Takeoff: the riskiest three minutes

But a 30 second reflection can sure help…

There has been a rash of takeoff accidents featured in the news. That cabin-class Cessna hitting the trees in Alabama was dramatic, as was the footage of the Beech Duchess in a yard in Florida. There have been a lot others and when I read of these I think about how unforgiving airplanes can be if you fly away without the old ducks all in a row.

My thoughts turned to the cockpit of my P210, on the ramp at Epps Aviation in Atlanta. Another pilot was flying, he started the engine, got a clearance to Savannah, and even though the airplane was parked on a slight downslope it didn’t start rolling with the application of a little power.

Duchess crash in Florida
Takeoffs seem easy, but they can quickly go wrong, as this Duchess pilot found out.

Both pilots had the same certificates and ratings and, as a matter of fact they had the same name. But when there is a transgression like that it goes down as sins of the father so, as fathers do, I started barking orders.

There was really nothing to do but tell ground control we’d be a minute, do a complete shutdown, get out and pull the chocks, and then start all over to get everything humming again. Then I did something I usually did but had omitted this morning. I said we’d just sit for a brief period and reflect on that screw up and wonder if we had left anything else undone.

No, leaving a chock under the nosewheel is not a life-threatening event, but it can portend bad things. The lapse comes from the same place that a neglected control lock, or inadequate sump draining, or not double-checking the fuel, or not properly latching doors comes from. And those things can all hurt.

Takeoffs are wonderful maneuvers and I never failed to think of it as pure magic when the weight of the airplane shifted from the ground to the wings. It’s still the same airplane but when it flies it comes alive.

That’s the happy part of taking off. What comes next is a period of flight with few options and where any problem can quickly become a serious problem.

One source I looked at indicated that ten percent of the fatal accidents happen on takeoff. Because the period of time is short relative to the rest of the flight, this suggests that the risk is quite high. Many or most of the things that lead to trouble in the first three minutes of flight can be anticipated and that is why it is important not to rush through the pre-flight work.

It is also probably safe to say that most accidents that occur in the first ten or fifteen minutes of flight can be traced back to something neglected before flight.

I’ll give you another anecdote that relates to this. The ramp this time was at Indianapolis, where I had stopped for fuel on my way to Wichita.

It was snowing when I landed and the snow got a lot heavier while the line folks gassed my airplane. It was almost like a snow burst. I had flown through one of those once and it was like having snowballs thrown at the airplane. Whoomp is what it sounded like when one of the big hunks of snow hit the airplane.

I surveyed the airplane and it was obvious that a lot of snow had accumulated in the short time it took for fueling. I knew that I couldn’t leave without clearing the surfaces.

Most of the FBO crew was at the terminal getting snow off of airline aircraft. The most help I got was the loan of a push broom and the promise that if I would wait for a lull in airline activity they would help out.

Cessna with snow
Yes, you do need to brush that snow off before takeoff.

I always carried spares and equipment for most eventualities in my baggage compartment. One of my staples was automotive deicer in spray cans. I heard many horror stories about damage this could do to the windows of airplanes but I used it many times for frost removal with no ill-effect.

The plan was to use the broom to get the snow off the flying surfaces and then douse everything with deicer in the hope that would keep things clear while I taxied out for takeoff.

Now I’ll treat you to an air fable. There was an airline 727 ahead of me for takeoff. Surely, I thought, the heat from his engines would help with my snow accumulation. I couldn’t get too close because the airplane was stopped and it would take a lot of power to get moving on the contaminated surface. I just wanted to get warmed up, not blown away, so I stayed back a ways. Dream on.

It was still snowing heavily. The 727 took off and then I was cleared for takeoff. I looked back at the horizontal tail and there was definitely some snow back there but I didn’t know how much.

The takeoff felt pretty normal but a minute or so after takeoff there was a whoomp, the airplane rolled left, and I corrected. Then another whoomp and the airplane rolled right, and I corrected.

The sounds were from the snow sliding first off the right wing and hitting the tail followed by the same thing from the left side.

Next, a strained voice from the 727 flight deck informed the controller that they needed to return and land. There was no mention of why this was necessary but I’d sure bet it was snow related.

