HISTORY OF FLIGHTOn October 25, 2018, about 1119 eastern daylight time, a Piper PA-31T, N555PM, was presumed to have impacted the Atlantic Ocean about 100 miles southeast of Charleston, South Carolina. The two pilots and three passengers were not found and presumed fatally injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. The family of the pilot/owner reported that the airplane departed from its home base, a private runway in Andrews, South Carolina, destined for Governors Harbour Airport (MYEM), Governors Harbour, Bahamas. Radar and air traffic control data provided by the Federal Aviation Administration (FAA) showed the airplane departed the private runway toward the southeast about 1047. The airplane crossed over the coastline and began a climb to an assigned altitude of 25,000 ft. The climb rate was consistent at 500 ft per minute (fpm), and the airplane remained on course flying toward the assigned airspace fix. When the airplane was about 12 miles from the fix (about 95 miles southeast of Charleston Air Force Base/International Airport (CHS), Charleston, South Carolina), while climbing through 24,300 ft, the pilot made a garbled radio transmission (that had a considerably increased background noise level) that included the word "emergency". The pilot of another airplane on the same frequency relayed that he heard N555PM say that he was diverting to CHS. At the time of the transmission, the airplane had drifted slightly to the right (less than 1/4 mile) of course. The airplane then began a descent about 1,000 fpm and returned on course toward the fix. About 23 seconds after the garbled emergency transmission, following several requests from air traffic control requests to repeat it, the pilot (or possibly the co-pilot) stated, "we're descending." About 15 seconds later, at an altitude of about 23,500 ft, the airplane turned sharply toward the left, and the descent rate increased to greater than 4,000 fpm. About 25 seconds after that, the radar data's altitude parameter became invalid; the last reported altitude was 21,500 ft. About 35 seconds later, at 1118:33, the pilot transmitted "emergency emergency," and no further transmissions were recorded. The last radar position was recorded at 1118:50, about 3 miles to the left (northeast) of the airplane's original course toward the fix. That position corresponded to a location about 100 nautical miles east southeast of CHS. The FAA issued an alert notice (ALNOT) and a search effort was conducted by the U.S. Coast Guard. One of the search airplanes reported an oil sheen on the surface of the water near the last known coordinates; however, neither the airplane nor debris were located. A nearby commercial tanker vessel was requested by the Coast Guard to divert to the area and search. Personnel reported briefly observing a white object just beneath the water surface, about 4.5 miles east of the last recorded radar position, but they were unable to identify it. The search effort was cancelled 2 days later at sunset. PERSONNEL INFORMATIONThe airplane was certificated for single pilot operation. The second pilot on board, who also held a flight instructor certificate, was serving as co-pilot. AIRCRAFT INFORMATIONA review of airplane records revealed that the airplane's most recent annual inspection included routine maintenance, the replacement of the starter generators on both engines, replacement of the cabin oxygen bottle, and compliance with several airworthiness directive (AD) inspections, including AD 2017-02-06, which addressed a potential issue with electrical wiring arcing and fire risk. The airplane was equipped with an airframe and propeller deicing system, an autopilot, airborne weather radar, and an satellite-based weather data receiver. METEOROLOGICAL INFORMATIONA review of weather records revealed that there were no convective or precipitation echoes in the area at the time of the accident. Satellite imagery depicted a mid-level layer of clouds in the area with tops estimated at 15,500 ft. An AIRMET advisory for moderate turbulence was in effect for the region. Atmospheric model results characterized the atmosphere as stable, with a freezing level around 13,000 ft. Icing conditions were likely between the freezing level and the estimated cloud tops. Additionally, a shallow layer favorable for light rime icing was present at 23,000 ft. Pilot reports from the region indicated light to moderate turbulence. AIRPORT INFORMATIONA review of airplane records revealed that the airplane's most recent annual inspection included routine maintenance, the replacement of the starter generators on both engines, replacement of the cabin oxygen bottle, and compliance with several airworthiness directive (AD) inspections, including AD 2017-02-06, which addressed a potential issue with electrical wiring arcing and fire risk. The airplane was equipped with an airframe and propeller deicing system, an autopilot, airborne weather radar, and an satellite-based weather data receiver. ADDITIONAL INFORMATIONAt the time of the pilot's initial emergency radio transmission, the air traffic control audio recording included the sound of an emergency locator transmitter (ELT) 'homing' signal, obscuring the pilot's voice. The only intelligible portion of the transmission was the call sign and the word "emergency." The pilot's intention to return to CHS was overheard by the pilot of another airplane in the area and relayed back to the ATC controller. The ATC communications recording provided by the FAA included audio sources from all of the radio frequencies and land line telephones used and monitored by the controllers during their shifts, mixed together into a single recorded channel or stream. As a result, the audio from any one radio frequency (or other source, such as land line telephone calls) cannot be separated or isolated from any other source during playback. In this case, the sound of the pilot's voice could not be separated from the simultaneous sound of the ELT signal. Efforts to filter and reduce the ELT noise failed to enhance the intelligibility or clarity of the pilot's initial "emergency" radio transmission. Further, the frequency on which the ELT signal was transmitting, could not be determined. Air traffic controllers monitor both the 121.5 and 243 MHz emergency frequencies through a network of remote antennas on the ground, which correspond to the geographic area or "sector" that they control. Because the recording mixed all frequencies together, it was not possible to determine if the ELT was transmitting on either (or both) of these frequencies. All ELTs transmit the same audible "homing signal" which all sound alike. Some military ELTs transmit exclusively on 243 MHz, while all civilian ELTs always transmit on at least 121.5, some also transmit simultaneously on 243 MHz. Some ELTs additionally transmit a unique digital code on 406 MHz, but also transmit the 'homing' signal on 121.5 MHz. The type of ELT installed on the accident airplane was not recorded in the available aircraft records. The ELT signal was recorded intermittently from 1039 (about 7 minutes before the airplane was first detected on radar, as it took off from Andrews, South Carolina) until 1124, about 5 minutes after the airplane's last recorded radar position. The signal was at times present for several seconds, and frequently only for a fraction of a second. The duration of time during which the signal was not heard, varied from a fraction of a second to several minutes. The overall duration spanned nearly the entire recording, which began at 1030 and ended at 1130. The ELT signal was captured only on the recording of the "R52" controller position at the Jacksonville air route traffic control center, which was the 2nd controller position at Jacksonville to talk to N555PM. As the flight continued south into Jacksonville airspace, it first encountered the "R71" controller position. There were no ELT signals heard on the R71 recording. As the flight proceeded further south, it was handed off to the R52 position. According to the Air Force Rescue Coordination Center 2018 annual report, 60 percent of the emergency beacons (about 5,900) detected in 2018 were non-distress related (accidental activation, improper testing, etc).