Probable Cause #1: Breakup In Flight
AVweb introduces a twice-monthly column investigating accidents and presenting lessons learned from them. This week in Probable Cause, we look at whether a faulty switch brought down a twin-turboprop. Quick recognition and a thorough scan could possibly have saved the day.
When we depart on an instrument flight we expect that the journey will progress normally to our intended destination. That, of course, is not a realistic expectation, as anybody at the NTSB can tell you. Because the chance of something going wrong is always there, we must constantly be aware of how our instruments and operating systems are working.
In this particular accident, the 62-year-old pilot-in-command had been flying for many years and had nearly 5,000 hours total time. According to the NTSB report, he had flown 86 hours in the 90 days preceding the accident and had 774 hours of total instrument time. He had also received a biennial flight review and an instrument proficiency check two months prior to the accident. He held a private pilot certificate with multiengine, single-engine and instrument ratings. He had approximately 400 hours in the accident airplane and nearly 2,500 hours of multiengine time, much of it in turbine aircraft.
It appears that this pilot was well-qualified to be flying the Cessna 425 Conquest turboprop that he had owned for about a year and a half. Yet, on an early March day in 2002, the pilot lost control of his airplane, causing it to break up in flight, killing him and his two passengers. After a thorough investigation, the NTSB determined that the accident was precipitated by the failure of a small but critical system. Had the pilot recognized the failure immediately for what it was, the odds are that you would be reading about a different accident right now.
The aircraft departed San Jose's Reid-Hillview (KRHV), Calif., airport at 10:29 a.m. on the day of the accident, bound for La Paz (MMLP), Mexico. An IFR flight plan was filed and the pilot had received an instrument clearance prior to his departure.
The aircraft was initially cleared to climb to 13,000 feet while being radar vectored to Victor 485, a southeast/northwest airway that runs between the San Jose (SJC) VOR and the Priest (ROM) VOR. The pilot was told to fly a heading of 110 degrees to intercept the airway and proceed on course.
Recorded radar information reveals that the airplane climbed at about 2,000 feet per minute through 6,700 feet. Then it began a series of unauthorized heading and altitude changes. It appears as though the airplane turned right and climbed to 8,600 feet before turning left and descending to 8,000 feet. The next turn was to the right while the airplane climbed to 8,500 feet followed by a rapidly descending right turn. At 10:34, the aircraft descended through 7,000 feet and the pilot told the ATC controller, "4JV, I just lost my needle. Give me ..." That was the last transmission from the aircraft. The last radar return showed the Conquest descending through 3,200 feet at a rate of 11,000 fpm. An analysis of the radar data concluded that the plane was nearly at Vmo before radar contact was lost.
A witness on the ground about 1.5 miles from the impact site heard a loud "screaming" jet sound. She looked up and saw the aircraft descend out of the clouds in a corkscrew pattern. The airplane's flight path changed from a descent to a climb while still turning. The witness told investigators that as the plane ascended, the corkscrew turns became much tighter. The airplane then leveled off momentarily before it began an arcing and spiraling turn, descending behind a hill.
Other witnesses on a golf course located about half a mile from the impact point heard a loud sound that they thought was an explosion or gunshot. They saw the airplane roll level in an arcing horizontal spin until it disappeared behind a hill. All the witnesses told investigators that they saw parts falling from the airplane and that it was "smoking."
The wreckage was distributed over an area approximately a third of a mile long in low, rolling, coastal hills about 10 miles southwest of KRHV. The fuselage came to rest at an elevation of 493 feet MSL.
The left wing and engine remained attached to the fuselage, with the propeller hub still attached to the engine. The inboard right wing with its engine was separated but remained attached to the fuselage by the control cables, fuel lines, and electrical wires. Both main landing gear were found in the extended position while the nose gear appeared to have been in an in-transit position.
The left elevator came to rest in a tree about 425 feet west of the horizontal stabilizer and about 1,275 feet northwest of the fuselage. A section of the right outboard wing was found on the north side of a 650-foot hill, approximately 1,300 feet from the fuselage while the left wing tip was found on the southern slope of the hill.
Structural examination of all of the recovered components revealed that the failures were the result of excessive loads on the airframe.
NTSB investigators interviewed the pilot's son, who said he had spoken with his father three or four days before the accident flight. He said his father did not mention any chronic or unresolved problems with the airplane.
A FAA airworthiness inspector visited the FBO that had refueled the airplane the day before the flight began and obtained a sample of jet fuel from the truck that had refueled the airplane.
The FBO employee who fueled the aircraft told the inspector that his truck had run out of fuel after 107 gallons were pumped into the aircraft's left tank. The FBO did not have any more fuel, nor did any other FBO on the airport have any JET A available. The aircraft owner ordered that 50 gallons be taken from the left tank and placed in the right. The capacity of each tank is 186 gallons, so splitting the 107 gallons pumped into the left wing probably evened out the fuel load between the two wing tanks, although the employee could not determine how much fuel each tank held prior to the transfer. A sample of fuel from the aircraft also was obtained. Both were submitted to a laboratory and no contaminants were reported in the samples.
