This is what Flight International wrote:
Flight International issue 15/2005
---
Wake turbulence led to overstressing and $3.6 million-worth of damage to airframe
A flight control system flaw allowed a Lockheed Martin/Boeing F/A-22 Raptor to exceed its load factor limits, seriously overstressing the airframe, after encountering wake turbulence from the target aircraft during air-to-air tracking tests. Flight-control software is being modified following the 28 September incident.
The stealth fighter reached peaks of +10.1g and -11.7g, and angle of attack (AoA) exceeded -60°, during a divergent oscillation that lasted little more than 3s, says the US Air Force investigation report. The load-factor limits for the test configuration, with two external fuel tanks, are +7.33/-0.5g, and AoA is limited to +26/-10°.
The aircraft landed uneventfully after automatically recovering from the incident, but the cost of repairing damage to the airframe – which includes overstressing wing and fin roots to more than 170% of design limit load – is estimated at $3.6 million. The USAF has not decided whether to repair the F/A-22 – Raptor 4003, one of the early aircraft assigned to the development test fleet at Edwards AFB, California.
The F/A-22 crossed the wake vortex of the Lockheed Martin F-16 target aircraft while repositioning for an air-to-air tracking test at 8,500ft (2,600m) and 500kt (925km/h) to determine the fighter’s handling qualities with two external tanks. Triggered by the wake encounter, the fly-by-wire fighter began a divergent oscillation with increasing positive and negative pitch rates, load factors and AoAs.
Throughout the event, the Raptor’s horizontal stabilator was moving trailing-edge up and trailing-edge down at maximum rate, but was unable to stop the chain of events. The pilot’s control inputs – while appropriate for the sensed motion, investigators say – ended up 180° out of phase with g, and contributed to the oscillation’s severity.
When the aircraft exceeded -36° AoA, the flight-control system initiated an auto-recovery as programmed, disregarding the pilot’s inputs and commanding constant full trailing-edge up stabilator, and – roughly 8.5s after the upset – the Raptor recovered to 1g flight.
Multiple previous wake encounters had showed no divergent oscillation, but investigators concluded the flight-control system was deficient because, despite having AoA and g limiters, it allowed the aircraft to exceed those limits and placed the pilot in a situation he should not have encountered.
