control and ease of the approach, faster is much better. One hundred sixty knots in a Lear makes the ILS a piece of cake, so try 100 knots in a 172 or Cheetah or Warrior. It is far easier than slower. And assuming that a miss is necessary, a gentle tug on the wheel results in a dramatic rate of climb. Slide that throttle forward and we'll see 1,000 fpm for a moment as we slow to Vy. We don't have to struggle to accelerate to Vy while trying to convince that sled to climb as we get rid of approach flaps. We'll be several hundred feet up, climbing nicely, by the time we decelerate to our desired climb speed and our risk of running into something is nicely reduced.In my humble opinion, what we do over the years in training and in situation normal ILS approaches matters hugely in our longevity. It's the old saw: On those days when we are flying around as if it doesn't matter, it really, truly matters. If we are in the habit of flying precisely, we will do so when the weather is bad or the airplane has a problem. If we are used to flying an ILS relatively fast, then it's not a new experience when we have to do it for real.

2)]So, OK, when do we have to fly an ILS fast? Being as it is autumn, the answer that comes to mind is, "When we have ice on the airplane." Suffice to say, when in cruise, the concern we have with ice is on the wing. Once we slow down a bit for the approach, the attention shifts to the tailplane. It ices up faster, with nastier effects, than the wing, due to its smaller leading edge radius. And when we slow down for the approach, it starts working at a much higher angle of attack than during cruise, moving it closer to stall. So we fly the approach fast in ice. Therefore, it's certainly nice to have flown fast recently, as recency of performing at task is the best indicator as to whether we will do it successfully the next time.When we have ice on the airplane, in our Warrior example, we are going to shoot the ILS at 100 KIAS, or faster if we can pull it off. Then, because we have broken in a new habit pattern, we will not make a power reduction when we see the approach lights. If we just leave things alone, while we transition to visual references, we increase our chances of surviving that icing encounter. It also allows us time to think as we approach the runway and recall what variable will be different with this particular landing: We won't slow down at all. We'll use that glideslope and localizer to make sure we get to the runway, not the adjacent taxiway, as we look at the outside world through the little peephole in the iced-up windshield. We will leave the power right where it is as we flare slightly and touch down, because we know that if we reduce the power in the flare, we will probably stall the airplane.It all works in lousy weather because we have created a habit pattern in good weather. We break out, observe and wait, because nothing is broken, so nothing needs to be done right away. We just continue to fly visually while cross checking the ILS needles. Maybe we should put a great big placard on the panel that simply says, "WAIT." We have plenty of time after we spot the runway to decide when and where we are going to slow the airplane for landing. And, no, we won't run off the end. We'll turn off about midfield and can pat ourselves on the back for flying an ILS for real and not doing anything dumb in the process.See you next month.