August 6, 2007
|About the Author ...
John Deakin is a 38,000-hour pilot who worked his way up the aviation food chain via charter, corporate, and cargo flying; spent five years in Southeast Asia with Air America; 33 years with Japan Airlines, mostly as a 747 captain; and is now flying Gulfstream IVs full time as a charter pilot. He also flies his own V35 Bonanza and is very active in the warbird and vintage aircraft scene, flying the C-46, C-131, DC-3, F8F Bearcat, Constellation, B-29, and others.
All the rest of John's "Pelican Perch" columns are available here.
After this column was written, two more Mustangs were involved in a crash during EAA AirVenture at Oshkosh, with one fatality. Readers may ask whether these accidents were related.
I don't think so. Both pilots in the Oshkosh crash were well-experienced, although I have no idea how current they were, possibly the key question. Formation flight is very demanding, and like carrier landings and low-level aerobatics, it requires skill, constant practice and currency. I am always shocked at how rusty I feel doing formation after a long layoff and I have to approach it as if starting all over again.
Many think the #2 airplane lost sight of #1. It is not impossible, but I consider it highly unlikely. When in a tight formation, the aircraft you are flying on is your whole world; you don't even want to blink.
You never look away. If you need to check something in the cockpit with your eyes, you move away a bit. More than one wingman has followed the lead right into terrain, for that reason.
Losing sight of the lead is the second-biggest sin in formation flying (the first is hitting someone) and is cause for an instant turn, pull-up, or whatever to clear the area.
Aircraft #1 touched down first and most likely would have had the throttle closed, making a normal landing. Aircraft #2 was higher, still flying, probably with a bit of power kept on just a second or two later. In fact, half a second before the collision, the #1 prop appeared to be turning more slowly than #2 (strobe effect, very brief), so there was not only a closure rate, but more power on #2. By the time he caught it, there was such an overtaking closure rate set up, with no way of stopping it, so he simply overran #1, without enough lateral separation to clear.
It's common enough to get an overtake like that, but most pilots agree that each will take his own side of the runway and move apart laterally in the final few seconds. Spectators can't see that.
One trick in formation landings is to brief the lead aircraft to keep just a touch of power on before, during and after the touchdown.
Normally, it's a very bad habit when not in formation, but this is one case where it gives the following aircraft some margin to play with: If he closes his own throttle, any closure will be stopped quickly. Or have the lead use partial flaps, and the followers use full. Better yet, use longitudinal separation, maybe 1/4 mile or more. Even then, if the lead aircraft blows a tire (and tires do blow), there will be a major effort required by the following aircraft to avoid running over the lead.
On takeoff, the lead aircraft normally uses less than full power (sometimes quite a bit less), so that the following aircraft has a margin in which to operate.
I'd be very surprised if the #2 pilot ever lost sight of #1, but now we'll never know.
(Video clip available here.)
Update: Aug. 8, 2007
In the interests of historical accuracy, the above remains as I wrote it -- with incomplete data.
Having now seen further data, and several messages from trusted people who saw the crash, I am now convinced that it was indeed a matter of #2 not seeing #1 until too late, if at all. I do not now believe there was any intent to make any kind of formation landing, beyond the briefed "in-trail" landings.
It is not the first time I've ever had to eat crow, nor will it be the last. It's an excellent example of why we need to wait a bit longer for the data, and why the NTSB often takes a year or more to release findings on accidents.
I was awakened by a phone call from a good friend. His opening words were, "Man, am I glad to hear your voice!"
My instant response to that is usually, "Oh %$#@! Who bought the farm this time?"
He said, "There was a fatal Mustang accident at Camarillo this morning and I'm happy to hear you weren't flying it."
So was I!
I will describe the events that led to the accident, the accident itself, what I believe to be the reasons for it, and what pilots can do to avoid similar accidents, including my reasoning for a slightly different "style" of go-around.
