Statement of Dennis Grossi, NTSB

MR. GROSSI: Good morning. In our joint presentation, Mike and I will give an overview, excuse me, of recorded information available to accident investigators.

In keeping with the theme of the symposium, to share the knowledge and experience gained from the use of recorded information, I will cover the technical and regulatory evolution of mandatory recorders, namely, flight data recorders and cockpit voice recorders, while Mike will discuss the role of recorded data, including air traffic control recordings of radar and communications, non-volatile memory devices, ACARS and satellite communications, and how this data played a role in TSB's investigation into the September fatal accident of SwissAir Flight 111 at Peggy's Cove, Halifax, Nova Scotia.

The first generation of flight recorders, the government introduced the recorder standards in the early 1940s and passed the first regulations at that time. This rule, however, was rescinded four years later due to lack of a suitable recorder. The rule was reissued after World War II, in 1947, only to be rescinded once again for the lack of a suitable recorder. It wasn't until 1953 that a commercial recorder was available.

The CAA dutifully dusted off the old rules and reissued them in 1957 with a 1958 compliance date. The first generation recorders used the oscillographic technique to record the information, by actually scribing the traces on to metal foil recording medium. The only parameters recorded were altitude, air speed, heading, vertical acceleration and microphone keying, all recorded as a function of time. Time was recorded by foil movement, which was generally around a tenth of an inch per minute, which means you could conceivably have an entire accident recorded within a tenth of an inch of the recording medium.

These recorders were eventually fitted to state-of-the-art aircraft that became very popular, such as the turbo-jet aircraft, such as the Boeing 727, 737, DC-9s and DC-8s.

The cockpit voice recorder was not introduced until late 1960s. The recorders were based on the conclusions of a 1960 feasibility study. The FAA issued the requirement for a cockpit voice recorder on turbo-jet aircraft by January 1, 1966, and one year later for four-engine reciprocating aircraft. CVRs record the last 30 minutes of cockpit audio on four separate channels, one channel for each crew member and one cockpit area microphone. The CVRs also marked the introduction of magnetic tape as a recording medium.

With the introduction of the CVRs and FDRs, we also had to have crash/fire survivability standards. These standards were set forth in two very similar technical standards, TSO-C-51 for FDRs and TSO-C-84 for CVRs. These standards were based on the very limited crash/fire survivability information, and as a result called for very mild requirements. In fact, the first recorders were actually mounted near the cockpit or in the main wheel wells of the aircraft.

As you can imagine, they didn't fare too well and were quickly moved to the back end of the aircraft. At the same time, the technical standards for crash/fire survivability were increased, and the the crush requirement was included. Fluid immersion was included, and penetration tests were included, and the impact requirements were increased from 100 to a thousand gs.

The introduction of digital recorders in December 1972 sparked one of the major milestones in the evolution of recorders. These recorders for the first time were recording digital information. Unfortunately, the rule change only affected aircraft with a type certification after September 30, 1969, and this included at the time DC-10s, L-1011s, and 747s.

However, these rule changes did not affect some very popular aircraft, like the 727, 737, DC-8, DC-9. These aircraft continued to be fitted with the old obsolete oscillographic type flight recorders.

With the exception of the introduction of digital recorders in 1972, the FDR recording requirements for most air transports remained virtually unchanged for better than 30 years, until flight recorder rule changes were instituted in 1987 and 1988. These rule changes called for the elimination of oscillographic recorders, increase of the original five-parameter recorders to record 11. For newly-manufactured aircraft, we expanded the parameter requirements would now be 28. We also expanded the requirements to include air taxis, and they expanded the CVR requirements to include aircraft as low as six passengers that required two crew members. We also instituted the rule to require the use of hot mike below 18,000 feet.

There were three technological milestones that need to be mentioned. The first occurred as mentioned earlier in the '70s with the introduction of digital recordings.

The second involved the introduction of digital data buses in the early 1980s which improved the quality and quantity of recorded data and greatly reduced the number of sensors dedicated to the FDR.

The next major development was the introduction of solid state recorders in the late '80s. This increased the recorder capacity and reliability and improved crash/fire survivability.

The introduction of solid state recorders also brought about the introduction of greatly-improved recorder standards. These standards were defined by the collective experience of accident investigators and recorder manufacturers from around the world.

These new standards called for two fire tests, where in the past they would only have the one high-intensity fire test. The second test was a low-intensity fire test that was brought about by years and years of experience of failed recorders. We found that the recorders were actually failing under long-duration/low-intensity fires.

We also increased the impact forces to 3,400 gs. The static crush/fluid immersion/seawater immersion and penetration tests remained unchanged. We also added at this time a hydrostatic pressure test to ensure that the solid state recorder memory would in fact survive a depth of 20,000 feet which was the limit of the recovery vehicles.

I might also add on the high-intensity fire test, we added a very strict test protocol. This changed the fire survivability outcome quite dramatically. In the past, the fire test did not have such a protocol.

The 1997 flight data recorder parameter requirements were changed, required new flight recorder parameters. These rule changes came about as a result of two fatal accidents involving Boeing 737 aircraft that were originally delivered with five-parameter recorders.

The NTSB was severely hindered by the lack of flight control parameters, and in fact, the basic premise that flight control surface position could be derived from control inputs and vice versa was also challenged and determined to be invalid.

As a result, airplanes are currently being retrofitted to record control surface position and control inputs. Transports originally fitted with five parameters and then upgraded to 11 parameters are now being required to record 18 parameters.

We also expanded, as you can see, the parameter requirements for new aircraft in stages. Now, for the aircraft manufactured after the year 2002, we'll require 88 parameters. As a footnote, it might be important to note that 88 parameters will take 700 data sources to meet that 88-parameter requirement, and that Boeing has expanded the parameter requirements to approximately 1,100 for this new generation aircraft.

Quickly, I can see I'm running out of time, we'll move on to the -- our next line of recommendations, which called for two-hour solid state recorders fitted with a 10-minute independent auxiliary power supply. You can see a demonstration of this capability in Booth 18 in the Exhibit Hall.

We also are calling for two combined voice and data recorders, and these recommendations were jointly issued by the NTSB and the Transportation Safety Board of Canada as a follow-on to NTSB's investigations and the TSB's investigation into SwissAir Flight 111, which brings me to Mike Poole's presentation.

Mike will go over and review his work in TSB's investigation and will cover that and give you exact examples of how this information has been used in that investigation.

Mike?