John Smith regularly took flying holidays with his family. His youngest daughter, Jacqueline, followed in his footsteps, getting her PPL once she turned 18. The father and daughter were in a syndicate of five people who owned a Piper Cherokee Arrow (PA-28R-201T) and the logs show that they took the plane most Easter weekends, booking it weeks in advance.

In 2007, they spent Easter weekend in Oban, flying every day along with wife and mother Angela in the back seat. A member of staff at Oban remembered them and posted to an online forum: “I spoke to the family shortly before they departed for home, and they came across as very nice people who were enjoying their flying in this part of the world.” But when it came time to return home, the weather was bad. They decided to depart Oban despite the clouds. The next morning, the wreckage was discovered in a farmer’s field in the west Highlands, just nine nautical miles from Oban Airport.

John Smith gained his UK PPL in March 1980 which was valid for life. He had 324 hours experience. His medical certificate expired on November 2006 and he had not submitted the annual exercise ECG required as a result of his heart condition. However, there is no evidence that he had any medical issues during the flight.

Jacqueline Smith’s was one of the first of the new-style JAR PPLs which only have a five year validity. In 2005, the CAA did not send any notifications that a renewal was required and numerous pilots were later found to be flying on expired licences. Jacqueline Smith did renew her Single Engine Piston (Land) rating and may well have overlooked the fact that her licence had expired.

Friday, 6 April 2007

The flight departs Andrewsfield Airfield in Essex with three on board. Father and daughter regularly flew together, with John Smith in the left seat and Jacqueline Smith in the right seat, regardless of who was the pilot in command. As such, it’s not possible to tell who flew the aircraft that day. Jacqueline Smith’s logbook shows that John Smith had flown the seven previous flight hours as commander and in the accident report, John Smith is referred to as the “commander” for simplicity’s sake. As he had more experience and was the one discussing the flight with airfield staff, it seems likely that he was, in fact, the pilot in command.

13:44 The family flight arrives at Blackpool Airport to refuel. John Smith is seemingly unaware that Blackpool required prior permission for all incoming flights. This was a temporary condition due to deal with heavy traffic over the bank holiday weekend so the Smiths would not have known this simply by looking at the plates. It’s not an uncommon mistake.

14:51 They departed Blackpool, arriving at Oban at 16:30. Oban Airport at the time required prior permission as standard and the Smiths were required to give at least three hours notice. John Smith expressed surprise when the Air/Ground Operator raised the matter with him after their landing.

The family toured around the local area every day in the aircraft. On Sunday the 8th, after returning from their flight, they refuelled to full ready for their flight home the following day. They went to a local hotel for dinner and drinks and arrived at the airfield the following morning.

Monday 9 April 2007

10:00 The family arrives at Oban airfield and loads the aircraft with their luggage. They report to the office and pick up the weather reports and forecasts.

The aftercast shows that between 10:00 and 12:00 hours, the cloud base was low with rain and drizzle around the mountains. Visibility at Oban varied between 4,000 metres and 10km and hill fog would have been extensive, with visibility less than 200 metres. Cloud was scattered or broken with a base varying from 400 feet to 1,500 feet and a top of 2,000 feet. But further broken or overcast clouds were above that laying starting at 2,000 to 3,000 feet with a top varying from 5,000 to 7,000 feet and further broken layers of cloud above that.

Later, a number of pilots around Scotland said that they had cancelled their flights on Monday owing to the rain and cloud. The Professional Pilots Rumour Network includes posts from pilots in the area in the days following the accident.

Light aircraft down near Oban – PPRuNe Forums:

“I know the area around Oban well and that afternoon I wouldn’t have been tempted to get airborne.”

“Ben Vorlich and the like were shrouded in mist, and the winds were very strong – you could certainly believe the weather guesser’s prediction of 28 knot+ gusts. I’ve never flown into Oban before (just overflys from my plank days), so all that, plus the surrounding terrain, was a big factor in my decision to go have lunch elsewhere and leave that for another day.”