John McKittrick, age 42, had something around 1,700 hours of flying time, including a few hours of dual in a North American T-6 in Northern California, and he had a high-peformance tail-wheel endorsement from someone in Santa Paula. He owned and flew his own King Air.
He bought a gorgeous, dual-control Mustang ("Lou IV"), fresh out of major rework by Paul Besterveld, a local pilot and entrepreneur who specializes in Mustang refurbishment, mostly as a labor of love, otherwise known as a hobby. Paul offered 20 hours of free dual in his own Harvard (a Canadian version of the North American AT-6) as part of the deal. Paul was rightly concerned that the new owner should get adequate training. McKittrick did not take advantage of that very generous offer, for he had already arranged for dual instruction in his new Mustang from Matt Jackson, a famous Reno Race pilot and a Warbird entrepreneur based in Van Nuys. He is well-qualified to instruct in the Mustang.
Is It Safe?
Now, let me say right up-front that I find all this reasonable enough. I have no trouble with the concept in general. Military pilots went into combat with the Mustang with less total time, and there were no dual-control Mustangs in those early days. I firmly believe that a talented pilot can do this safely enough with 500 hours, or even less, with proper training. On the other hand, I've seen too many pilots with more money than brains kill themselves in airplanes that are beyond their abilities. That doesn't appear to be the case this time. While I think McKittrick would have been better off to take the T-6 time offered, I can understand his eagerness to fly his new toy and get his training in the airplane he wanted to fly. With Matt in the back seat, it was a reasonable operation. Definitely not as safe as taking lessons in a 172, but given that someone wants to fly a Mustang, it's a good way to do it. The insurance company demanded at least 25 hours of dual instruction before he would be insured for solo flight, and they went well beyond that.
(Gone are the days when just about anyone could buy a warbird and go fly without insurance, or with just liability insurance. Hull insurance is now required right along with liability, and I'm told the premiums are roughly 4% of the value.)
At the other end of the scale, an experienced warbird pilot can familiarize himself with the Mustang cockpit, read the manual, and go fly it safely with no training at all, for it is among the easiest of the warbirds to fly. Since dual-control Mustangs are fairly hard to find, most did it just that way in days gone by. But most of those pilots were well-seasoned, with time in many different types. They had "salt," mostly. Today's crop of pilots seems to me to be excessively book-oriented, slaves to the regulations, and quick to cite "policy" and "SOP." Flying comes with an infinite set of variables, and most of them are not covered in writing.
While the Mustang is a pussycat in many ways, it does have at least one very nasty tendency in common with all the big-engined fighters, and that is control at low airspeed and high power. This "feature" killed a lot of pilots in the military, and continues to do so today. This accident is almost certainly just one more example.
The training was indeed accomplished and, by all reports, McKittrick did very well during 31.5 hours of dual instruction with 64 landings. Matt has stated to me that it was "15 hours too much," but insurance requirements dictated more, and McKittrick himself had a great attitude, not pushing for solo at all.
Time for Solo
On a lovely Sunday morning, the two of them flew to the Camarillo airport, which is perfect for this sort of thing. They taxied to the CAF ramp, where Matt advised the tower that he was releasing a pilot on his first solo in the Mustang, and he climbed out, telling McKittrick to make one pattern and taxi back. This is the normal procedure, which allows a de-brief and comments.
Matt told me that the takeoff was fine, but that McKittrick snatched the gear up very, very quickly, which he considered odd, for he had not taught him that show-off stunt, used only in airshows when a roll after takeoff is intended. Matt made a mental note to de-brief the point.
The pattern was normal, but Matt noted that, when turning final, McKittrick increased RPM to somewhere around the "book value" of 3,000, contrary to the 2,700 that Matt teaches. (The difference in sound is very distinctive). McKittrick had commented to Matt that morning that he had again reviewed the manual, which calls for 3,000.
