Flybe DH8D at Exeter on Sep 11th 2010, distracted crew causes EGPWS warning

Last Update: June 14, 2012 / 13:51:50 GMT/Zulu time

Bookmark this article

Photo of Flybe G-JECF, De Havilland Dash 8 (400) — Flybe G-JECF, De Havilland Dash 8 (400) (Photo credit: Global_5000 / Flickr / License: CC by-nd)

Incident Facts

Date of incident
Sep 11, 2010

Airline
Flybe

Flight number
BE-1794

Departure
Bergerac, France

Destination
Exeter, United Kingdom

Aircraft Registration
G-JECF

Aircraft Type
De Havilland Dash 8 (400)

ICAO Type Designator
DH8D

A Flybe de Havilland Dash 8-400, registration G-JECF performing flight BE-1794 from Bergerac (France) to Exeter,EN (UK) with 49 passengers and 4 crew, was on approach to Exter's runway 26 descending through 3300 feet MSL towards 2600 feet MSL with autopilot engaged descending on vertical speed and with the approach mode armed when the crew received an "IOP1 FAIL" (Input Output Processor #1 Fail) message at the engine display, the speed bugs and minimum descent altitude was removed from the captain's displays while the first officer's displays remained normal. The captain attempted to restore his displays by switching the source to air data computer #2, however, without success, then returned to normal setting. When the captain selected the data source to #2, the approach mode disarmed, after reselecting to normal the captain realised the approach mode was no longer armed and armed the mode again. The captain subsequently decided due to lack of bug speeds on his displays to hand control to the first officer. Shortly after the first officer took control the enhanced ground proximity system (EGPWS) issued a "CAUTION TERRAIN". Both pilots had been working to resolve the IOP1 message and now looked up. They were in visual meteorologic conditions and could clearly see the runway ahead, however, the EGPWS now issued a "TERRAIN! TERRAIN! PULL UP!" warning prompting the first officer to disconnect the autopilot and advanced the power levers to about 80% being confident this would be sufficient to maintain terrain clearance. At that point ATC inquired whether they were descending with the glideslope and the captain realised that the pre-selected altitude had been cleared at the control panel permitting the aircraft to descend below the selected altitude. The captain informed ATC they had an instrument failure and were climbing to capture the glideslope again. The aircraft climbed to about 2200 feet, captured the glideslope and continued for a safe landing.

The United Kingdom's AAIB released their bulletin concluding:

This serious incident was the culmination of a sequence of events. The initiating factor was an avionics failure which led to a loss of cockpit indications during a critical phase of flight.

Existing operational procedures did not provide clear guidance for flight crews to deal with this failure. This situation was exacerbated in this case by a departure from standard operating procedures, resulting in the loss of previously selected flight director modes. A breakdown in the monitoring of the approach profile led to a descent below the glide path and the triggering of a GPWS warning.

This incident, once again, highlights the importance of monitoring the flight profile, especially when dealing with unfamiliar situations, and the need to react appropriately to GPWS warnings, particularly when the cause is not immediately apparent.

The AAIB reported that the aircraft had descended to 1417 feet MSL/ approximately 700 feet AGL about 8nm before touch down before it started to climb again in reaction to the EGPWS warning.

The aircraft logs showed a first IOP1 FAIL message had occurred on Aug 22nd 2010, maintenance performed the troubleshooting procedures as per manual without findings however. The next IOP1 FAIL message occurred during the incident flight, maintenance reset the relevant circuit breaker and IOP1 returned to normal operation. The aircraft was released with the request for crew reports of IOP FAIL messages. Two more IOP1 FAIL messages occurred on Sep 20th 2010, maintenance swapped IOP1 and IOP2 as a result. On Sep 23rd 2010 an IOP2 FAIL message occurred, however was not logged as the indication disappeared after flight crew powered down and powered up again on the ground. On Oct 1st 2010 another IOP2 FAIL occurred, an operational test performed satisfactorily. On Oct 8th 2010 another IOP2 FAIL message occurred, subsequent maintenance checks finally confirmed IOP2 showing a fault, the IOP was replaced with no further FAIL messages occurring.

The AAIB quoted the operator stating that IOP FAIL messages were common amongst their fleet causing operational delays. However, IOP are not amongst the most frequently removed components. Only about 20% of the IOP FAIL messages result in confirmed IOP fault detections and replacements of the unit. In the vast majority of cases the fault clears after a reset of the unit. A high number of units removed from the aircraft and sent to the manufacturer for repair were returned with the notification "no fault found" but continued to cause trouble as soon as they were installed on the aircraft. As a result, the operator had adopted a policy which would label a unit, that had been sent for repairs and returned with "no fault found" for three times, as "rogue" and to be prevented from entering the spares inventory again. At the time of the incident three rogue units had been identified.

The manufacturer of the IOPs was aware of the issues and had established an investigation to detect faults that can not be reproduced by the acceptance test procedures in the workshop. This investigation found a number of components, including the secondary power supply module, to contribute to IOP failures.

The incident unit was tested at the manufacturer's (Thales Aerospace) workshop under supervision of the French BEA in consultation with the AAIB, and initially tested "No Fault Found". The AAIB continued: "As it was not possible to reproduce the IOP failure on the test bench it was considered that an intermittent fault may exist so a further more robust and iterative test schedule was devised and performed on a dedicated systems test rig, to simulate the aircraft environment and flight conditions during the incident. The IOP was subject to long operating periods and varying temperatures on the test rig; an ATP test was also performed before and after each temperature endurance test. Following many iterations of these tests, an intermittent fault was identified."

These extended tests finally identified an intermittent failure of the secondary power supply module's -15V DC power supply. The fault was subsequently reproduced during ambient temperatures and after further tests became permanent. A x-ray examination of the module found cracks in the solder of surface mounted components on one of the circuit boards, especially the number 1 transformer. The AAIB reported: "It was concluded that the cracked solder would have caused intermittent electrical contact in the -15 V DC power supply path and led to the intermittent fault on IOP S/N 364 experienced during the incident flight and repeated during subsequent testing."

As result of determination what indications would be lost with the failure of IOP1 or IOP2 as well as IOP1 and IOP2 together it was found, that the failure of both IOPs together would lose more indications that the sum of failures by either IOP1 or IOP2. The AAIB analysed that the extent of information loss caused by either IOP1, IOP2 and IOP1+IOP2 was not fully reflected in the manuals the AAIB registered a technical query with the manufacturer. As a result the manufacturer responded they would adopt the QRH accordingly following investigation of all failure modes.

The AAIB analysed that following a PULL UP warning by the GPWS a full energy pull-up maneouvre must be flown unless the aircraft was fully visual below 1000 feet AGL with the runway visible, the aircraft established on final approach track, established on correct vertical profile as confirmed by glideslope indication or PAPI/VASI, aircraft is stabilized in landing configuration with approach power set and it is immediately obvious to the crew that the aircraft is not in danger.

The AAIB annotated they were dissatisfied with the incident reporting and handling of the incident, which failed to preserve the cockpit voice recorder after the incident flight. The incident was actually reported to the AAIB only four days after the occurrence at which time the aircraft had completed more flights and the cockpit voice recorder had been overwritten. Six days after the occurrence it became apparent that both flight crew had not correctly reacted to the EGPWS warning by initiating a go-around and both flight crew were grounded for remedial training.

Three safety recommendations were released as result of the report.