However, John Smith hopes that he can get his family home. He comments that the weather is clearer in England and mentions as a part of conversation that he is not instrument rated. He decides that they will go up to have a look at the weather but, if it isn’t suitable, they will return to the airfield. The Air/Ground operator tells him that they won’t be charged for an additional landing if they have to come back. The family board the aircraft and do an engine check before heading for the runway.

10:35 The aircraft departs west and at approximately 1,000 feet above mean sea level it disappears into the cloud.

John Smith transmits that they are at 1,500 feet amsl and changing to the en-route frequency.

Oban Air/Ground give them the frequency for the appropriate Scottish ATC station and warns them that they’re unlikely to receive a reply until they are further south.

John Smith acknowledges this without reading back the frequency.

10:42 The aircraft is high enough to appear on radar tracking at Tiree in a climbing left-hand turn passing through 3,300 ft. It then does a complete circle to approximately 5,800 feet over the Isle of Kerrera, 8 nautical miles south-west of Oban. The aircraft turns were becoming tighter and its altitude seems to vary between 5,600 and 5,900 feet. The flight then appears to “wander” in a roughly south-easterly track.

From the Air Accident Investigation Branch official report Piper PA-28R-201T Turbo Cherokee Arrow III, G-JMTT:

The weather at takeoff and the forecast for the first part of the flight over south-western Scotland was not suitable for the intended Visual Flight Rules (VFR) flight. It is possible that the aircraft climbed in a hole in the cloud over the Isle of Kerrera. Once the aircraft had climbed to height the pilots would not have been able to keep in sight of the surface, as the privileges of their licences required them to do.

Neither pilot was trained in instrument flying other than the basic familiarisation training as a part of their initial training, which is meant to allow a pilot who accidentally flies into instrument conditions enough information to be able to fly back out. It is unlikely that either of the Smiths was experienced enough to maintain control of the aircraft in cloud.

10:48 The aircraft altitude reduces to approximately 5,300 feet and about thirty seconds later, the aircraft enters a left turn. During this turn, Tiree radar records that the altitude drops to 4,700 feet and then to 3,700 feet in under 16 seconds.

Lowther loses secondary radar tracking for 23 seconds. There’s no mechanical reason for this and it is likely that there was no line of sight between the radar and the aircraft transponder, which is consistent with the aircraft in a severe nose-down attitude.

In summary, the radar data shows the aircraft climbing to, and holding, a relatively stable cruise altitude but with no set direction. Turns were initiated, culminating in a relatively tight turn associated with a large descent rate and unusual aircraft attitudes. Electrical power was available at least until nearly the end of the last recorded turn, well after the tight descending turn was initiated. Given the location of the end of the radar track relative to the accident site location and disparity between the direction of the last recorded track and the estimated impact direction at the accident site, the aircraft carried out at least one further half turn between loss of the radar track and impact. It is also possible that it carried out further complete turns or other manoeuvres below radar coverage. The time between the loss of radar track information and impact is not known.

11:55 Oban’s Air/Ground Operator phones Scottish ATC to enquire whether G-JMTT has made contact with them. It has not.

There is no reason to file a flight plan for this flight, thus the Air/Ground Operator cannot know the route they are taking back to Andrewsfield. He thought they maybe planning to stop at Blackpool to refuel again, but the Smith’s have not contacted Blackpool for prior permission, which they were made aware was required on Friday’s flight. If ATC had been formally notified of the flight, overdue action would have been initiated within an hour of G-JMTT’s non-arrival. In practice, it would not have made a difference.

Tuesday 10 April 2007

One of the other pilots of the plane becomes concerned as he is unable to contact John Smith by phone. He contacts Oban and confirms that the flight departed the day before. He contacts John Smith’s office, who tell him that John Smith was expected in for 10am meeting but he has not arrived. He then contacts Andrewsfield who state that they have not heard from G-JMTT. He phones Distress and Diversion to inform them that G-JMTT is overdue. A full overdue action is instigated at 14:08.

