First there is the sound of a plane's engines. Then a crackling noise, as though the plane has hit a power line. Then, crashing sounds, as parts of the aircraft landed on homes; a loud bang as the plane impacted with the ground; and a few seconds after the crash, people screaming as the plane fuselage skidded down Beech Street and plowed into walls and cars in the neighborhood.
The 11 seconds of Wednesday's plane crash into an East Palo Alto residential neighborhood were recorded by the city's ShotSpotter gun-shot detection system, East Palo Alto Police Chief Ronald Davis announced during a press conference on Thursday.
It is the first time in aviation history that such a recording will be used for forensic purposes, according to Joshua Cawthra, lead investigator for the National Transportation Safety Board, which is looking into the causes of the accident.
Three Tesla Motors employees, pilot Doug Bourn and passengers Brian Finn and Andrew Ingram, were killed in the 8 a.m. crash one minute after takeoff from the Palo Alto Airport. The twin-engine Cessna 310R struck high-tension power lines and a transmission tower at the edge of the Baylands and then slammed into the Beech Street neighborhood. Parts of the plane hit homes, destroying one and damaging three others. No other persons were injured, officials said.
The gun-shot detection technology was developed by Mountain View-based company ShotSpotter. The system triggers only on loud, impulsive noises -- things that go "bang," according to James Bedlock, company president.
The recording starts a few seconds before the sound of the crash.
"When we heard that a plane had crashed in an East Palo Alto neighborhood, all of us at ShotSpotter knew there was a high probability the city's ShotSpotter Gunshot Location System had detected the incident. (Wednesday's) plane crash created such a noise and it did trigger more than one ShotSpotter sensor deployed in East Palo Alto," he said in a company statement.
ShotSpotter filters sounds to separate gun shots from other noise then reports the gunfire and location to police. The system automatically classified the crash as loud and impulsive but not gunfire and did not report the incident in real time to the East Palo Alto Police dispatch.
For forensic purposes, all loud, impulsive noises are logged by ShotSpotter systems, even if they do not trigger an automatic alert, in case those noises needed to be reviewed after-the-fact, he said.
"Once we determined that the system had registered a loud, impulsive, non-gunfire noise at the time of the crash, we assisted the East Palo Alto Police Department with the retrieval and storage of the audio captured by their system's ShotSpotter sensors for the seconds surrounding the impulsive noise (the crash).
"The East Palo Alto Police Department then provided that data to representatives of the National Transportation Safety Board to support their investigation of the crash," he said.
"Because the ShotSpotter sensors each contain a GPS receiver with a precision clock, the NTSB now has a precise, millisecond-by-millisecond recording of the incident, as captured by several ShotSpotter sensors deployed throughout East Palo Alto. In total, five ShotSpotter sensors generated data which contribute consistently to the mathematical location of the crash. The sensors were located at various distances from the crash, the closest being just over 600 feet away and the furthest being roughly 1,500 feet away," he said.
Cawthra said during a late afternoon press conference Thursday that the data will help determine the speed of the engines and whether they were working properly at the time of the crash. A preliminary report will be available on the NTSB website within about five days, he said.