Otherwise, the approach and landing were all fine, ending with a normal wheel landing with the tail slightly low, just as he was taught, and "by the book." Perfect, except that it sounded like he did not fully close the throttle to idle, even after touchdown. That is normally done in the flare or slightly before, but some pilots will touch down with a little power when flying a new airplane. Not a bad technique to "cheat" a little, but it is a bad, bad habit if used all the time.
After rolling along for perhaps 100 feet (a second or two), Matt noticed the tail come up a bit, perhaps to "pin" the aircraft to the ground. One local Mustang pilot theorized that the tail may have come up unexpectedly because Matt was no longer in the back seat. Matt is a very large man, and the back seat is aft of the CG, so this may or may not be a valid comment. I wonder if perhaps McKittrick felt the tail might be on the way up more quickly than he expected, with the nose coming down, and feared a prop strike? This probably was not a problem, but there is no way of knowing what he was thinking in the excitement of his first Mustang solo. The small amount of power he was still getting would also have made the controls more effective than he was used to at that point, perhaps exaggerating his pitch inputs.
At that point, he jerked the stick back abruptly, with the tail wheel striking the ground hard enough to make an audible bang. This placed the airplane in a perfect attitude for flight, and since there was just enough speed left (and a touch of power), the airplane lifted off again, attaining a height of 10 feet or more. The normal recovery for this would be to hold or regain the three-point attitude, allow the aircraft to settle, then pull the stick fully back to make a three-point landing. A tiny shot of power would have been appropriate to blow a little air over the wings, create a little lift, and cushion the landing. That would be a perfectly normal bounce recovery, which McKittrick had done many, many times with Matt in back.
For whatever reason, McKittrick jammed on full power, absolutely the worst thing he could have done, in my opinion. The airplane pitched to a very nose-high attitude, rolled to the left in the classic "torque roll" and hit the ground almost fully inverted, killing the pilot instantly. There was no fire.
The NTSB and FAA reportedly investigated the accident the same day, and since all rules were complied with, all paperwork was in order, and no one else was hurt, with no "collateral damage," the case was closed within hours, and will most likely be ruled some variation of "pilot error."
That's probably true, but I think there's more to it. I would say, "Pilot error caused by half a century of improper training in the go-around maneuver."
The Usual Procedure
The nearly-universal training for go-arounds is:
- Full power, as quickly as possible ("balls to the wall");
- Pitch up to some angle that will give maximum climb (angle or rate);
- Retract flaps when able; and
- Retract gear when "positive rate" is attained.
I have previously written about go-arounds in Baby on the Runway. That column adequately covered items 2, 3 and 4 above, but I did not cover item 1 as well as I should have. The column was written with the average GA (General Aviation) airplane in mind, and also for jets, and with that in mind, I stand by it. It was not particularly relevant for high-powered warbirds,
Prop Lever(s) Forward?
Conventional wisdom suggests that the prop control should be fully forward (full high RPM, a.k.a., low pitch, a.k.a., fine pitch) for landing, with widely differing advice on just when to do that. Some will do it well before entering the traffic pattern (noisy!), some will suggest in the pattern somewhere, and a lot of the warbird pilots strongly state that it should be pushed fully forward no later than on short final. All have strong opinions, and typically, many consider "their way" to be the only correct way. Any way will work, provided the pilot knows the airplane, and what to do, but I don't think this is even close to the best way.
I have disagreed with that for more than 40 years, and I do not do it myself. I have been guilty of tolerating the procedure when instructing others, because it has become so strongly ingrained in the fraternity that it's just not worth the hassle to try and change more than 50 years of hard-core training, poor checklists, and group-think. I've also faithfully pushed the prop forward for landing when receiving instruction, or on the check rides I've taken, because some would consider it a "bust" not to do so, and few will listen to "logic" that is contrary to well-entrenched "conventional wisdom."
In recent years, leaving the prop set for cruise all the way through the landing and rollout has gained some favor, and I've been suggesting it as a better way. For one thing, it is quieter for the neighbors.
