Back ...

Actually, the results of TAM 3 were encouraging. I wanted to keep trying, and, with some reservations about how much time they could give, the team decided to try again in 2003.

I started building slightly modified models almost immediately. I paid the drag penalty of a slightly fatter fuselage to put all of the fuel in a single tank. I shifted the wing position forward to better accommodate a CG (center of gravity) shift as fuel was consumed. I also moved the autopilot to midway between the wing trailing edge and the stabilizer leading edge; this ensured that there was no way rainwater could be sucked in to foul the electronics. We were flight-testing this version two months later and found the drag penalty to be negligible.

I rang in 2003 with three new airframes—numbers 23, 24, and 25—ready for engine testing. In early February I took on a small consulting job with some of my former colleagues at the Johns Hopkins University APL. They had a high-school senior serving as an intern, and they told me that Cyrus Abdollahi would learn more, and be more useful to their UAV [Unmanned Aerial Vehicle] project, if he came on board as an intern on our TAM effort.

That was a classic understatement. Cyrus was a knowledgeable model builder, a good RC pilot, and a whiz on computers. We started working together in February, and he spent nearly 25 hours a week helping the blind man! We built three more airplanes and did some good flight-testing for STAR and for APL. It is a delight to work with this young man, and I miss him now that he is in college and able to come by only on Sundays.

During the winter and spring of 2002-2003, I continued to test engines down in my shed. You would think that after 12 years of work and five records I would know everything there is to know about an O.S. .61 FS engine. Not so.

After several months of mysteries, I found a flaw in two of my five homemade rear power takeoffs that are used to drive an alternator that powers the whole system. I have acquired 26 O.S. .61 FSs at flea markets and on the Internet. The flaw wandered from one engine to another as I moved it to test engines. It takes as much as 12 hours of running to see if a change has a significant effect, and if you’re working to get six engines in proper condition, it’s easy to lose track of some the variables.

The culprits were units I had machined back when I had good eyes. Tweed Cottrell made six new ones of superb quality.

By June there were six airframes; some had been through preliminary flight-testing. Autopilot construction had fallen behind, so I was continually transferring the one workable unit from one airframe to another. Les Hamilton and Ron Bozzonetti gave up many hours of their hobby time to come out for flight tests.

On July 18, a virgin model (number 26) with a virgin autopilot failed while being manually flown. In the rush, my shake-and-bake routine was apparently inadequate to avoid an infant mortality; the model went down in some woods on the edge of a wheat field approximately a minute after launch. Dedicated STAR member Andy Gutow flew over the woods in a Cessna with Ron Bozzonetti as a second pair of eyes. Nothing was seen; 26, with alternator, gyro, good engine, and autopilot, was gone. Period!

On July 29, Gay and I set out toward Saint John’s in a Chevy Astro minivan loaded with five TAM models: numbers 23, 24, 25, 27, and 28. First, Gay first drove out to Lexington, Kentucky, to trade our Subaru for the van that our son-in-law Larry Snipes owned—a trade that saved $3,000 of STAR’s funds. The cost underrun was a huge help to the budget.

We arrived in Saint John’s on Sunday, August 3, and were cheerily greeted by friend and benefactor Nelson Sherren. The rest of the crew—Joe Foster, Les Hamilton, and Cyrus Abdollahi—flew in the next day.

I honed the engine on number 24 for two days, and we went out for trim tests Friday afternoon, August 8. The first launch—now referenced as TAM 4—went up into a fogless sky over Cape Spear at 8 that evening. There was a gentle wind from the west and the dirt road sloped down to the west, so this 77-year-old geezer had no trouble heaving it into the air.

People told me that the climbout and smooth straight-line departure into the sky still-painted by the setting sun were beautiful things to see. I have to take their word for it because it is gone from my eyes if it is more than 200 feet away.

Satellite data flowed into our operations room via E-mail messages for the next eight hours. TAM 4 was on course, the engine was fine, the speed was right, and the altitude was correct at 1,000 feet. Then after 430 miles, a bit short of the place where we lost TAM 3 the year before, there was nothing. No further report.

The cause could not be determined because the model went down during a period when all satellites were absent from the north Atlantic. There was no rain, winds were modest, and storms were hundreds of miles to the south. It was a mystery.

Joe Foster postulated carburetor ice because TAM 4 disappeared at roughly the same time that TAM 3 did. My hundreds of hours of bench tests during cold and hot, humid weather didn’t support that idea. Somebody suggested that the Bermuda Triangle has a cousin up in Greenland. Cyrus, in his quiet, humorous manner, said that maybe the Icelandic Navy’s gunboats needed target practice. We had no clue.

The weather reports Saturday morning, August 9, were favorable. Tailwinds of 10-15 mph in the middle of the flight would help, but there would be crosswinds for a couple hundred miles after launch and the last couple hundred miles near Ireland. Conditions were expected to deteriorate two days hence, so we decided to try for another launch that night. The hustle to uncrate number 25, undo the safety packing, and stuff it with autopilot and engine put a bit of a strain on a crew that hadn’t slept much the night before.