I was okay so I kept going but had I returned and run into the captain of that 727 I think we would likely have agreed that our takeoffs should have waited for a lull in the snowfall.

I was preparing for a low-visibility takeoff one foggy morning. That meant moving deliberately through the pre-flight, checking everything at least twice. Such a takeoff is demanding but I never thought it particularly risky. Something being askew would change that, thus all the double-checking.

I had an epiphany when I thought about my preparation on this foggy morning. The need for careful preparation might not be quite as important on a clear day but it can still be pretty important. From that day forward I treated all pre-flight preparations equally, regardless of the weather.

Accidents that occur on or soon after takeoff actually fall into two separate columns.

When something happens after an IFR departure, it usually relates to the pilot losing control of the airplane in roll.

Takeoff by Cessna
The moment of truth.

When something happens after a VFR departure, it usually relates to power—not enough or not running well or at all. Not enough power would cover those accidents where a pilot tries to fly out of a place that is simply not big enough given the conditions that exist. A loss of control often follows a loss of power so on a visual departure problem the flaw is in pitch control. The pilot stalls the airplane after having a problem with the available performance.

There are proportionately a lot of IFR departure accidents. If I had to describe a typical accident of this type, I would go for night, a somewhat rushed departure, an unfamiliar airport, and a pilot of moderate flying experience but, given the airplane involved, a lot of experience making money. I’m not jumping on the bandwagon to poke at one-percenters but a lot of folks who reach that level do seem to feel skill in that area might equate to skill at something like flying. It doesn’t.

I have always been a proponent of, on departures in IMC, climbing straight ahead at full power in the takeoff configuration until 1,000 feet above the ground. (The landing gear would be retracted, if applicable, when a positive rate of climb is indicated.) The transition from visual to instrument flying must be made when the airplane is rotated to a flying attitude. Staying visual for as long as you can see is not the way to do it.

Doing it that way means that you can give your absolute attention to the attitude of the airplane in that period when you are adapting to the sometimes disorienting sensations of flight.

This can’t be done off some runways because of departure procedures. One I used frequently and that required a high degree of concentration was Runway 16 at White Plains, NY (HPN). The procedure there is runway heading to 800 feet, which is about 400 AGL, then a climbing right turn to a heading of 320 and climb to 3,000 feet. I always knew that it was a condition ripe for spatial disorientation so I always felt I had to be locked and loaded to do some of that flying stuff, as Goose sort of said in Top Gun.

A pilot from a prominent family had his Piper Meridian pulled out and prepared for takeoff one foggy morning at HPN. You guessed it, Runway 16 for departure. You also guessed that he lost control soon after takeoff and crashed or I wouldn’t be telling this story.

Loss of control is a hot button with the Feds now but most of the emphasis is on low speed events like stall/spin accidents. It is actually pretty easy to come up with ideas there. Hey, we’ve been doing just that in Air Facts since 1938 which just illustrates that any solution to this old problem is elusive.

Even more elusive is any answer to the question of control losses on IFR departures.

One difficulty comes because it is something infrequently done and something you basically have to learn on your own. Simulators aren’t much good as might have been illustrated one scuzzy evening. The pilot of a Cessna P210 couldn’t keep it under control and crashed a quarter of a mile past the end of the 8,100 foot long runway.

The day prior to the accident the pilot had an instrument competency check in a simulator.

Fog on runway
Foggy takeoffs are particularly demanding.

Simulators are good for a lot of things but this is not one of them because they can’t simulate how acceleration affects a pilot’s sense of what is going on with the airplane or with himself.

So how do we approach this problem? The only way is with a deal the pilot makes with himself. In that pause before takeoff the pilot has to acknowledge that the next three minutes will be totally critical and that the mindset can and will be on nothing but controlling the airplane. There might be squirrelly and disorienting sensations to overcome. That takes concentration.

What about the autopilot? This can certainly be useful but only if its use is the same as on VFR departures. Certainly if a pilot decides on the spur of the moment to use an autopilot for a low-visibility departure the chances of setting it up correctly will not be nearly as good as they would be if the autopilot were used that way all the time.

I remember one low-weather departure accident where autopilot confusion likely had a role. The airplane was new, both to the world and to the pilot, and it had a full autopilot.

