AINS-70 Tells All
It was very early December, 1979, when a large wooden crate arrived at McDonnell Douglas in Long Beach, Calif, from the ANZ Antarctica crash. The Douglas cognizant engineer for AINS-70 and the local Collins rep opened it. The Collins Ground Proximity Warning Computer and a Navigation Computer Unit were inside, padded by clothing. The shipment had been from Christchurch, but it made the engineers wonder if the clothing hadn't come from the crash site.
It was later learned the onsite investigators had no real idea what they had found, but the units looked important.
The Collins GPWS computer has a row of four or five latching magnetic indicators that trip when a warning is sounded. That was its nonvolatile memory. The engineers examined it just long enough to determine that the expected warnings had been given.
On to the NCU - a busted mess it was. The front was missing; The whole left side containing the circuit cards was bashed in; the back was hanging by about 300 wires that ran between the innards and the rear connectors. The power supply modules on the right rear half appeared physically intact, as were the two 8K magnetic core memory modules just in ahead of them.
Again, Non-Volatile Memory in the NCU consisted of two magnetic core memory modules, roughly 8 inches tall, 3 inches wide, and six inches deep. They had the appearance of aluminum bricks. Each one had a capacity of 8 kilobytes of data, and they were 4-bit bytes, IIRC. Each bit of data was held by a tiny ferrite donut with five wires passing through it, and all the donuts were arrayed in a three dimensional grid. Being magnetic ferrites suspended by wires, they were not expected to survive a severe mechanical shock.
While the DAC engineer looked on, the Collins rep gingerly removed the two core modules from the battered NCU. The DAC engineer put them in his briefcase and the next day flew them to Cedar Rapids, Iowa, and the Collins factory.
The Collins team applied just enough voltage to tickle the memory modules, and copied their contents onto two like modules.
The copies were then installed into an NCU and the system powered up. Amazingly, the modules revealed the whole last half hour of the flight; the (31,000?) cruise altitude, the race track while descending, the 260 knots at 700 feet over the ocean, the latitude and longitude...
With this evidence that the airplane had performed exactly as programmed, there was no point trying to blame the DC-10, McDonnell Douglas, Collins Avionics, or Litton, the maker of the Inertial Sensor Units.
Recovering that flight data was a singular event in the careers of some of those involved.
GB