It was 11:30 a.m. before I could start running fuel tests. Ordinarily I have to run and measure for three or four hours to make sure the needle and filters are okay. For the next two hours fuel flows swung from crazy highs to saggy lows. I use a fussy medical filter in the fuel line that likes to stay wet. Apparently this one had dried out in spite of isolation clamps during the seven-day trip. I knew the routine; I had to put in a new filter and start over. But by that time it was 2:30 p.m., and we had to go off on a trim flight.

From 4 p.m. to 5 p.m. I measured fuel again, but I was far from sure that it was okay. Soon we would have to start the fueling and weighing process. Carl Layden, the Canadian observer, was waiting in the shop with Contest Director (CD) Les Hamilton to attest to the weigh-in’s accuracy. Expletives flashed through my brain before I finally said, “Throw this flukey thing and hope!”

We did just that at 7:45 p.m. local time. The launch was easy because of a mild west wind. Joe’s climbout was quick and smooth. TAM 5 did a graceful turn toward the northern waypoint then beelined out of sight on the 62° heading to Ireland.

All was well at 11 p.m. Data indicated engine rpm okay, altitude okay, speed approximately 43 mph, no tailwind. There was nothing I could do but hope, so I packed off to bed. Les, Joe, and Cyrus stayed on post to sift the incoming data.

There was good news when I entered the TAM shop at 8:30 Sunday morning. The Spirit was still flying and was roughly 560 miles out. However, there were some ominous aspects of the data.

The engine was supposed to be regulated and smooth at 3,900 rpm, but readings ranging from 4,100 to 3,100 were coming in. Altitude was also a bit unsteady, varying from 280 to 320 meters (plus or minus 70 feet). Data points came in roughly once a minute for periods of up to 12 minutes.

We couldn’t plot any meaningful graphs, but it was obvious that the model was porpoising from a shallow climb to a speedy dip in the flight paths. We had seen this behavior in Maryland flight tests; I was fairly certain that the engine was lean, making insufficient power to pull the model into its usual smooth cruise condition. If the porpoising stayed mild, the penalty in speed was not severe.

The Spirit trotted along all day Sunday. Over the midocean it picked up a 5-10 mph tailwind and was cruising at 50-55 mph. This tailwind was supposed to be 10-15 mph, but Mother Nature doesn’t always do what man predicts. I went to bed at roughly 10 p.m., fearful that the cool of night would increase the viscosity of the fuel, taking the engine from lean to dead.

Sure enough, when I returned to the shop at 4 a.m., Les and Cyrus said they had had no satellite data for three hours. Les and I agreed that we should call the Irish officials—Joe Dible and John Molloy—and tell them to turn around. They were on a six-hour trip from Dublin. Likewise, we called AMA President Dave Brown, who had volunteered to land the model. He was at a bed-and-breakfast near the Mannin Beach site.

“We have a satellite!”

Within minutes of my saying goodbye to Dave, Cyrus turned away from the data computer he’d been monitoring and announced, “We have a satellite!” Cyrus had been a heartwarming example of unusual patience and youthful optimism from the start. The old geezers had given up, but not Cyrus.

We quickly called the Ireland crew. None of them had turned around. (Later we learned that a couple of satellites had not dumped their data and had to go around the earth one more time before reporting to the system.)

Not only was the Spirit still flying, but it was flying better. It was 5 a.m. Monday morning in Newfoundland, but the Spirit was far enough east to be in warming sunshine. It had flown over the Gulf Stream during the night. Perhaps warmer air voided the viscosity problem. It was obviously happy to be rid of a lot of fuel weight. The elevator had gone from nearly full up-trim to a little below neutral. The engine rpm was a comforting, steady 3,900.

I buried my head on Gay’s shoulder and wept unashamedly for joy.

At 9 a.m. Newfoundland time (12:30 Ireland time), or 371⁄4 hours into the flight, the model was approximately 70 miles from the Irish coast. Its speed was down to 43 mph. Its heading was right on target at 95° true heading. We had a real fingernail-chewing cliffhanger on our hands. The engine was supposed to go roughly 37 hours if I had set the needle at the intended reliable, slightly rich setting. I do not chew my nails, but I, and everyone else waiting in the Newfoundland shop, was very tense.

My thoughts drifted to a 1999 cross-country attempt to fly from Leesburg, Virginia, to Savannah, Georgia—a distance of roughly 540 miles. We had flown for 71⁄2 hours. We had passed Ron Clem’s distance record of 507 miles when a Niagara Falls-type thunderstorm swamped the model, drenched the crew, and put two to three inches of water in the footwells of my 1972 Ford LTD convertible.

At that point I realized that the engine might stop a couple miles offshore. I mused at my well-worn slogan “You can’t win ’em all!” and hoped I wouldn’t have to use it this time. My fears didn’t materialize into disaster; the model came into sight at Mannin Beach at roughly 2 p.m. Ireland time on Monday afternoon.

Dave Brown, a member of six US World Championships teams in the 1980s, confidently toggled the landing-gear switch to gain manual control of the airplane. He banged the rudder stick hard right to kill the engine. Dave’s wife Sally was cell-phoning the happiness. Dave glided the Spirit into a dead-stick landing approximately five feet from the designated spot. At 2:08 p.m. Ireland time, Sally’s report—“It’s on the ground!”—raised a whooping cheer in the Newfoundland shop. I buried my head on Gay’s shoulder and wept unashamedly for joy.

Continued ...