The pilot got his IFR clearance, took off, wandered about a bit at low altitude and crashed. Pilots who knew him surmised that he was wondering how this thing (the autopilot) worked the whole time he was airborne. He did not have the reputation of a planner but of one who did things on the fly.

I had a simple (roll only) S-Tec autopilot in my P210 and always thought it one of the better devices to use in reducing risk on IFR departures.

One of the problems pilots have had with full autopilot systems comes when the autopilot is not set correctly and it startles the pilot by doing something unexpected when engaged.

The correct procedure here is to turn the autopilot back off and hand-fly while evaluating the situation. The natural reaction is to grab the wheel and try to make the airplane do what you want, even though the autopilot is on. The sensations and control pressures that come with that are pretty astounding and if any force is applied to the pitch control system the autopilot will roll trim against the force and continue to do so as long as the force is applied. That could be spatially disorienting, to say the least.

S-Tec 55X autopilot
The autopilot is really helpful, but only if you know exactly how to use it.

With my simple S-Tec, the device had absolutely nothing to do with pitch control. This procedure is not actually recommended but runway heading could be set, the autopilot activated in the heading hold mode, and the takeoff made in this configuration. If some aileron were needed on takeoff, as in a crosswind, the autopilot could be easily overpowered and it would revert to holding heading using roll when the need for aileron was finished.

Alternately, and the way I did it most of the time, the autopilot could be activated in the heading hold mode right after liftoff. Either way, the autopilot would be keeping the wings level while the airplane climbed on runway heading. The first turn after takeoff, a frequent time of trouble, could also be handled by the simple autopilot.

Many years ago, when autopilots were first becoming common, Piper actually offered models with two autopilots, a simple one and a full-service deal. I flew a Piper for a while with that feature and actually used the simple autopilot a lot while seldom using the more complex one which I didn’t care for.

To me, low visibility departures were an enjoyable challenge. I felt like I was prepared and ready to perform and the reward was great. If a pilot is at all apprehensive about doing this, best watch TV and drink a Coke in the pilot’s lounge until the visibility improves to one mile.

I want to offer the ultimate example of a disastrous IFR departure. I’ve used this example before, but it such a good example of how pilots can put themselves in impossible situations that it bears repeating. This is as it was related in the NTSB accident report. The airplane was a Cessna 551 Citation flown by an owner-pilot.

The pilot called from home at 0909 for an IFR clearance which became valid until 0930. He said he needed 15 minutes to get to the airport. The call ended at 0914.

The chief pilot preflighted the aircraft while the pilot was en route to the airport. The pilot arrived between 0920 and 0925.

After enplaning and loading he made an immediate start and taxied a short distance to the runway.

The chief pilot noted that approximately two minutes had elapsed from engine start until the takeoff roll was started. Up to three minutes was needed for the gyro to be up to speed after the generator had been selected.

Witnesses stated that the ceiling was between 20 and 100 feet and the visibility was reduced by fog.

The aircraft took off at 0930. At 0934 notification was received that the aircraft had crashed. It impacted 1.75 miles north of the airport in about a 30 degree nose down, 90 degree left bank attitude. No preimpact failures were found.

The chief pilot stated there were no mechanical deficiencies on the previous flight but he noted that some avionics had been slow to warm up. Also that the pilot had previously used the copilot’s HSI during takeoff.

There’s an old saying, that dates back to the year 1575, that seems applicable here: Haste makes waste.

All takeoffs have a lot in common and even though instrument and visual departures have different requirements they share one strong similarity: if you don’t have all those ducks in a neat row before advancing the power, those first three minutes could be long and potentially painful.


  • Good points—I’m certainly familiar with “that pause before takeoff” and making sure the mind is concentrating on the task at hand. “Lights, Camera, Action” as one of my instructors used to say (as in Lights—strobes etc, Camera—transponder, Action—fuel/mixture/props/controls free & correct). As you say, there’s no doubt takeoffs are at the top of the list of those special things about flying…ripe with failure modes too. Mr. Collins, have you bought that LSA Cub yet?!?

  • Very nice article. I especially appreciate your comment “staying visual for as long as you can see is not the way to do it.”

    This should be emphasized in training, but is not. The act of entering the clouds is disorienting, especially if suddenly looking down / turning head once references are lost. I would encourage everybody to practice, with a safety pilot, donning the hood shortly after rotation and practicing low IFR takeoffs.