With high-powered airplanes at low airspeed, pushing the prop forward for the landing is a deadly, dangerous habit, and that habit probably helped to kill many pilots. I'm sure McKittrick received instruction in low-speed/ high-powered flight, and had been told many times to be gentle with power at low airspeed. All of us who fly these magnificent old airplanes know the danger, and most of us have scared ourselves a time or two getting a little too frisky with the throttle when it is hooked to somewhere around 2000 horsepower. There are at least three situations that will trap the new warbird pilot, and even a few experienced pilots have been killed when they forgot in the heat of the moment, reverting to instinct.
The first trap is getting a bit too low and slow on the base-to-final turn. Very insidious, very deadly in any aircraft. Impossible to demonstrate or train for it, the only training is to stay out of that situation.
The second is the last-second go-around that is entirely unexpected. This can be, and should be, a major training item, and McKittrick had a lot of these.
The third is unexpectedly becoming airborne again after landing, either from a sudden gust of wind, or a bounce, or (as in this case) getting the nose too high after touchdown.
Any of these will rattle the pilot, even more so in a strange airplane, and far more so in a very high-performance thrill-ride like the Mustang. There is an incredible adrenaline rush under these conditions. In these situations, when something unexpected happens, pilots will instinctively revert to the way they were trained from their student pilot days, and simply go "balls to the wall," with a gut feeling of "Oh, %$#@!, let's get out of here!"
That's the last mistake they'll ever make, and things will go so bad, so fast, they'll never even know it. We need to change that instinct, I think.
This is not limited to warbirds. My Bonanza with the Robertson STOL kit will fly so slowly that, at full power, it takes absolutely full right rudder, and a fair amount of right-wing roll control (there are no ailerons) to hold it straight. I'm a little surprised the FAA certificated it. When I upgraded from the 285-hp IO-520 to the 300-hp IO-550, it got a little worse, and I will not haul it off the ground at minimum speed. I'm sure there are other examples like this in the GA fleet. Some of the "hot" little experimentals like the Lancair IV-P show this tendency, so it's not as rare as you might think.
A brief refresher on this important subject may be in order here, and then I'll get back to the accident, and go-arounds in general.
There are at least four "forces" that cause this problem. All four produce a left turn/roll/yaw tendency in aircraft with the prop turning clockwise (viewed from behind). Of course, the British, bless them, make engines that turn the other way, and those airplanes turn/yaw/roll to the right.
I will limit my comments in this column to the clockwise-turning prop airplanes, unless otherwise noted.
First, There Is Torque
The word "torque" is often misused to mean the total effect of all four forces. In normal conversation this may be useful, but keep in mind that it is a misuse.
In reality, torque is a simple reaction to the engine pouring all that power into the prop, and it is a pure roll force. Open the hood of your car and race the engine a little. You'll see the engine try to roll in the opposite direction from rotation, and quite vigorously. All that motion is from the horsepower accelerating the crankshaft one way, with the reaction twisting the engine the other way.
In the airplane, if 2,000 hp is turning the prop one way, that same 2,000 hp is trying to roll the airplane the other way ("equal and opposite reaction"). In most GA aircraft, this force is not very significant because the horsepower is (relatively) low, the wings are relatively long (good leverage), and roll control is generally pretty effective. The effects from this force are mostly lost within the other forces. But in something like the Bearcat, the power is much higher, the torque is correspondingly greater, and the wings are shorter, giving less leverage. I assure you, the roll is there, and it's very noticeable. Left tires wear out much sooner on the Bear, and most pilots start the takeoff roll with full right aileron, regardless of other factors. It just feels better. Without that, it really feels like the airplane will lift the right gear right off the ground, and roll it over on the runway. It probably won't, but I'd rather not give it a chance.
