Melvin's digital garden

Optimising underwater modems

speaker: Shan, Subnero event: Friday Hacks, NUS Hackers ** No RF gets absorbed in a couple of meters ** Subs in WW2 3-300Hz wavelengths of 3500km radius of the earth is 6378km ** underwater modem dry – wet end connector – transducer 14 watts used by the digital components when listenings goal is to reduce that to 3 watts intermediate goal is 6 watts ** snapping shrimps creates a cavitation bubble that produce a sound across a wide range of frequencies when it collapses, acts like a jammer ** ships moving past the modems produces bubbles that move down into the water between the modems speed of sound is 1500m/s in water, water with air reduces the speed to less than 330m/s ** one way ranging for the AUV to figure out how far it is from the modem timestamp each packet, AUV listens to the packet and compares the timestamp with its internal clock assumes precise clocks ** DC-DC converters were over speced and not operating at the highest efficiency ** oven controlled clock crystal ocsillators affected by temperature put it in an oven that maintains at 60 dc draws 2 watts ** elephants in the room

  • FPGA - 1W
  • CPLD - 0.8W complex programmable logic device
  • DSP - 2W
  • Fan - 1.5W replace with OpenCL??

Links to this note