It's a question that usually follows the discussion of tantalizing performance and wow-'em technology: Are homebuilt-experimental amateur- built-airplanes safe? First off, it should be clear to even the most casual watcher of the homebuilt world that the airplanes are not failing out of the sky with atomic-clock regularity. Considering the tremendous leeway afforded to pilots and builders of experimental-class airplanes-and the propensity for these models to stretch the aeronautical envelope-by most measures, the current standards work.
But we as pilots tend to want hard numbers. The AOPA Air Safety Foundation maintains an accident database, culling from National Transportation Safety Board information. Each year's accidents are tallied and presented in the Nall report (comprehensively titled: "Joseph T. Nall General Aviation Safety Report, a safety project of the AOPA Air Safety Foundation"). We looked at the numbers for 1993, which include mostly preliminary reports from NTSB. As the board files its final reports in the next few months, the grand tally might change.
According to the Nall figures, the overall accident rate for general aviation was 8.79 per 100,000 flight hours, with a rate of 1.67 fatal accidents per 100,000 hours. Pat yourself on the back: General aviation has steadily improved in safety. Figures for 1983 show a total rate of 10.73 and a fatal rate of 1.94 per 100,000 flight hours. It's important to have an accident rate qualified by realistic hours-flown numbers, and that's where homebuilts fall through the cracks. There's no approved, sanitized hours-flown figures for this class, so knowing the number of accidents and fatalities provides only half the picture. What's more, we won't even try to wildguess the usage; homebuilts, tend to be recreational devices rather than workhorses, so the traditional means of deriving a number-surveys and educated wetted digits thrust into the slipstream-don't necessarily apply. Suffice it to say that there are about 22,000 airplanes on the Federal Aviation Administration registry under the category of experimental-Amateur-built. (That number, by the way, has increased by about 840 airplanes a year since 1982, but it's still anyone's guess how many of those are active.)
If it is statistically dangerous to attempt constructing a rate, then it might be worthwhile to look at the basic accident causes and how they compare to those of production airplanes. In rough terms, the major causes of accidents resemble those for factory-built steeds. According to the Nall report, in homebuilts, the pilot was the major factor in 52.6 percent of the accidents, with mechanical failures/poor maintenance accounting for 13.3 percent. Accidents whose causes were unknown tallied 32.9 percent, and that equally nebulous category of "other" consumed 1.2 percent. On the production-airplane side, the pilot received credit for 67.1 percent of the accidents, and mechanical/ maintenance causes represented 7.8 percent. Unknown scored 23.2 percent and other, 1.9 percent. Among the possible conclusions: The maintenance issue is almost twice as prevalent in the homebuilt accidents, and those crashes labeled unknown also play a larger part.
Statistics shift when you focus on serious accidents, or those in which the pilot or passengers received serious or fatal injuries. Here, the homebuilts and factory- constructed airplanes are much closer. Pilots were the main cause of 66.2 percent of homebuilt serious accidents and 67.1 percent in production models. Mechanical and maintenance-caused accidents were 8.5 percent for homebuilts and 7.8 percent for factory-builts. Unknown and other causes chalked up 22.5 and 2.8 percent in experimentals and 23.2 and 1.9 percent in production aircraft, respectively. One might surmise, then, that the makeup of serious accidents follows suit with production airplanes, but homebuilts tend to have different kinds of incidents where the pilot and passengers are not gravely injured.
What's more telling, perhaps, is the breakdown with regard to phase of flight. Compared to factory iron, homebuilts scored more serious accidents during maneuvering (40 percent versus 30), landing (13 percent against 7), and takeoff and climb (19 percent next to 14). Significantly, weather played a much smaller role in homebuilt serious accidents, 11 percent versus 26 percent for the production birds.
These numbers are fairly easy to explain. Maneuvering accidents likely come from the differences in handling and performance from production airplanes. Typically, homebuilts have more installed power for the weight and often maintain higher wing loading than common production models. For the transitioning pilot, these characteristics, combined with the need for higher approach speeds and the presence of generally lighter controls, prove simply too much. Also, and this is an important point, not all homebuilts have traditional stability characteristics, which are generally built into factory airplanes for safety reasons and to ease the transition from one model in the maker's line to another. It's no mere coincidence that the Cessna 172 feels just like a heavier 152 and that the 182 responds like a bigger Skyhawk. Also, homebuilts are more likely to be engaged in casual aerobatics, fleetwide, than production airplanes, further increasing exposure.