You can demonstrate gyroscopic force by lifting one wheel of a bicycle clear of the ground, and having someone spin it. Try tilting the bike left and right, and you'll feel the gyroscopic effect. This effect is 90 degrees to the input, and proportional to the mass of the spinning object, the rotational speed and the effort put into changing the plane of rotation. It's very powerful on the bike wheel, even at low RPM. The force is immense when a pilot changes the plane of rotation of the prop by lifting the tail, and it causes pure yaw. This one is most noticeable on the Mustang because everyone seems to raise the tail to a near-level attitude on takeoff. Raising the tail quickly translates to a sharp left yaw as the tail comes up. I'm aware of one "expert" who insists on lifting the tail all the way to a nose-down attitude, a really dumb move. If the tail is lifted a bit slowly, no problem. But if someone lifts the tail prematurely, and too quickly, he'll be heading for the buckwheat to his left before he knows it. Best in all the high-powered airplanes is to just lift the tail a little, and slowly, perhaps half-way to the level attitude, and let the airplane fly off. Many who fly the Bearcat simply make three-point takeoffs, and this works because the airplane accelerates so quickly. With this technique, there is no gyroscopic effect to speak of, because the pilot is not trying to change the plane of rotation of a 1,000-pound, 12-foot flywheel turning at well over 1,000 RPM. Besides, anywhere near a level attitude in the Bearcat on the runway is likely to result in a prop strike.
At low airspeeds and high angles of attack (of the wing), the descending prop blade has a much greater angle of attack than the rising blade, creating much more "lift," which translates into more "forward pull" on the right half of the disk, or pure left yaw. As the airplane picks up speed, and the tail comes up (taildragger), the left-pulling force from p-factor is reduced. On the other hand, the nosedragger has very little p-factor while accelerating on the runway in a level attitude, but the effect will increase as the nose comes up to flying attitude.
Props impart a clockwise swirling motion to the slipstream, which corkscrews its way around the airplane. When this spiraling slipstream hits the vertical fin and rudder, it will be pushing against the left side, tending to yaw the airplane left. The part of the slipstream that passes under the tail finds nothing to push, so there is no opposing yaw effect. Personally, I'm a little doubtful this one has much effect, but I mention it for completeness. I cannot recall ever seeing good data, or real evidence, supporting this theory.
There have been many tricks to counter these effects. The Bearcat is the most modern prop fighter, and Grumman knew all the tricks by the time it was designed (too late to fly in WWII). I understand the engine is canted something like six degrees to the right and two degrees down, nicely reducing the "left pull" so common to the rest of the fleet.
Most aircraft have a vertical stabilizer that is mounted to the fuselage with the leading edge canted, but that is for efficiency in cruise, not so much for control on takeoff.
Mustang pilots usually pre-set six degrees of right rudder trim for takeoff, in the badly mistaken belief it will help to control the airplane on takeoff. It does not, it actually reduces the maximum available effectiveness of the rudder. It does create a lighter feel in the rudder, which is a horse of another color entirely. If I'm taking off in a left crosswind, I'll set left trim, which means I'll have to push harder with my right foot, but I'll have more "available rudder." Why the right trim, then? It's simply a good preset position for the climb portion of flight. By the time the aircraft is cleaned up, the power is reduced a bit, and the speed is 160 knots or so, the rudder trim will be perfect.
(As a side note, it is generally not correct to say "degrees" when setting trim tabs. They are usually simple arbitrary marks near the knob or handle. Some are further marked with numbers for easy reference. Mustangs may be one exception: They are definitely marked with numbers and the degree symbol. I have not measured the actual deflection with a protractor.)
The Spitfire is an interesting airplane, and not just because the prop turns "the other way." All of the above effects make this airplane turn/roll/yaw to the right. There is so much power available in the Mk XIV that the pilot runs out of rudder control at slow speeds during takeoff before attaining full power. It's actually possible in this airplane to apply power until the rudder hits the stop, then control the airplane left and right by changing the power. More power, it goes right, less power, it goes left. As the speed increases, more power can be applied, probably attaining full power right around liftoff. Update: Aug. 10, 2007 -- A number of alert readers have taken me to task over the prop rotation on the Spitfire. The one I'm most familiar with is the Griffon-powered Mk. XIV, which does does turn counter-clockwise (as seen from behind). The Merlin-powered Spitfire props (the vast majority) turn clockwise.