13:40 Angus McFadyen, a farmer near Loch Scammadale, is out with his son, checking the sheep in the hills above Bragleenmore Farm. They discover the wreckage on a hillside at 963 feet above sea level and contact the Strathclyde police to report the crash.

Both wings were sheared off but McFadyen later tells the BBC that he recognised the fuselage as being from a plane. He finds the engine and the remnants of the cockpit 32 metres (104 feet) away.

The impact crater and other features are consistent with a high-speed nose-down impact at a speed of between 140 and 200 knots.

14:50 Distress and Diversion are contacted regarding the wreckage. The search is called off and the accident investigation begins.

So What Happened?

Post-mortems were carried out on all the occupants of the aircraft. All three died on impact. There were no obvious signs of disease and all the occupants tested negative for drugs. Although it is possible that a medical situation incapacitated one or both of the pilots in the flight, there was no evidence that this was the case.

However, both John and Jacquelyn Smith tested positive for alcohol. John’s result was the equivalent of a blood alcohol concentration of 99mg/100ml and Jacquelyn’s was 48mg/100ml. Angela Smith, in the backseat, tested negative, making it unlikely that the alcohol detected could have been produced post-mortem.

It is an offence to perform an aviation function where the proportion of alcohol in your blood is over 20mg/100ml. As a comparison, the legal limit for driving in the UK is 80mg/100ml. John Smith was legally too drunk to drive and yet appears to have been in command of the aircraft. Jacquelyn Smith had a lower result but still was not legal to fly.

Based on the weather reports, the flight took place in cloud (Instrument Meteorological Conditions or IMC), which meant that the pilots were unable to determine the attitude of the aircraft visually by looking at the horizon. They had to rely on instruments as they were not in sight of the ground.

The flight instruments in the plane consisted of electric Horizontal Situation Indicators and Turn Coordinators and a vacuum-driven Attitude Indicator (Artificial Horizon).

The vacuum pressure was supplied by a vacuum pump with no back-up. On the right side of the instrument panel was a gauge to show the level of suction and on the left side of the instrument panel was a warning light which illuminates if the suction drops too low.

The Century III autopilot on the plane relied on the vacuum-driven Attitude Indicator as its attitude reference source. The pressure from the altimeter is used to command the autopilot to maintain the set altitude. The autopilot is reliant on the pitch information as given by the Attitude Indicator, which is also the primary instrument for safe flight in IMC. Without the Attitude Indicator, the autopilot would not function correctly and would not be able to hold a heading or an attitude.

The GPS unit was damaged in the crash and so it is not possible to determine the aircraft’s last recorded position nor see what communication and navigation frequencies were set.

The flight was tracked externally by two radar installations. The radar tracking is not completely accurate and in fact, the tracks from Lowther don’t quite coincide with the tracks from Tiree. However, there’s certainly enough data to understand the “motion trends” and get a general feel for what the aircraft was doing.

The plane appears to have been “under reasonably precise control” until the final tight turn when the aircraft starts to descend rapidly. Lowther then loses secondary radar, indicative of the aircraft being in an unusual attitude. From the data, it appears that the Smiths had lost control of the aircraft and were plummeting nose first towards the ground.

The vacuum pump was severely damaged, the rotor and vane assembly had shattered. However, close examination showed rub marks on the fracture faces … which meant that that the engine-driven end of the pump must have still been rotating after the coupling fractured. That meant the engine was still running.

That means that the fracture must have occurred before the aircraft crashed into the ground.

The vacuum pump was eleven years old and had done approximately 994 hours. The pump manufacturer had sent out a Service Letter which limited its use to 6 years and 500 hours but the aircraft owners were not aware of this. The service manual was unclear and it could be understood that the pump needed to be replaced at the 1,000 hour inspection.

The vacuum pump had only one function on this aircraft: to supply vacuum pressure to operate the Attitude Indicator.

