Iridium Data Transfer from North Pole Deployed

1
Iridium Data Transfer from North Pole Deployed
Ocean Flux Buoys Tim Stanton Oceanography
Department, Naval Postgraduate School
2
OVERVIEW

• Flux buoy objectives
• Data transmission requirements
• Hardware Design
• Sampling design
• Iridium Transmission protocol
• Testing / Evaluation
• Sample statistics
• Summary

3

• SCIENTIFIC OBJECTIVES
• Measure turbulent stresses and
• heat fluxes in the ocean mixed layer
• Use direct eddy-correlation techniques
• ltuw,vwgt, ltwTgt
• Return both mean, spectral
• covariances, and scalar spectra to
• look at turbulent length scales
• Optionally bring back raw (4 Hz) sensor
• data sections for QC, sensor evaluation
• and other analyses
• Measure over at least annual cycles with
• selectable sampling (typically 45 minutes
• every 3 hours

4
Data Transmission Requirements

• 2-way coms to provide
• Adaptable sampling strategies
• Selectable power conservation
• Sampling experiments
• Remote diagnostics of sensors and processor
  system
• 20 to 200 Kbytes / day data transmission
• Return basic buoy status, position, ice velocity,
  mean fluxes
• Return selected raw data blocks of (u,v,w,T,S)
• Fallback to summary messages via Argos

5

• Hardware Design
• 8 port switched power
• RS232 main processor
• Iridium modem (primary,
• 2-way comms)
• Argos ptt (secondary, 1 way)
• Snow / ice / cold tolerance

6
Main Buoy Electronics 68332-based
controller (Onset Model 8) High efficiency, low
noise switched- mode power supplies for each of
8 ports Both solar panel/gell cell and 450 AH,
10.5 volt lithium battery pack support NAL
9500 Iridium modem
7
Sampling Design

• Remote selection of data types to output
• Remote selection of sample intervals, sample
  duration, sub-intervals for flux estimation and
  reporting intervals
• Remote programming of sample-doubling threshold
  (eg if currents exceed 15 cm/s, half the off-time
• Remote monitoring of buoy performance parameters
  including battery voltages, mean and peak current
  consumption, upper and lower buoy internal
  temperature
• Settings and monitor values updated with every
  data transmission (typically 2/day)

8
Data Communication Solution (2001 design)

• Iridium direct dial-in to a NPS workstation
• Quick connect, low overhead protocol
• Relatively simple software design
• Under local control (early days of Iridiums
  come-back)
• Build in tolerance to no connects, dropped
  connects
• Build in a hand-shaking block by block data
  transfer protocol
• Large rotary buffer for outbound blocks to
  overcome service drops
• Fall-back to one way, summary data transfer via
  Argos

9
(No Transcript)
10
Testing / Evaluation

• As usual, short testing cycle before the first
  NPEO deployment
• Needed to evaluate the main processor / power
  board, NAL modem and the 3000 line main program
• End to end test sequences of pseudo-data to
  establish connect and dropped connect performance
• Needed nearly open horizon conditions to simulate
  ice-deployed conditions
• Extreme cold performance tested with a cryogenic
  freezer on top of our building (still there)
• Difficult to test long fall-back protocols

11
Iridium Data Transfer Performance

• Statistics for a 6 month period on the 2003 buoy
• Dial-in success rate 94.9
• Of 350 connect attempts, 325 had full data
  transfer success, 12 had no successful data
  transmission, 13 transmitted at least 1 data
  block.
• 93 of calls were fully successful, 7 had
  dropped calls
• Effective throughput on 25Kbyte transfers was
  1979 baud
• Effective baud rate on 132Kbyte transfers was
  2949 baud

12
Full buoy drift since the April 2002 deployment,
and the winter mix-down
13
SUMMARY

• Iridium is an excellent solution for the 20-200
  Kbyte / day
• data transfer requirement in polar regions
• The direct dial-in protocol was quick to develop
  and effective, but does not scale well to large
  deployments
• Two-way coms exploited to provide adaptive
  sampling and diagnostic capabilities
• Care needed with snow/ice covering of the antenna
• It would be great if there were a slow-charge
  method for the super capacitor in the 9500 This
  is an unnecessary burden on batteries / switchers
  at turn-on.