(Another side note: If the pilot has to think about which rudder pedal to push on takeoff, or how much, he's dead meat. The feet need to move instinctively, without conscious thought, "as needed." At most, the thought process will be "Gee, this feels funny," while barreling down the runway. With a few takeoffs, it won't feel funny anymore.)
Some aircraft (like most Mustangs) have a locking tailwheel that locks to the rudder when the control stick is not fully forward. This gives the pilot a very positive "tailwheel steering" during taxi with the stick neutral or back, and this is also helpful early in the takeoff. Some T-6s also have this locking system. But it can also bite, as when trying to hold the tail up during landing (bad technique). That last inch or two of forward stick movement also unlocks the tailwheel, and there will be very little directional control when the tail touches down.
Tailwheels that lock straight fore and aft do not help with control (contrary to common belief). The system is merely to keep the tailwheel from shimmying, and to keep the assembly straight as it retracts. Many airplanes have no locking tailwheel at all, and pilots will not notice any difference in directional control. I've even heard rumors that if pilots forget to lock the tailwheel on aircraft so equipped, they won't even know it. Not that I've ever done that myself. Of course not.
Most of the fighters can handle full power early in the takeoff roll, provided the pilot is aggressive with the rudder, but in all the high-powered aircraft, it's best to be gentle with the power. It's easier on all those moving parts in the engine, too. By being gentle with the power, the pilot buys precious time to react to the changing forces leading him astray.
All the above claptrap aside, it comes down to this: If you apply too much power too quickly, you need to be very fast and aggressive with the flight controls, and in some cases, there simply may not be enough control available. It is extremely difficult and dangerous to train for this close to the ground, and most training is to stay away from that area. McKittrick probably had full flaps or nearly full flaps down (properly so), and this would have allowed the airplane to still be flying at an even slower speed, which further reduced flight control effectiveness. The little bit of power he kept on would have made this even worse, allowing a slower airspeed before the stall. "Pouring on the coal," a perfectly natural thing to do in light of GA training in general, made the situation impossible, and even applying full control-deflection wouldn't have been enough.
Let me belabor this. Matt Jackson certainly gave him instruction on applying full power suddenly, and how to control it. He certainly gave him instruction on recovery from bounced landings, either a small shot of power to "cushion" the next touchdown, or gently adding power for the "go." Heck, no one can do that many landings in any airplane without bouncing! What he could not train for is the sequence of events peculiar to this event. A bit of power still left on, the airplane unexpectedly rising into the air again at a very low airspeed, adrenaline at full flow, and perhaps a thought process of "I'm running out of runway!" Add to this the fact that it was all unexpected, right out of the blue. Recovery from that would have taken a very gentle application of power and a delicate touch on the controls, perhaps even deliberately "skipping" (deliberately bouncing) the next touchdown while the power was still coming up for a go-around.
Once the airplane was high in the air, with the extreme nose-up attitude reported, nothing could have saved the airplane. Crashing straight ahead might have saved the pilot.
Now, finally, back to the subject of how we can do better go-arounds, and reduce these risks.
First, let me say that we in the aviation community have a terrible tendency to work out some "universal method" that works best for the new student pilot in a Cherokee, and we use that forevermore without thinking much about it. Much of what we "know" comes from the massive flight training effort in World War II, when the goals and dynamics were very different. We then tend to apply that to all situations, in all aircraft, with all pilots. It's called "primacy," because, "You learn best what you learn first." That's fine for good habits, good techniques and good procedures, but can be lousy otherwise. The classic go-around is one example, in my opinion. If you're the "one size fits all" type, you probably won't like the rest of this column.
