MoD knew about safety hazard on Chinook MkII

The documents also raise new questions about whether ministers, briefed by officials, have inadvertently misled Parliament over the crash. The documents are likely to be discussed in the House of Lords next Monday.

Peers are due to vote on a motion, put forward by former Labour foreign affairs minister Lord Chalfont, for a select committee to investigate the Chinook crash. If the motion is passed, it will be the first time in recent years that the Lords has formed a select committee.

The crash killed four crew and 25 senior police and intelligence officers. An RAF Board of Inquiry found insufficient evidence to blame the pilots. But an air marshal, supported by his superior, disputed the findings and ruled that there was absolutely no doubt whatsoever that the pilots were grossly negligent. This was in effect the final ruling.

One of two leaked MoD Air Publications referred to a Block One upgrade to the Chinook's Full Authority Digital Control (Fadec) system which was introduced in late 1994 - several months after the crash - to address nine software anomalies.

One of the changes, said the document, was to address the problem of pilots encountering an engine acceleration without their command when the Fadec was in "reversionary mode", a back-up to the system's primary mode.

Another of the changes was designed to deal with the opposite problem of "uncommanded engine shutdown in reversionary".

A Fadec's hydromechanical unit was found in reversionary mode on the crashed Chinook, but this was said to have been a normal consequence of fuel shutdown at or after impact.

In a Parliamentary reply to MP Desmond Swayne in August 1999, armed forces minister John Spellar listed details of the nine Block One software changes without making any mention of uncommanded engine acceleration or shutdown. Spellar said the Block One upgrade was part of a "product improvement programme".

He said one of the software changes addressed an "error" which occurred when "switching to the reversionary system during pre-flight ground checks". Another change was the fitting of a "software filter" to "reduce the number of incorrect computer computations experienced at very low speeds during engine start up".

In a statement to the Public Accounts Committee, published in November last year, the MoD said of the Block One update that it addressed "nuisance faults".

The second leaked document, dated 1997, said that an erroneous signal to the Fadec may limit the engine power available when maximum power is needed or only one of the two engines is running.

The MoD said the problems highlighted by the leaked documents were "unlikely" to affect safety.