What happened
On 28 January 1999, a Canadair CL600-2B19, registration G-MSKK, was performing a scheduled passenger flight from Birmingham to Belfast. While climbing through FL 200 at a speed of 240 kt, the aircraft's autopilot engaged, the central glass ply of the right-hand main windscreen suddenly failed with a loud cracking sound. Despite this failure, the remaining plies of the windscreen held their integrity, ensuring that the aircraft's pressurisation remained stable.
Upon the failure, the crew immediately levelled the aircraft. The first officer declared a PAN-PAN and requested vectors to return to the departure airfield. The captain transferred control to the first officer and initiated the Quick Reference Handbook (QRH) procedures for a shattered windscreen. After informing the 25 passengers and 5 crew members of the situation, the aircraft returned to Birmingham, where it landed without further incident. There were no injuries to anyone on board.
The investigation
The investigation examined the structural integrity of the windscreen and the broader history of similar incidents within the fleet. It was noted that the operator had experienced three separate windscreen centre ply failures across various main and side windows since November 1998. Furthermore, the investigation identified that at least 18 other operators had been impacted by similar issues since 1994.
Initial theories suggested that the failures were caused by manufacturing defects, such as drilling errors or improperly applied Teflon tape that left glass edges exposed. However, the investigation established that the root cause was more likely related to disbonds and voids within the laminated screen. While newer windscreen units have shown a reduced defect rate, some failures have persisted.
Manufacturer records from Bombardier indicated that 190 CRJ windscreens had been replaced due to such issues, though none of these specific instances resulted in a loss of cabin pressure. The manufacturer and windscreen producers have been conducting ongoing research into these early-life failures, which have shown no specific correlation to aircraft age, geography, or operator.