The first flight of our new RV-3 had gone very well. The morning was overcast, but the ceilings were much higher than the floor of the Class B airspace above us. The winds were predicted to be above the limits that we had set for the day, but at the time, they were fine at both our departure field and the recovery field a few miles away. We planned to make the first landing elsewhere to take advantage of a much wider runway. Since we also had friends with hangar space there, it made sense that if we had any problems with the airplane or weather, that would be a fine place to be. As it turned out, that first flight was by the book, and the landing worked out well. During the time it took to remove the cowl and check the powerplant, however, the winds had come up to the point where we all felt that it would be best to wait until later in the afternoon for a second flight, assuming that the forecast was correct and the winds would drop.
That second flight was primarily to get time on the engine and do minimal systems and handling evaluations. The chase plane was being flown by my wife and carried her as PIC and a flight testing work colleague in the right seat with checklists acting as the mission director. My job was to fly and let Steve manage the test plan, watch for traffic, and generally keep track of the things secondary to the flight. We launched about an hour and a half before evening twilight into light winds and nice, smooth skies. Our flight plan called for long racetrack patterns over the Texas coastal plains—smooth, flat terrain almost entirely covered with pastureland, fields, and oil-field roads. Because we live at an airpark underneath that Class B, circling the field is a somewhat frenetic activity—flying the tests where we knew we could land—even “off field” was a better proposition.
Those oil field roads are really nice, by the way. Oil rigs are big, heavy pieces of equipment—therefore, the roads they must use are heavy duty, smooth with nice gravel, wide, long and straight. They have no wires anywhere nearby that might snag a big rig, and generally are closed with locked gates at the ends to keep folks out. In other words, they are pre-made gravel runways. As I said—it was no accident that we centered our test area on that sort of terrain—the next best thing to having a dry lakebed.
The recorded flight path tells the story—in this case, (with careful planning) the roads were far better than the fields.
After the second takeoff, we followed a short route to the test area that allowed us constant access to possible landing sites—roads, field, and private runways. Even though it took less than ten minutes to get to our preferred area, we wanted no gaps. The airplane was flying beautifully and I was enjoying the fruits of two and a half years of building effort that had been shared with my wife. The engine was smooth, temperatures were good, and I anticipated that it would break in quickly—if it wasn’t broken in already. We’d been aloft a little less than thirty minutes when Steve radioed, “So Paul, have you switched fuel tanks yet?” Now this was a question, not a command. We didn’t really have the tank switch programmed into the plan—but then again, there should be no harm in keeping things balanced, and it was an obvious thing to do…so I keyed the com and said, “Not yet, but I’ll do that right now,” reaching down at the same time and turning the selector knob in front of the stick from the left tank to the right. For fifteen seconds, everything was fine—and then it got real quiet. The throaty roar of the four-pipe exhaust went away, replaced only by the smooth sound of the windmilling three-blade, constant-speed prop.
“OK, so the engine just quit!” I said, in as calm a voice as I could muster as I keyed the mic. “Roger that, engine just quit,” came the reply from Steve, “I’ll listen, you talk.” Years of training as a flight-test team was paying off for the two of us. You see, I had spent 30 years in NASA’s Mission Control, flying the space shuttle as a flight controller at first, and then as the flight director for close to twenty years. Steve was an experienced astronaut and had served as a capcom (spacecraft communicator) on and off for the 15 years he had been in the program. We had trained together for many hundreds of hours, handling simulated emergencies over and over again (with the help of our flight control teams and astronauts in the simulator cockpit), usually many at one time. We had handled many real problems during actual missions. And now, here we were with a simple, one-problem day—nothing to get too worked up about.
“Steve, I am going to try for a relight, would you check my landing options?” I said as I began scanning the cockpit for a reason for the silence. Mags, fuel pump, fuel selector back to the other tank…”Paul, how much power have you lost?” came through my helmet. I pumped the throttle forward and back several times and nothing changed; the windmilling up front stayed constant. “Looks like all of it, I’m trimming for best glide,” I responded. “See if we can make the Creasys’ strip,” I added, referring to a private airstrip owned by a married couple who worked with us on the space program. The engine had failed at about 6000 feet msl (which was also agl on the coast), and we were now down to about 4000 feet. Even as I asked, I knew that the Creasy strip was out of reach with our descent rate, so I picked out the nicest stretch of oil field road below me and set up for a power-off pattern.