If there was no vacuum pressure, the gyro rotor within the Attitude Indicator would slow down and then start to topple. This takes quite some time and minutes could pass before the gyro was moving slow enough to become unbalanced.

The Attitude Indicator will give false information based on the movement of the gyro: it can’t tell the difference between the plane movements and the gyro toppling. If the autopilot were engaged, it would follow the false indications of pitching and rolling shown by the Attitude Indicator.

If the plane has lost vacuum pressure, the Attitude Indicator will not show it. The suction gauge should show a zero reading and the vacuum pressure failure light should illuminate.

As a result of the impact damage, it was impossible to determine whether the light was illuminated before the plane crashed. Nor is it possible to determine whether the autopilot was engaged at the time of the crash. But the faulty readings of the Attitude Indicator could confuse a pilot as easily as they do the autopilot. Flying in cloud, a pilot relies on the Attitude Indicator because his own responses cannot be trusted.

With the absence of outside visual references, physical sensations can produce compelling perceptions of the aircraft’s attitude and manoeuvres that differ markedly from those indicated by the flight instruments and spatial disorientation can occur. This tends to be more likely when recent and/or total instrument flying experience is low and in a high stress situation, or with alcohol in the pilot’s blood.

The flying conditions through cloud would have been stressful. Neither pilot had been trained in instrument flying. Both pilots had blood alcohol levels over the legal limit. Score three out of three for spatial disorientation.

Flying through clouds, the pilot would have suffered from “the leans”. A major breakthrough in aviation safety in the early 20th century was understanding this phenomenon and teaching pilots to use instruments for visual input, rather than trusting the false sensation of leaning. Instrument training teaches a pilot to use all the instruments in order to get a clear understanding of the aircraft’s attitude, including what to do under circumstances like this where a primary instrument has failed.

Without visual input, a pilot can inadvertently put the aircraft into a turn in order to try to straighten out a non-existent turn. This inadvertent turn then develops into a spiral dive and the plane accelerates nose-first into the ground, just as the radar tracking shows.

Although the Air Accidents Investigation Branch did not have records of accidents caused by vacuum system failures, their US counterpart, the National Transport Safety Board, had 62 accidents/incidents over the past twenty years which cited the vacuum system as one of the factors. Of those incidents, 40 were fatal. Reports showed that in many, the pilot was aware of the loss of vacuum pressure before losing control in IMC conditions – that is to say, flying through cloud.

It’s still possible that something else happened, that the pilot was incapacitated in such a way that did not leave medical evidence … and it is impossible to prove without a doubt that the vacuum pump failed completely; however the final spiral dive combined with high airspeed are standard symptoms of spatial disorientation and there is clear evidence that the pump must have broken before the impact.

The Air Accidents Investigation Branch concluded that the failed vacuum pump was the only failure discovered in the wreckage that could have caused the resulting loss of control of the aircraft. It is unclear when the vacuum pump failed but they estimate that the resulting erroneous readings probably started just before the aircraft entered the final left turn, approximately 24 seconds before the radar track was lost.

Conclusion

The aircraft crashed after control was lost while in IMC. The characteristics of the final flight path, particularly the high airspeed, the rapid descent and the rate of turn, were consistent with the effects of spatial disorientation. The pilots were not IMC or Instrument Rated, and alcohol was present in both pilots. It is likely that the accident resulted from loss of control as a result of the pilots following unreliable indications from the AI, whilst in IMC. The AAIB has made four Safety Recommendations relating to the maintenance of vacuum pumps.

The pilots were not IMC or Instrument rated. Had they been flying under VFR conditions, in sight of the surface, they would probably have been able to maintain control of the aircraft.

Four recommendations were made, focused on the correct maintenance of vacuum pumps. As for VFR pilots who fly in bad weather because they just really want to get home, well, there’s not much that the AAIB can recommend other than Don’t!

---

If you enjoy this blog, consider buying my ebook: You Fly Like a Woman for less than a Starbuck’s coffee at Amazon, Barnes and Nobel or Smashwords.