Probable Cause #3: Trapped On Top

  • E-Mail this Article
  • View Printable Article
  • Text size:

    • A
    • A
    • A

There are some among us who like to tempt fate by skirting the rules. But a Bonanza pilot who did that got more than he bargained for, as this report of Probable Cause explains.

This article originally appeared in IFR Refresher, August 2005.

Probable Cause

Flying in the clouds requires currency and proficiency at flying on the gauges. It's the pilot's responsibility to meet the legal requirements before taking off into or flying through instrument conditions.

A problem we see too often, though, is when pilots who are not current or proficient decide to fly above a cloud deck when the weather is good VFR below the clouds. To them, that doesn't seem like a matter for real concern because they figure they will find a hole to drop down through when they get close to their destination, or they will simply get an IFR clearance down through the cloud deck to VFR conditions below. In their minds, they will not do any "hard" IFR flying, only a short descent through the clouds to VFR conditions below.

What they fail to consider, though, is that the forecast could be wrong and that the weather below them could close in with little or no warning. They disregard completely the possibility that they might have to descend through clouds and fly an instrument approach to get back on the ground safely.

One pilot I knew would sometimes fly above the cloud decks, even though he wasn't instrument rated. Several times he had to get assistance from ATC to get down through the deck, but he never caused enough of a problem to draw the attention of Oklahoma City. He was lucky. But what would he have done if his vacuum pump quit while he was on top?

I asked him that question several times and he always had an answer. He said he believed he could handle whatever came up, and he had no intention of getting an instrument rating. Luckily for him, he never got into a situation where he had control problems. But others were not so fortunate.

Trouble At Altitude

On a spring day in 2000, a Beech N35 V-tail Bonanza departed Caro, Mich., for Newport News, Va. Cruising at 9,000 feet on an IFR flight plan, all seemed normal until the flight was in the vicinity of Conicville, Va., when the pilot reported that he had an electrical-system failure. Shortly after that, the pilot amended the failure, telling the ATC controller that he had now lost his suction.

"I guess my gyro is going round and round," the pilot told the controller, according to ATC transcripts that are included in the final NTSB report.

The pilot then asked the controller if he knew of any airports in the area that had broken clouds, adding that he was currently on top of the clouds and would stay that way. The controller informed the pilot that the nearest airport that showed broken conditions was Charlottesville, Va., while nearby Shenandoah and Petersburg were both overcast.

"Yeah, give me a heading to somebody that's broken. That's what I'm gonna have to do," the pilot said.

"You do have your gyro back. You just want to go to the closest airport with a broken ceiling. Is that the way I understand it?" the controller asked.

"I do not have my artificial horizon, my gyro. I got the compass and the airspeed indicator." Then he said, "I have to stay on top because I haven't got my artificial horizon."

The controller then issued a series of vectors to take the pilot over the Charlottesville Airport. A couple of minutes later the pilot said, "Alright. Just stay on this heading until something shows up. Long as I can stay on top I can keep the darn thing right side up 'cause I lost my artificial horizon." Then the pilot asked the controller what altitude he was indicating on the radar screen.

"I'm showing you at 9,800 feet," the controller replied. "I don't have any traffic in that area. If you can't hold an altitude, that's fine. Just do the best you can."

When the controller asked the pilot to clarify the problem he was having, the pilot responded, "Looks like it's gonna be the gyro or suction pump cause I don't have the artificial horizon. I do not have the directional gyro. The GPS is working. The airspeed is working. The altimeter is working."

A few minutes later the controller asked the pilot if he could descend to 7,000 feet or if he wanted to stay at 9,000 "I'm almost in the clouds right now," the pilot said.

The controller then asked if he could climb to 11,000 feet, to which the pilot responded that he could. At that point, the controller declared an emergency for the pilot and advised him to maintain any altitude that would keep him clear of the clouds.

But things weren't going well for the pilot. A few minutes later, he again asked the controller what altitude he was showing.

"Your Mode C is coming and going. Last time I saw was 9,700. I'm not picking up a Mode C right now," the controller replied.

"I'm showing myself on an easterly heading. All my instruments have gone haywire," the pilot said.

The controller replied, "It just popped up at 10,000. Are you on top now?"

"Negative," was the reply. "And my vertical speed is going."

"I'm showing you at 10,400," the controller said.

The pilot replied, "Washington Center, my airspeed indicator is going from ... Oh, man, I'm gonna ..."

That was the last transmission from the aircraft.


There were two witnesses to the accident, one of whom had to run to get clear of the falling aircraft. The main wreckage impacted 15 feet from where he had been standing. He told investigators that he heard an "engine popping noise that became progressively slower." He looked up into the sky and saw the airplane falling in pieces. He then heard two loud bangs and ran because the aircraft was heading directly towards him.

The other witness was about three-quarters of a mile from the impact site. He said he noticed an airplane doing "air tricks" and making sputtering noises. He focused on the airplane and saw it break into pieces. He also recalled hearing three loud booms coming from the direction of the aircraft.

