Prior to the flight the pilot-rated passenger and first pilot discussed who would operate the aircraft. The passenger would occupy the left seat and taxi the airplane to the active runway, while the first pilot would occupy the right seat and do the flying. The passenger did not remember if the parking brake was set or released prior to beginning his taxi, but did remember that the aircraft did not pull to the left or right while taxiing. Upon reaching the active runway and getting the airplane lined up, the first pilot took control of the airplane and advanced the throttle. The passenger said he didn't feel like the airplane was accelerating properly, and when he asked the first pilot if the airplane was accelerating properly, the first pilot replied, "Yes." As the airplane continued down the runway on its takeoff roll, and sensing the airplane was still not accelerating, the passenger asked the first pilot if he wanted him to put some flaps on. The first pilot replied, "No," and there were no further communications between the passenger and first pilot. The airplane continued off the end of the runway and through a 66 foot grass overrun before going over an embankment and impacting trees. The aircraft subsequently came to rest inverted at the bottom of a ravine. The fuselage and tail section of the aircraft were oriented up and parallel to the slope of the ravine, while the nose and left wing of the aircraft were partially submerged in the adjacent canal which borders the edge of the ravine. Witnesses to the accident thought the pilot was doing a high speed taxi, that the airplane reached a maximum speed of approximately 50 miles per hour, and that there was no attempt to abort the takeoff roll. Several witnesses reported that the engine didn't sound right, was making noises, and was not running smoothly. A post accident examination of the airframe, engine, and propeller did not reveal any anomalies which would have precluded normal operation of the aircraft.Handling an engine failure in a single-engine aircraft during the takeoff roll might seem like a no-brainer, but it requires immediate and sensible action by the pilot if there is to be a favorable outcome. I'm always surprised by the number of certificated pilots with whom I fly in single-engine aircraft who do not brief the handling of emergencies before takeoff and climb-out. Not only is it important to self-brief if you are the only pilot on board, it's just as crucial to discuss the possibility when there's another pilot or an instructor on board.
The National Transportation Safety Board determines the probable cause(s) of this accident as follows:
The failure of the pilot to abort the takeoff roll. Factors contributing to the accident included the failure of the aircraft to accelerate for undetermined reasons and the trees.
In fairness to GA pilots, the Practical Test Standards do not require a rejected takeoff demonstration for single-engine private pilot and commercial candidates. Student pilots learn to do countless takeoffs and landings, so it's no wonder that we're spring-loaded to commit to a takeoff. However in the accident cited above, both occupants were retired airline pilots, each with several thousands of hours of flight experience. Each pilot knew how to brief a takeoff and how to abort a takeoff in an airliner. The accident illustrates that in a GA aircraft, the pilot in command needs decision making skills based on something other than Kentucky windage.
Rejecting the takeoff in a single engine aircraft while you're still on the runway is easy - close the throttle, maintain directional control, brake as needed, and exit the runway or evacuate the aircraft. What you need is some objective and subjective criteria to aid your decision as well a plan of action should you decide to abort.
Computing takeoff and landing performance is actually a 14 CFR 91.103 requirement, so if you are not at sea level on a standard day, or if the plane is close to gross weight, it's time to get out the handbook and run the numbers. Departing a big city airport with 6000 to 10,000 foot runways, it's easy to become complacent. Takeoff calculations in these cases helps you identify an abort landmark - a point down the runway where you should be airborne. Read the fine print in the performance charts carefully, since many GA aircraft manuals are famous for the large print giveth and the small print taketh away. Remember, too, that this performance data tends to be biased toward a new aircraft and a highly-skilled and current pilot. When in doubt, add a fudge factor of at least 50% to the suggested takeoff distance.
Identify an abort landmark where you predict the aircraft should become airborne. If you reach that landmark and you're not off the ground, reject the takeoff and keep the plane on the ground, where you'll have plenty of time to troubleshoot the issue.
In addition to objective criteria for accepting a takeoff - RPM in the green, manifold pressure above 25 inches, oil pressure and temperature normal, airspeed indicator alive - it helps to have subjective criteria for aborting. The engine running roughly or back firing, the plane pulling to one side, excessive nose wheel shimmy, or the feeling that something just isn't right are all good reasons to keep the plane on the ground.
What about the landing data for the airport from which you are departing? Computing the landing data for your departure airport might come in very handy if you have to return for an immediate landing or if you have to perform an off-airport, forced landing. Landing data is also important because light, single-engine aircraft can get into some pretty short runways. With high density altitude, it's not uncommon to be able to land on a small strip in the mountains, only to find you do not have enough runway to takeoff.
Airborne in a single-engine aircraft and still at a low altitude or in the clouds, your options are limited if something goes wrong shortly after takeoff. You are probably going perform an off-airport landing, the question is where? If you fly out of the same airport(s) on a regular basis, why not learn the location of the flattest, least populated, most obstruction-free areas near the airport? This is as much for the protection of people on the ground (who have not accepted the risks of flight that you have) as it is for your own protection. The survivability of a forced landing in smaller aircraft is quite good if you land on a flat, firm surface with few obstructions. The smaller the aircraft, the greater the odds you'll walk away provided you fly the plane all the way to a landing. The longer you maintain control, the better your odds.
If you are departing IFR, have an approach selected that you'll use to return if something goes awry. Departing Oakland on runway 27 left or right, I often see the few pilots who even have bothered to think about an unplanned return choose the ILS RWY 27R as their emergency approach. If you are departing to the west in a twin-engine aircraft, it's quite possible you could limp around 8 miles or so back to the east to begin the ILS. But with a single-engine, you need something better. You're departing to the West, so why not select an approach that arrives from the West - such as the VOR RWY 9R or GPS RWY 9L? Or if all else fails, know a VOR radial or heading that will take you back toward the airport or your emergency landing site.
A great exercise is to fly to a quiet airport that has a reasonably long runway and practice aborting takeoffs. One of my favorite places to do this sort of thing is Castle Airport in Merced. A former B52 base, now a public use airport, it has an incredibly long runway.
It's tough to make the decision to not even start the engine if the numbers just don't add up. It's even tougher to reject a takeoff when you're rolling down the runway, you haven't estimated the takeoff performance, and you haven't come up with a plan of action as the trees at the end of the runway loom larger and larger.