Don't Trim for Landing
Don't get in the habit of trimming for the landing. It may look cool, and it may help in the flare if you have a really weak left arm. You may even convince yourself that it helps you make better landings, but it's a killer in some airplanes on the go-around. Most airplanes will pitch up all by themselves when power is applied, some pitch up really hard. Having excessive nose-up trim makes this much worse, and may well make the airplane unflyable. If you experiment with your airplane, and find this effect benign, please don't fall into the trap of doing it because it works. Someday you may fly a different airplane in which it doesn't. Stay off the trim once over the airport boundary at the latest.
Flaps Don't Have to be Down Full
Think about the use of flaps. Just because there is a position marked "FULL" doesn't mean you should use it. Flaps don't have a major effect on some or even most GA aircraft, and most will land better (and just as short) with half flaps, or even none at all. Partial flaps will usually allow a slightly more nose-up attitude at touchdown, which may be beneficial on airplanes that tend to land on the nosewheel first (Comanches, Mooneys). There is probably no good reason to use flaps at all on any ILS approach in small GA aircraft. Some may say it makes the airplane more stable, but in reality, it simply feels different. Once you become accustomed to not using flaps for the ILS, the "feel" will be just fine. You won't need them for the landing either -- all ILS runways are more than adequate for a flapless landing. Take a look at the stalling speed of your aircraft with and without flaps, you may be surprised at how small the difference is. Try it different ways, enough times to get over the "strangeness factor," and see how the airplane flies. Landing with less flap will make speed control a bit more important, but it decomplexifies the go-around. (I know that's not a proper word, but I like it, and it's my column.)
(The legendary Paul Soderlind disagreed with Boeing so strongly on the use of 40-degrees of flaps on the 727 that he had all the Northwest 727 flap lever tracks bolted to prevent extension beyond 30 degrees. It didn't change the landing requirements more than a couple of hundred feet, and got away from a number of bad things. He's gone now, and I miss him a lot. We used to get into some real shouting matches in email, which we both loved.)
Learn to slow down for the touchdown, and plan for the touchdown 100 or 200 feet beyond the threshold. (Never aim for the end; you'll stub your toe, one day.) I ask you, how many accidents have you heard of where the airplane stalled short of the runway? Now, how many have you heard of where the landing was "hot and long," and the airplane went off the far end of the runway? The big-iron pilots are terrible at this, with a really lousy record, sometimes on nice long runways. Why? Because we're all landing too fast. Pick your point on the runway, and put the airplane there, at the proper speed, which is just above the stall in most GA airplanes. A bit faster in jets, but well under Vref. The throttle should be fully closed at or before touchdown. A touch of power in a strange airplane, or after a long layoff, is a good method of "cheating," but it's a very, very poor habit to use it all the time.
At Last, The Prop (And My Point)
Seems such a small thing, but this really makes a difference: I suggest that you do not make a habit of pushing the prop all the way forward for the landing. In fact, I cannot think of a reason to do it at all, unless you're much too fast on final, and need the drag ... when a go-around is more appropriate, anyway.
Is "prop forward" needed for the go-around? No! If you leave the prop set for some sort-of cruise RPM, you're all set to return to cruise manifold pressure, and that's more than enough power to arrest the descent and start the go-around process in just about all cases, except anemic airplanes at high elevations (where full throttle won't kill you anyway).
Key point; With the prop set for lower RPM, if you do panic and jam on full throttle (as McKittrick did), the low RPM will not let you get anywhere near full power, although it will probably give you more than you need. This means you will not get the "torque roll" so strongly in high-powered airplanes, and you will not get the strong nose-up pitch we see in so many airplanes, especially when trimmed too much. It won't scare passengers nearly as much, either, for it makes a very gentle transition from a normal approach angle to the go-around.
