USCG Mishap Investigation - SN 21120 & 21121

By: John C. Evans

This page is dedicated to the USCG & Garrett Mishap Investigation on Engines SN 21120 and 21121.  These Engines came off a USCG HU-25 Aircraft that experienced a double engine failure while flying into a rain storm in the Bermuda Triangle.

USCG Engine Mishap Investigation SN’s 21120 & 21121

    Shortly after delivery of a new USCG HU-25 aircraft with the new LPC bronze polyester abradeable shroud material the aircraft flew into cold rain in the Bermuda Triangle, and experienced a double engine failure.  Both engines began to surge (pop and bang) but continued to run allowing the aircraft to remain aloft and make an emergency landing on an island.  Ground inspection of the engine and multiple engine runs over a three day period were unable to detect or resolve the engines surge (compressor stall) problems. 

    A USCG maintenance team was dispatched to borescope the engine inlet and low pressure compressor.  The borescope inspection revealed that both engines had severe 2nd stage low pressure compressor blade damage including blade airfoils that had broken off at the blade platform.  Two fresh engines were dispatched to the site on a USCG C130, the engines changed and the aircraft returned to service.  The USCG then dispatched a four man team to Garrett to assist in and document tear-down investigation results of both of the damaged engines.  A  USCG “Mishap Investigation Team” led by Captain Scurria and team members LTCDR Ned Peak, J. Currier, G. Gerard, and R. Sanford was dispatched to Garrett Phoenix AZ for engine teardown investigations of engines SN 21120 and 21121.  Two additional U.S. Air Force Wright Patterson Personnel, Paul Deutchle and Brett Hauber assisted the USCG with data analysis.  The author John C. Evans ran the investigation for Garrett, and used the two most experienced ATF3 O&R tear-down mechanics, Will Haynes and Wayne SteeleBill Edwards, Document & Data Management, covered the engine teardowns.  Other Garrett members supporting the engine mishap investigation and corrective actions were T. Abbot, R. Ashby, R. Bleakley FAA DER, R. Clarke, M. James, K. Keegan, L. Meyer, C. Noyes, D. Rogers (R&O Engineering), D. Tyler (Chief Engineer), J. Westerhausen, and G. WoelkiGarrett FAA DER Manager B. Moring covered the investigation for the FAA.  Bill Moring previously worked at FAA Western Region as an Engineer and signed off the ATF3 type certificate for ATF3 engines.  The failures of both engines were identical, 2nd stage low pressure compressor blade failures with downstream domestic object damage.

    This failure mode appeared to be an infant mortality problem.  Engines either failed very early in their life or ran forever.  AMD Dassault discovered a way to screen engines before aircraft delivery.  They would climb to high altitudes where the air was very cold, and decelerate the engines to flight idle.  After 30-seconds they would perform a snap acceleration and some engines would fail while others could be delivered with confidence that they would not fail.  While this flight test procedure resolved Dassault’s engine reliability problem, it did not resolve Garrett’s problem.

    Garrett’s conclusion was that the very hard bronze/polyester based LPC abradeable material eliminated the USCG corrosion problem, but presented a new problem.  The shroud material was much harder than the previous aluminum/polyester material (and not as abradeable).  When the USCG flew into cold rain the super cooling of the compressor shrouds caused them to shrink into and break the LPC 2nd stage blades.  The broken blades would then proceed downstream through the engine causing substantial hardware damage.  Shortening the 2nd stage compressor blades by 0.015-inches and reinforcing the Titanium 1st & 2nd stage compressor shrouds with stiffening rings on the shroud outer diameter corrected this problem.  

    Garrett engineer Mike James conducted high altitude flight testing on an instrumented ATF3 engine on the Garrett Falcon 20C flying test bed.  Mike discovered that during a recreation of the Dassault Screening test the 2nd LPC exhibited a transient decrease in 2nd stage blade tip clearance of about 0.020-inches, resulting in a very heavy blade tip rub.  The rub was sever enough to bend blades past the material yield threshold.  The decision was to increase blade tip clearance by immediately decreasing 2nd stage blade length’s by 0.015-inches.  Subsequent back to back performance testing indicated that the expected performance loss was so small as to be undetectable.  The blade tip reduction corrected the blade breakage problem but Garrett decided to use the belt-and-suspenders approach.  The Titanium 1st and 2nd LPC shrouds were both stiffened with three ribs on the outer diameters to prevent subsequent shroud induced blade failures.  The 3rd, 4th, & 5th LPC blade tip shouds were made of steel with much lower thermal expansion rates and deemed not to require redesign.  After the fixes were installed ATF3 engines never experienced another LPC blade failure.

Additional ATF3 engine design improvements identified.

    Garrett Repair & Overhaul Engineering and ATF3 Project Engineering recognized that much of the hardware in the ATF3 engine turbine section was heat distressed requiring replacement and negatively impacting engine durability and adding to the repair cost and of ownership.

    Assistant Project Engineer Fred Fuller took the lead in addressing and resolving these and any other ATF3 shortcomings that could be identified.  Fred’s group consisted of three engineers, Kurt lammon (a recent University of Florida engineering graduate), Greg Hansen, and the author John C. Evans.  The efforts of this team would ultimately led to the “Silver Bullet Modifications.”  This name was chosen “tongue in cheek” to kill any hardware sucking vampires still lurking inside the ATF3 engine.


Created on: 6/9/2009, Updated 7/10/2009

Please email me with any suggestions or corrections on this page.