    Then, taking off with a low ceiling is no different than what you have practiced. If you don’t “cheat” out of the hood/foggles, you won’t even realize that you are in IMC – the sensations are basically identical as in simulation.

    In actual low IFR I stop my outside scan starting 100′ below the clouds, and don’t even realize when entering the clouds, as by then concentrating on the instruments. I’ll look up again when stabilized, or when the sunlight tells me it is VMC and I’m above the layer.

  • Good article, Dick! I’ve enjoyed reading your stuff for years. I have about 50 hours in a C210, although not recent. Great airplane! Blue skies and tailwinds.

  • Reminds me of the time in the seventies when i was flying the V tail Bonanza. Touch & Goes on a 2500 ft runway were no big deal until right after a rolling takeoff the right side door (all there was) popped open when i was about 200 ft off the ground. i remembered my flight instructors words, ” leave it along until you are above 500 ft” .

  • Excellent points !!!!
    Taking the time to prepare and verify safe ORM/SRM takes some discipline and knowledge. This article is a gift to the sensible and responsible pilots, we get the knowledge without the risk of trial and error, all we need to do is appreciate the fact that Dick is a selfless teacher.
    Most of the accidents mentioned could have been prevented.
    US air carriers have minimal accidents in recent years due to strict ORM/CRM and rigorous crew training. Sterile cockpit,” No crew conversation bellow 10.000 feet”.
    My students remember things related to their focus in flight: Roger, a young student was a gifted manipulator of aircraft controls from the start and I always praised him for that. On the final approach to 17L we could hear the wild roar of engines from a Lear holding short for us. With very low engine noise the C-152 was quite and the Lear was very loud revving the engines like a hot rod. Roger lost focus and his stabilized approach became unstable…..wow what is that ? he asked with excitement……
    My reaction was very sharp as I called for go-around ….fly your airplane first Roger !!!!!!
    After securing the airplane, Roger apologized for the incident and promised not to let himself be distracted. Discipline !!!! We need it,our best friend in life and in flight.
    Many thanks dear Dick Collins !!!!

  • Nice article, although I’d be careful about putting into print the technique of using the a/p during takeoff, litigation crazy country that we live in and everyone wanting to point the finger elsewhere for their own misdeeds.
    I fly a Boeing for an air carrier, and like all airlines we go through a fairly comprehensive briefing prior to pushback and another prior to takeoff. This chance to “talk through” the departure and what actions the pilot flying (PF) will be
    taking throughout the takeoff/departure phase helps immensely. A “pre thinking” if you will, of what you’ll be doing. I do this same “pre think” in the trusty Skylane that I share with three others, going over in my head just what I’ll do and when. I do this as part of the preflight. After getting the clearance, doing the preflight, etc…just sit in the cockpit, visualize what it is I’ll be doing. It helps me immensely, and I suspect it would help others…..especially those of us that are over 50…60?..70?..whose memories and reactions aren’t what they once were. It takes a few extra minutes, it could buy you many more years of aviatin’.

  • Another factor that I think contributes to GA takeoff accidents is failure to calculate takeoff performance on a regular basis. Because I trained at fairly large airports with towers and long runways, we almost never calculated takeoff performance outside of prep for the checkride. It occured to me sometime later that I had no idea on each takeoff whether the aircraft was performing properly and rotating at the expected distance, and I had little confidence as time went on after my checkride that I could accurately interpret (without error) the somewhat arcane takeoff distance charts if I really needed to. I think CFIs should incorpoate an airline-style brief at the beginning of each flight and force the student to confidently calculate takeoff performance data, and demonstrate a plan if the aircraft isn’t performing as expected, instead of teaching the student to “eyeball it” or just figure “it’s a long runway, who cares?”

    • Adam,

      It’s neither practical or helpful to drag out the charts and calculator for every single takeoff unless the pilot is just newly transitioning and doesn’t have a good understanding of the aircraft’s capability. But for most pilots who do understand what their aircraft is capable of, it’s merely necessary to pre-define an envelope of safe takeoff operations. That envelope needs to account for runway length, slope, and surface condition; aircraft weight and balance; wind speed and direction; density altitude; and the general condition of the aircraft itself.