Gliding was nothing particularly new to me, having helped develop many of the landing procedures and techniques for NASA’s “heavy glider” (as we sometimes called the shuttle) over my decades with the agency. Even the relatively steep descent angle of a short-winged RV seemed more akin to a sailplane than the brick-like descent of the shuttle orbiter. It was ironic that it was only a couple of years earlier that I had actually added a glider rating on to my license (a gift from my wife). All in all, the landing was fairly simple, even though it was only my second landing ever in an RV-3. I played the turn to final to hit my touchdown point approximately one third of the way down what I would guess was about 8000 feet of straight gravel road with rice fields on each side. The touchdown was smooth, and I got right on the brakes to bring us to a halt as quickly as I could, in case the road ahead wasn’t as good as what I was on. Only after I stopped did I look farther down and realize that now I was going to have to walk more than half a mile to the end which opened on to the highway to meet my ground team.
Louise and Steve flew overhead to mark my position on the GPS, and I called on the radio that I was down and safe, collecting coordinates off my own GPS, and was getting ready to call the ground crew to come to our aid. Chase stayed overhead until I had made that contact, then made the five-minute flight home to get onboard with the ground crew. Meanwhile, my job as pilot was finished—it was time to become a builder/mechanic. And it didn’t take long to find the problem.
We had built the airplane with the standard fuel valve supplied by Van’s in the kit. It is a simple, brass-body valve that some refer to as the “tractor valve” because it would be right at home on the farm. Not as pretty as an aerospace valve, it has done the job in thousands of airplanes for many years, requiring only a little cleaning and lubrication now and then. The only thing I don’t really like about it is the handle. Fortunately, there are a couple of aftermarket options for better-looking handles, and the one that I chose fit either of the two configurations of the valve—one with the “flat” of the shaft in one direction, and the other with it 180 degrees opposite. The handle accommodates this by use of a set screw that can be positioned either from the tip of the arrow, or from the tail. Once the builder has decided which way it needs to be, they are supposed to use Loctite to fasten it permanently in place—this keys the handle to the shaft. The handle is actually held on with a screw that is tapped down the center of the shaft.
What I forgot was the simple act of using Loctite on the screw—one of the thousand details that come at the very end of a project. The set screw backed out just enough so that when I turned the fuel valve, it started to move the shaft—just enough to get off of the tank I was on, but not enough to get onto the new tank before it slipped. This effectively trapped me between tanks. This was a simple builder error that cut off the fuel and lead to a forced landing. But there is even more to it than that. In Van’s design, the valve is plumbed so that the left and right tank positions are 90 degrees from each other, and the valve is cut so that anywhere in that 90 degrees of throw, fuel will be conducted from one tank or the other—or partially from both if it is in the middle. Without realizing that this was an important design feature, I had changed the plumbing so that the left and right positions were 180 degrees apart—and there is an “off” in between. I never thought to ask why the design had what, to me, seemed an odd configuration. I made a change without first fully understanding the established design.
The culprit—the author forgot to use Loctite on the set screw that “clocked” the valve handle to the valve.
So, there were two lessons to be learned by this builder. The first was that about 60 or 70 years ago, airplane systems designers figured out that a good fuel valve shouldn’t have an “off” position between tanks—it should be off to one side or the other. The second was that there are no unimportant details when dealing with a fuel system—and that set screw needed to be secured.
I guess there is a third lesson that can be shared, one that I learned in the aerospace business when I was a young engineer, and that is to always have a backup plan. There has never been a perfect airplane built; there is always the potential for error. For those unanticipated errors, we have to have plans—and in my case, it was to have a good landing option should the unexpected occur. Because just as there are no perfect airplanes, there are certainly no perfect builders.