While the weather wasn't terrible, it was far from ideal. There were several AIRMETs issued for mountain obscuration, turbulence and icing. AIRMET Zulu, update 2, was issued for icing conditions over portions of Washington, Oregon, California, Idaho, Montana, Wyoming, Nevada, and Utah. It warned of moderate rime to mixed icing in clouds and in precipitation below 18,000 feet. The closest upper-air sounding came from near Oakland, about 39 miles northwest of the accident site. The 1200Z sounding on the day of the accident indicated that a saturated environment existed, with the temperature and dew-point spread less than two degrees Celsius and a relative humidity greater than 85 percent from the surface to 10,800 feet. The freezing level was located at 8,790 feet.
An airborne traffic reporter who located the downed aircraft at the request of ATC said the weather in the vicinity at the time of the accident was 1,600 feet (MSL) scattered with a higher overcast layer at 2,000 feet. The visibility below the clouds was good, in the range of 10 miles, he reported. No rain was falling and there was some turbulence consisting of "light chop."
The Conquest was certified to operate in known-icing conditions and was equipped with flight instruments on both pilot and co-pilot sides. This included two pitot tubes, each feeding one of the two airspeed indicators in the cockpit. Suspecting that icing may have played a role in the accident, NTSB investigators focused on the positions of all instrument panel switches. Except for the left pitot heat, the anti-ice and de-ice system switches were positioned normally for flight in known icing. The right pitot heat switch was found in the ON position while the left pitot heat switch was in the OFF position. Additionally, the left pitot heat switch was "noticeably displaced" to the left, an indication that it may have been struck by something during the crash.
In a laboratory, stereoscopic examination of the switches found that the right pitot heat switch was intact. The toggle lever mechanism of the left pitot heat switch, however, was broken loose from its housing. The report noted that a microscopic examination of the left switch's housing revealed imbedded debris and wear marks that were indicative of an old fracture. No other failure-related external anomalies were noted. The switch was noted to be intermittently open with the switch in the ON position.
The switch was then disassembled and examined further. It had a significant amount of "large, coarse, fibrous, lint-like debris" in it. The flexible copper conductor of the left switch circuit breaker section had several broken strands and the electrical contacts were dirty.
The conclusion was that the toggle portion of the switch was bent to the left during the impact sequence. But the housing fracture predated the accident and allowed an internal buildup of the lint-like debris. This resulted in the switch becoming unreliable, causing intermittently open electrical contacts. The elements of the pitot tubes and heated static sources were connected to a battery and were found to operate normally.
Check It Out
The NTSB concluded that the cause of the in-flight breakup and accident was the pilot's loss of control that resulted in the design limits of the aircraft being exceeded. The loss of control was due in part to the loss of primary airspeed reference resulting from pitot-tube icing that was caused by the faulty switch. Other factors were the pilot's distraction caused by the airspeed reading anomaly and spatial disorientation.
If this is indeed what happened, there were several actions the pilot could have taken that would have averted the accident. First, had he checked the pitot heat during preflight, it is possible that he would have found the pilot's side to be inoperative. The report on the pitot-heat switch said it was possible to get the switch to make contact if the switch were manually held in the on position. Had he just turned the switch on during his preflight and checked the heating element it most likely would not have worked.
In flight, had the pilot recognized the loss of the airspeed indicator for what it was, chances are the accident could have been avoided. By looking at the airspeed indicator on the right side instrument panel he would have seen a normal reading and concluded that the airspeed indicator on his side was inaccurate.
Remember, basic instrument flying consists of scanning all of the instruments so that we develop a mental picture of what the airplane is doing. If all of the instruments except one appear to be operating normally, it is still possible to develop the mental picture required for continued flight.
A pilot must never allow himself or herself to be distracted by the failure of one instrument. You must concentrate on the rest to determine where the fault really lies. Chances are the 425 pilot never realized what the real problem was that day.
He may have thought that the airplane's attitude was the issue. If the autopilot was turned on when he noticed the problem, he probably disconnected it, thinking the problem was with it.
The comment he made to ATC that "I just lost my needle. . ." could indicate that he thought he had an instrument failure that included his turn-and-bank indicator. Again, he could have looked at the instruments on the right side instrument panel and seen that all of the gauges except for the airspeed indicator offered the same readings as those on his panel.
This is not the first time that a pilot in instrument conditions misinterpreted an airspeed indicator failure (this one caused by the loss of pitot heat), and it probably won't be the last. You can prevent this type of accident from happening to you by improving your instrument scan. Do not allow yourself to get into the habit of just looking at the attitude indicator, direction indicator, altimeter, and VSI with occasional glances at the airspeed indicator and turn-and-bank instrument.
You need to scan all of the instruments constantly. The turn-and-bank indicator or turn coordinator is one of the most important instruments in the panel. Yet, many pilots ignore it altogether. The more information you derive from your instrument scan, the easier it is to develop the mental image of what is going on. Once control of the aircraft is lost because of a misunderstanding of what is occurring, it is even harder to regain it.
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.