PRESS RELEASE FROM THE
FEDERATION INTERNATIONALE DE L'AUTOMOBILE (FIA)

According to LDRA Ltd., the company appointed by the FIA to investigate
Formula One electronic systems, the best evidence is that Benetton
Formula Ltd. was not using "launch control" (an automatic start system)
at the 1994 San Marino Grand Prix. Had the evidence proved they were,
the World Motor Sport Council would have been invited to exclude them
from the World Championship. Given the evidence available, such a course
of action would obviously have been wrong.

To avoid speculation, the report of the FIA Formula One Technical
Delegate submitted to the World Motor Sport Council on 26 July is
attached

Hockenheim, 29 July 1994

Report by the FIA Formula One Technical Delegate on the investigations
carried out on the electrical systems on Car Number 5 in the 1994 San
Marino Grand Prix.


An investigation into the software used in the computer systems of the
cars finishing in the first three places at the 1994 San Marino Grand
Prix was undertaken by Liverpool Data Research Associates Ltd. (LDRA).

LDRA is a company which specializes in the analysis, validation and
verification of highly complex computer software such as that used
in modern civilian and military aircraft and a wide range of safety
critical applications.

On race day (1st May 1994), each of the teams was requested to supply
the source code* for the software on board the car and schematic circuit
diagrams of the electrical system. (Appendix 1 )

One team complied in full with this request and a demonstration of
the complete electrical system was set up with entirely satisfactory
results.

Having received nothing from the other two teams, a fax was sent on 9th
May (Appendix 2) asking for urgent action.

An alternative suggestion was received from Benetton Formula Ltd. In
this letter dated 10th May (Appendix 3), they stated the source codes
could not be made available for commercial reasons.

In a fax to Benetton Formula dated 15th May (Appendix 4), we accepted
this proposal, on the condition that Article 2.6 of the Technical
Regulations was satisfied.

On 27th May we received a detailed program for the demonstration at
Cosworth Engineering. (Appendix 5)

The tests which were scheduled to take place on 28th June were canceled,
by Benetton, after some discussion between Ford and themselves
concerning non-disclosure agreements

By a fax dated 28th June, we again requested the tests take place as a
matter of urgency. (Appendix 6)

The demonstration and tests took place on 6th July. We received a
report from LDRA on 11th July (Appendix 7) which left a number of
unanswered questions which we were advised could only be addressed by
close examination of the source code.

In a letter to Benetton dated 13th July (Appendix 8) we made it clear
the demonstration had been unsatisfactory and we required the source
code for the software.

Following another exchange of letters on the 13th and 14th July
(Appendices 9 and 10) a meeting was set up at the Benetton factory on
19th July, an agenda for which was received on 18th July (Appendix 11)
which gave our advisors full access to all the source code, but only on
Benetton's premises and subject to the instructions set out in Appendix
11.

Analysis of this software, which had been used at the San Marino Grand
Prix, revealed that it included a facility called "launch control". This
is a system which, when armed, allows the driver to initiate a start
with a single action. The system will control the clutch, gear shift and
engine speed fully automatically to a predetermined pattern.

Benetton stated that this system is used only during testing. Benetton
further stated that "it (the system) can only be switched on by
recompilation of the code". This means recompilation of the source code.
Detailed analysis by the LDRA experts of this complex code revealed that
this statement was untrue. "Launch control" could in fact be switched on
using a lap-top personal computer (PC) connected to the gearbox control
unit (GCU).

When confronted with this information, the Benetton representatives
conceded that it was possible to switch on the "launch control" using a
lap-top PC but indicated that the availability of this feature of the
software came as a surprise to them.

In order to enable "launch control", a particular menu with ten options,
has to be selected on the PC screen. "Launch control" is not visibly
listed as an option. The menu was so arranged that, after ten items,
nothing further appeared. If however, the operator scrolled down the
menu beyond the tenth listed option, to option 13, launch control
can be enabled, even though this is not visible on the screen. No
satisfactory explanation was offered for this apparent attempt to
conceal the feature.

Two conditions had to be satisfied before the computer would apply
"launch control": First, the software had to be enabled either by
recompiling the code, which would take some minutes, or by connecting
the lap-top PC as outlined above, which could be done in a matter of
seconds.

Secondly, the driver had to work through a particular sequence of
up-down gear shift paddle positions, a specific gear position had to be
selected and the clutch and throttle pedals had also to be in certain
positions. Only if all these actions were carried out would the "launch
control" become available.

Having thus initiated "launch control", the driver would be able to
make a fully automatic start. Such a start is clearly a driver aid
as it operates the clutch, changes gear and uses traction control by
modulating engine power (by changing ignition or fuel settings), in
response to wheel speed.

When asked why, if this system was only used in testing, such an
elaborate procedure was necessary in order to switch it on, we were told
it was to prevent it being switched on accidentally.

A full copy of the LDRA report of the 9 July meeting can be seen in
Appendix 12.

In the circumstances, I am not satisfied in accordance with Article
2.6 of the Formula One Technical Regulations that car number 5
(M.Schumacher) complied with the Regulations at all times during the
San Marino Grand Prix and I therefore submit this matter to the World
Council for their consideration.

Charlie Whiting
FIA Formula One Technical Delegate


*Source code

Computer instructions are usually called machine code and are
represented internally as a series of noughts and ones known as binary
numbers. This form of instruction is very difficult for humans to
understand, so computer languages have been devised that enable us to
express instructions in a form that is more natural to us. Programs
written in these languages are known as "source code". A computer can
not use them directly but they can be translated to machine code that
it can understand by using another program called a compiler. When the
machine code is loaded into the computer's memory the processor can then
execute the instructions that are described in the source code.