The fuselage came to rest, inverted, on top of a power transformer, killing the pilot and his single passenger. Both wings were broken off the plane and located 400 and 550 feet from the fuselage. More ominously, sections of the left elevator were found approximately 2,500 and 4,100 feet southwest of the impact site. The left and right horizontal stabilator and the right elevator were never located.

The Bonanza pilot held a private pilot certificate with single-engine land and instrument ratings. He had accumulated 1,554 hours of flight time, and 89 hours of actual instrument time. However, his last recorded entry for IFR flight was 1.3 hours flown in September 1997 and his last instrument competency check was in April 1995 (respectively 30 months and five years before the accident).

Upon examination of the engine-driven vacuum pump, it was discovered that the pump's drive shaft had sheared and that two of the pump vanes on the rotor were broken. As a result, the rotor could only be turned small amounts. Torsional stress marks on the drive shaft suggest that this failure occurred prior to impact and resulted in the pilot losing his gyroscopic instruments. This is further backed up when no rotational scoring was found on either the inside housing or the rotors of the attitude indicator or the directional gyro. In other words, they weren't spinning when the Bonanza impacted the ground.

The aircraft had 4,589 airframe hours in service and the engine had 1,524 hours in service.

Tempting Fate

This accident reminds me of an incident that took place many years ago when a client of mine went to Wichita, Kan., to pick up his new Cessna 182 at the Cessna factory. He and his three passengers left Wichita early in the morning for the flight to Worcester, Mass. They stopped for fuel later in the day but the pilot did not check the weather because he had been given a rosy forecast for the entire trip earlier that morning.

But during the day a stationary front that had been lingering for several days started to move. Some very heavy thunderstorms developed and moved through Massachusetts and Connecticut late in the afternoon. The frontal passage was followed by some very thick fog that lasted the entire night and into the next day.

My client was flying over a broken layer of clouds that eventually became overcast. But he didn't turn around, nor did he check the weather further. He kept going until early in the evening he arrived over Windsor Locks, Conn., and decided he had to do something about getting down. He checked the weather and found that the ceilings were 500 feet or less below him with visibilities ranging from zero miles at Worcester to three-quarters of a mile at Windsor Locks.

Fortunately, my friend had not lost a vacuum pump and all systems continued to work properly as he called Approach Control and confessed his predicament. Fortunately, the controller he was talking to was also a pilot. He gave my client, who was not instrument-rated, vectors to line him up with the ILS, told him to turn on his wing leveler and set the airplane up for a 500-fpm descent.

The controller was good. My client did exactly as he was told and as the airplane broke out approximately 500 feet above the surface, he spotted the runway and was able to make a normal landing.

Could the Bonanza pilot have done something similar? The weather report some 35 miles away at Staunton, Va., at the time of the crash showed a ceiling of broken clouds at 8,000 feet and another broken layer at 10,000 feet. Perhaps if he had trimmed the airplane for a 500 fpm descent and minimized the control inputs, the airplane could have descended through the clouds and into VFR conditions with plenty of room to spare. The airplane did not know the vacuum pump was not working and that would not have effected how the airplane flew.

But the pilot knew the pump was not working. Perhaps through instinct and compulsion, he tried too hard to fly the airplane partial panel and his lack of proficiency worked against him. His weather briefing probably indicated that he might have a layer of clouds to climb and descend through, but he didn't plan on his vacuum pump failing while he was on top of the clouds. Likewise, the inoperable gyros probably rendered his autopilot -- if there was one -- unusable, adding another level of difficulty to the situation. Still, allowing the airplane to descend through the clouds on its own with minor roll and pitch inputs would have been a better solution than over-controlling the plane and causing it to break up in flight.

Most of the accidents we read about that involve the loss of a vacuum pump occur in much worse weather than the Bonanza pilot faced. Some kind of instrument approach may be required, which means the pilot has to be capable of flying the airplane on a partial instrument panel. Unfortunately, there are too many instrument pilots who barely maintain their instrument proficiency, never mind partial-panel proficiency.

And it's liable to get worse as new technology continues to infiltrate our cockpits. Eventually, our main instruments will be CRTs that show us what the ground looks like below us, exactly where it is, and what's down there for us to hit. But what if the fancy electronics quit for some reason? We are going to have attitude and directional indicators for backups.

In an ironic twist of fate, the same instruments we now use as our primary source of instrument flying will be the partial panel of the future. And it will be just as imperative then that pilots practice flying on those instruments as it is now for them to practice flying on the "needle, ball, and airspeed," a skill this Bonanza pilot apparently had lost.

The NTSB cited the failed vacuum pump, but more importantly, the pilot's lack of recent instrument experiences as factors that caused this accident.

More accident analyses are available in AVweb's Probable Cause Index. And for monthly articles about IFR flying including accident reports like this one, subscribe to AVweb's sister publication, IFR Refresher.