Noise evaluation The RAMESSES Experiment III: The experiment

Data reduction

Except at the closest ranges the signal is at or below the receiver noise level and synchronous detection or stacking must be employed. Source fields must therefore be well characterised. The intention was to transmit frequencies of 0.25 Hz, 8 Hz and 64 Hz but a transmitter malfunction modified this.

High voltage 256 Hz power from the ship is fully rectified and the resulting half sinusoids switched in DASI to give blocks of positive and negative polarity. The transmission frequency is controlled by switching the polarity of the rectified signal after a specified number of zero crossings of the 256 Hz carrier have been detected. A preset program within DASI determines the specific frequencies and duration of each frequency burst. Due to instabilities in the 256 Hz carrier signal and errors in the part of DASI controlling the zero crossing count, the polarity of the rectified sinusoids was switched too frequently, thus increasing the frequency of the signal and shortening the length of single frequency bursts. The switching error was larger for the positive going half sinusoids, making the transmitted waveform slightly asymmetric.

The DASI logger, which monitored the transmitted signal, logged for at most three minutes in every half hour, and so provided an incomplete record of transmission. A source model was constructed by interpolating between points at which logger records were available, supplemented by short range data from instrument Quail during the first tow. Transmission frequencies during the first tow were 0.35 Hz and 11 Hz, in burst lengths of 92 seconds. During the second tow the transmission frequency drifted between 0.69 Hz and 0.76 Hz, with an average of 0.75 Hz, in bursts of 44 seconds.

A least squares fit at the transmission frequency during each burst was used to extract the amplitude of the fundamental sinusoid from the raw time series recorded by each receiver channel, and to estimate the error on this value. The phase of the transmitter changed between bursts by an unknown amount, making it impossible to stack several bursts to improve the signal to noise ratio. Because of this, none of the long range data from the LEM instruments, which was at or beneath the ambient noise level, could be extracted. The LEMURs employed a stacking algorithm to reduce data volume, so no data could be recovered from these instruments.

Noise evaluation The RAMESSES Experiment III: The experiment

Go to first page