Gentle transitions are good. The usual go-around (from decades of training and lack of practice) are all too often a slam-bang, semi-botched maneuver. Let's slow it down, folks! None of us do them enough to get away with "quick!"
If you get "too much" manifold pressure with low RPM, it won't hurt a thing for a short time, even in "delicate" engines like the big Merlins. It takes time for the engine to heat up enough to produce detonation, the big fear. An overboost for 10 to 30 seconds isn't going to hurt a thing, in my opinion. The detonation many fear is at power settings well beyond any sane limits. (Full throttle -- at the physical stop -- in the normal Mustang is 61 inches at 3,100 RPM, and the Reno racers run 120 inches or more, and 3,600 RPM, and probably in detonation all the time.)
Matt taught McKittrick to use only 2,700 RPM in the pattern and for the landing, for some or all of the reasons above. And he taught the use of only 46 inches of manifold pressure for go-arounds and climb, also to get a more benign response. We'll never know why McKittrick chose to use 3,000 RPM for the fatal landing, and probably full power. Perhaps he felt "The Book Knows Best," and decided that, since it was his airplane, and he was now in command, he'd follow the book. Matt's no dummy, and he's got several thousand hours in these airplanes, much of it under "demanding" conditions, to say the least.
The New Procedure
My suggested go-around in any airplane, in all but the most difficult conditions:
Add throttle gently. At a minimum, you want enough to stop the descent, fly level, and maintain speed. You should probably limit yourself to roughly cruise power for that initial power setting in most airplanes. No need to be exact, no need to even look at the manifold pressure, just shove in "enough" by feel.
Fly level. Do not allow climb for a few seconds, until the airplane and your mental condition have stabilized and caught up. If you are not descending, and not losing speed, you're in fine shape. Take your time; this is the point where most pilots get rushed and botch the whole process. There is no hurry. You will almost certainly need to push on the yoke to hold level flight in almost all airplanes, and in some you will need to push hard. Re-trim as needed, while you evaluate whether you need more power.
(For those who use "canned" power settings for approaches, remember the very small difference in MP between "level flight" and "normal descent?" Three inches or so? You don't need much to stop the descent!)
With this modest addition of power, you won't get nearly as much pitch up, or need the trim change you would if you'd gone to full power. This is a good thing. If you do need more power, push the prop control forward gently, then add throttle as needed.
When the power is set to your satisfaction (enough to produce at least level flight, or perhaps a little acceleration or climb), pull the flaps up to the normal takeoff position while still flying level, or perhaps starting a very slight climb if the airspeed is coming up. If you've done the above, you will not need to "milk" them up; just select the position, and adjust pitch as needed as they come up to maintain the level flight, or begin the climb. As the drag goes away, the speed will come up faster, and you'll be right back in the normal situation you see right after takeoff. No rush, no fuss, no scared passengers, no rattled pilots. Do not allow climb until you have the speed, and the flaps up or coming up. Do not go for the insane Vx or Vy speeds -- they are much too slow.
Leave the gear alone until you're sure you will not contact the runway. The timing for retraction is up to you. I do it fairly early, but some like to leave it down for a time. Do not make the common mistake of rushing to get the gear up; it's the least important item of all. Besides, you want it down in case you do happen to touch down.
Finally, if you do a last-second go-around, a "skip" or a "bounce" is perfectly acceptable, and is not cause for a "bust" on a check ride, or criticism at all. Perfectly normal. Of course if you do a missed approach from 200 feet and bounce your wheels on the runway, you've done something wrong!
To summarize, the procedure becomes:
- Add enough power to stop the descent and maintain speed;
- Catch your breath, re-trim as needed, evaluate power, add more if needed;
- Flaps up, pitch to maintain altitude for a few seconds (do not climb);
- Add more power if needed, and begin the climb;
- Gear up when no contact with the runway will occur.
If you allow yourself to get rushed, you will botch it. Any pilot will.
Slow it down, take your time ... and be careful, up there!
More from AVweb's Pelican is available here.