      For most light single engine prop aircraft, runway length is rarely the limiting factor in a successful takeoff. In other words, most of the time, if a given light SE prop plane can’t successfully take off in, say, 2,500 feet, it probably also couldn’t take off with 12,000 ft of runway, due to all the other factors.

      So a given pilot can define for himself/herself the shortest runway for a given set of conditions that is clearly OK, and then drag out the charts for any conditions that are outside the envelope. For my Cherokee, any paved runway of 2,000 feet or longer at zero slope at a density altitude of under 3,000 feet, loaded as I typically fly at least 200 pounds under gross, with either a crosswind or a headwind down the runway, with a paved surface and no snow or puddled water … then I’m going to be just fine.

      But if I’m flying out of a high elevation airport on a hot afternoon, anywhere near gross, with no headwind … I’m gonna drag out the charts and verify.

      As the last safety check I use Sparky Imeson’s guideline, i.e.., if I haven’t achieved 70% of rotation speed (or 40 mph) by halfway down the runway, then I’m shutting down the takeoff. And then evaluate why I couldn’t successfully take off (wait for cooler temps, offload passengers, fuel or cargo, verify that the engine is working OK, etc.).

      • I respectfully disagree. One of the reasons why commercial aviation is so safe is that they do the same thing, each and every flight, regardless of the preceived risk. Why in GA do we accept “eyeballing” things when we would never tolerate it within the commercial environment?

        How would we feel if the pilot of our next airline trip came over the intercom right before takeoff and said “good evening folks…seems like a prettty nice day out there, and I think I’m pretty light today with a nice long runway, so I decided to skip the normal weight, balance, and takeoff performance calculations because it’s pretty inconvienant. Have a great flight!” Of course we would find that unacceptable and unprofessional.

        In my personal flying, I’ve often found that I start to make mistakes when I think I can “skip steps” because something about the flight seems standard.

        In this day and age, it should take no more than a few seconds to use an automated tool to calculate weight, balance, takeoff distance, landing distance, etc. and provides the pilot an opportunity to really think about what they are doing.

        I caveat all this by saying that of course, a Cessna at an international airport is probably going to get off the ground, and if it’s severe clear for a local trip around the pattern you probably don’t need a full live weather breifing. But I think it’s a good idea to “practice how you play.”

        I think the accident statistics bear out that once you start going “ehh, this seems OK” it starts to really infect all areas of flying, leading eventually to high time pilots taking off overweight at short airports on hot days because they slowly crept over the course of their career to a place where they felt comfortable guessing whether a particular flight was going to be successful.

        Either way, great discussion and I’m glad this topic came to light. I think so much emphasis is placed on other areas of flight during training and flying that takeoff is often seen as an afterthought.

        • Commercial airline flying and personal private flying are entirely different enterprises. If we are to burden private pilots with all the procedures, protections, backups, and rules imposed on airline pilots, private aviation would completely disappear.

          That’s the nature of private flying. If you want to pretend to be an airline pilot, and go through all they go through … then you would not be able to fly at all unless you possess an ATP ticket, thousands of hours of left seat time, you’d be required to have a full time second officer with nearly equal qualifications, and so on and so on.

          Nope – the nature of private flying is necessarily vastly less restrictive. Because we’re not carrying hundreds of fare-paying passengers in the back, flying aircraft that weigh hundreds of thousands of pounds, and we are not getting paid six figure salaries for the privilege of flying.