Experimental airplanes do not need to meet Federal Aviation Regulations Part 23 certification rules, including those for stability, control response or harmony, or minimum stall speed. (Although it's true that were many of them subjected to the tests, they would indeed pass.) These differences should put the experimental-airplane pilot on notice that extraordinary skills may be required. Experience is the key, and statistics show that low time in type plays a critical role in the frequency of accidents. Stoddard-Hamilton saw a dramatic decrease in Glasair III accidents after about half of the fleet's owners attended type-specific training. Such rigid guidelines will likely become the norm for high-performance kitbuilts.
Another aspect of experimentals influences the number of wrecks: flight test. Any way you cut it, flight test is a high-risk endeavor in even the simplest and most docile of airplanes. Each homebuilt is different, and there is no assurance that the individual project will come out with the same handling characteristics or performance as the kit-maker's version or a similar model down the hangar row. So the test pilot has to be ready for anything; he must be capable of flying the airplane and troubleshooting under adverse conditions. The test pilot also must be current in the type (or at least something similar) and remain as unemotional as possible about the prospect of conducting the flight test-that means wearing a parachute and being prepared to use it if something goes horribly wrong. All of these requirements work against the average homebuilder.
Should you as a homebuilder hire a test pilot? For most, that answer ought to be yes. Historically, homebuilders perform poorly as test pilots. The numbers are hard to pin down, but industry estimates range from 15 to 40 percent of the total homebuilt accidents occur on the initial flights or, just as important, during the builder's first few flights. Often, the years spent constructing have not also been filled with useful flight time; some stop aviating completely to indulge in the creative process. And many builders lack time in airplanes similar to those they build. Add inexperience, low (or no) time in type, and the typical builder's myopia — I built it, it's got to be right and you've got accidents almost literally waiting to happen.
Unfortunately, the preliminary NTSB reports don't include aircraft time or the pilot's time in type, so one must read carefully through the 174 narratives to find the uncracked nuts. We counted 10 accidents in which the summary showed the airplane to be on its initial test flights (either after completion or a rebuild). These accidents also include cases in which the builder crashed with very low time in type.
Accidents like this one are all too common, from the NTSB preliminaries: "A homebuilt Seahawk struck the ground in an uncontrolled descent shortly after taking off ... for [the pilot's] initial test flight of the experimental airplane. Witnesses reported that the airplane got airborne, bounced on the runway, and then assumed a nose-high attitude with pitch oscillations. It then rolled left to an inverted attitude, hit the runway, slid off to the side, and burned. The pilot had roughly 300 hours of flight time. The airplane was destroyed and the private pilot/builder was seriously injured."
Mechanical turmoil also influences the accident makeup with homebuilts. Out of the 174 accidents we reviewed, 57 of the summaries contained the phrase "lost engine power" or something like it. Of these, the direct breakdown includes 53 related to the engine or propeller, two to the fuel system, one to the oil system, and one to the electrical or ignition system. It's true that the builder of an experimental may overhaul and maintain everything on the airplane, including the engine. It's also true that many homebuilders opt for used engines (most often true when buying certified powerplants). These numbers may well be pointing toward a lack of familiarity with the engine and its subsystems; in many homebuilts, thanks to a range of powerplant options, installation details are often left to the builder's discretion. Further, in looking at the accidents individually, it seems the powerplant failures are fairly evenly spread among certified models and nontraditional engines, two-stroke as well as four-stroke types.
Some accidents point to questionable design or maintenance, like this one. From the narrative: "A homebuilt Dragonfly lost control while landing ... the pilot reported that the airplane touched down hard while landing and bounced. During the recovery he applied aft stick but ... the pistol grip broke off from the stick and he lost control of the airplane." In this case, perhaps, a second set of eyes during. the most recent inspection might have caught whatever caused the grip to depart the control system.
In reading through 1993's roster of accidents, there are a few items to note. Homebuilt safety should be a state of mind. Poorly maintained experimentals can be just as deadly as poorly maintained production airplanes; gravity doesn't know the difference. And most important, training is critical. As Monday-morning quarterbacks, we can guess that more time in type or basic flight currency may have kept at least a handful of these accidents out of the Nall report altogether.
Related Articles
