DAVID LEARMOUNT / LONDON
Why didn't ACAS prevent last week's mid-air crash?
Differences have emerged between versions of what pilots were told and did in the minute before their aircraft collided over Germany on 1 July in Europe's first mid-air collision since 1976 and the world's first which airborne collision avoidance systems (ACAS) should have prevented.
One of the key questions facing investigators studying the collision between a DHL Boeing 757 freighter and a Bashkirian Airlines Tupolev Tu-154M is how two ACAS-equipped aircraft came to descend at the same time instead of being safely directed apart. The 757 pilot reported that he was descending in response to an ACAS direction, but it is not known whether the Tu-154M pilot was descending in response to the controller's issued instruction, or for another reason. German accident investigation agency BFU was last week still trying to establish whether the Tu-154's ACAS was working. If it were not, then that could help explain the sequence of events.
Initial analysis of the pilots' action was based only on the air traffic control voice and radar tapes without the aircraft recorders, meaning that the timeline was uncertain in an investigation in which precise timing is likely to play a major role.
In a statement immediately after the crash, at 23:35 at flight level (FL) 354 (35,400ft/10,500m) over Lake Constance near the Swiss border, the Swiss air traffic services provider Skyguide said Zurich air traffic control centre (ATCC), which had control of both aircraft, had told the Tu-154 to descend 50s before their flight paths were to cross. It insisted this was "normal", but said controllers had to call twice before the pilot began his descent 30s later.
But BFU had a different version of events in its statement on 4 July. It said the descent order was given 44s before the collision and the pilots began the descent 14s later as ATC issued the instruction again.
In addition to investigating ATC tactical planning, pilot responses to ATC and ACAS alerts, the BFU is also studying:
* Why Zurich ATCC switched off its short-term conflict alert (STCA) system that night for maintenance while one of its primary system computers was out of service for uploading new software.
* What difference it made that one of the two duty sector controllers had taken a break at the time of the accident. Zurich ATCC says this is not against regulations.
Swiss authorities are asking if there is a case for prosecuting the controller involved or Skyguide.
The two aircraft had been at FL360 (36,000ft) with their tracks due to cross over Lake Constance. The Tu-154, en route from Moscow to Barcelona, had established contact with Zurich 5min before the impact. The 757 was flying north from Bergamo, Italy, to Brussels. Skyguide says it was necessary for the Tu-154 to be cleared to FL350 to avoid the 757 and because that was the correct level for its continued flight from that point.
The collision, near waypoint AKABI on the 757's track, killed all 71 passengers and crew on both aircraft. A tragedy on the ground was narrowly averted as wreckage missed the town of Uberlingen.
Another problem for investigators is the damage to cockpit voice recorders (CVR) and flight data recorders (FDR), particularly the Tu-154's CVR. Late last week, investigators had recovered CVRs and FDRs from both aircraft, but, according to BFU, the only one in good condition is the 757's FDR.
On 26 June, before the collision, the BFU had completed a report on Skyguide's radar systems which found that the Geneva centre's secondary radar was, when first checked, outside acceptable tolerances for accurate lateral separation. This, the BFU report says, has since been improved. A BFU recommendation is that Skyguide should be able to guarantee that secondary radar in the ATCCs always has a primary radar back-up.