  • Adam makes a good point and you do too Duane.
    As a CFI who instructed students mostly in a large airport I can see what Adam means. A large airport with long runways can spoil a pilot, but my approach is to emphasize the chart calculations that students find boring. I can show the calculations on the ground numerous times and the student will say later something similar to what Adam mentioned.( no offence ).
    On a 12.000 ft runway, I discourage intersection take offs when they leave most of the runway behind. If a student chose to do just that, he/she will experience a simulated engine failure…..need I say more ?
    Take off always using all available length and use optimum climb, altitude is our best friend.
    Did you by any chance think that I will not pull the power back if someone does take off with plenty of runway left ? I will pull the power back definitely when the timing is correct, after a thorough ground school and pre- take off briefing that is…no nasty surprises they are counter productive.The student should be able to maintain full control and land flawlessly….and then if I chose to apply full power, he/she is expected to instantly calculate whether can take off again safely and what their options are, enough runway to safely touch down or land straight ahead regardless of what is ahead the airspeed should be the ACs best gliding speed.The last portion is mostly mental calculation and discussion ( we don’ t get dirty).
    Needless to say the above is hard physical and mental work requiring great focus and discipline. The students eventually love the challenge and become confident.
    Any time a short or grass runway was available my students enjoyed the eye opening experience and appreciated the time spent on calculations.
    That is part of my job, to show them where to find the information and become thoroughly familiar with all available information, including ” runway lengths”. ( I remember when the FAA put that on the written tests)
    If your instructor failed to introduce charts and calculations (very unusual) and you pass the practical test, now is up to you to hit the books and charts for your safety and the safety of others.
    Fly Safe !!!!


    • I think you bring up a great point – students are often confused as to why CFIs are sticklers for seemingly illogical “rules” (like no intersection takeoffs or always taxiing to the last possible portion of the runway and then lining up on centerline).

      A student might think “why can’t I do an intersection takeoff with 8000′ of remaning runway, when we just took off from a 2500′ runway to get here?”

      And I think the answer ultimately lies in the fact that a good CFI wants their students to 1) fly the same way (with maximum safety in mind) each time, and 2) that it builds in the student’s mind the notion that they are responsible for all aspects of the flight as a future PIC.

      A student answering a question like “could you land on the remaining runway” with “I don’t know” or “I guess so” is unacceptable. Yet many CFIs don’t really force students to take that level of responsiblity for their flight.

    • Chris,

      I appreciate your remarks. But training is not real world flying, for reasons that are neither negative in nature, and some that are negative in nature. Statistically, the first couple hundred hours of flight time after primary training have relatively low accident rates, then the rate climbs from roughly 200 hours to 500 hours where it peaks. The relatively safe window right after getting the ticket means new pilots still freshly remember their lessons, and they’re also generally not as subject to complacency as are more experienced pilots.

      Yet, most of us don’t spend most of our flight time after getting our tickets doing pattern work, practicing turns about a point, doing the stall series, and so forth. We enter the real world where we do lots of things they never taught us in primary training, and we have lots of agendas to fulfill as pilots.

      In any event, learning how to verify the charts a few times at different airports under varying conditions is great training for how to do it when necessary. But it is NOT an argument to pretend that every flight has to be programmed to follow our training syllabus. That’s simply impractical.

      The method I outlined above that I have always used, and in fact learned during primary flight training, is to define safe or “green” envelopes in which is it NOT necessary to go through the charts … and as long as we understand the limitations of the envelope, no problem, period. In engineering, we call this “bracketing the problem”, which allows us to create perfectly safe and useful solutions without having to manually run all the calculations.

  • ” Train Like We Fight, Fight Like We Train.”
    “It’s said that Mother Nature is a hard teacher because she gives the test first and the lesson afterward. When the danger is clear and present, and especially when one is entrusted with the lives of others, discretion is the better part of valor.”
    Below is a sad story but true….
    It took place close to where I live, one of many preventable tragedies.

  • Here’s another “I learned flying from that” from your erstwhile, and very naughty Gringo Bush Pilot. The story:

    One stunning and beautiful Sunday morning I’m climbing out of Fallbrook Airpark (my home field) to fly up to the “Sin Fernando Valley” to sneak into my office in Encino while nobody is there. You know – a devious boss checking on the troops.
    At any rate, as the quaint little village of Fallbrook passes beneath me, I smell the unmistakable aroma of fresh eggs and bacon as it comes wafting into the cockpit. My saliva glands are bursting – when my addled brain comes to the realization that breakfast is not served in the cockpit – at least, not in a pauper’s Piper.
    Dimly aware that breakfast is cooking under the cowl, I do a 180 and descend to the field. With trepidation, I open the cowl – no breakfast … but on closer examination, and cleverly hidden by mama bird, is a nest with two eggs, just out of view over number one cylinder.
    As I clean my kitchen, I add one more item to my pre-flight mental checklist.

    Check the kitchen…

Leave a Reply

Your email address will not be published. Required fields are marked *