Wireless Sensor Networks:
Partitioning Power
Wireless Sensor Networks (WSNs) generally consist of miniature,
self-contained, electronic modules laden with sensors for recording
environmental parameters. Power is usually the
constraining factor of many WSN implementations.
One effective method for minimising power
consumption is to partition the power supply into domains allowing the
microcontroller to selectively power various units:
- The microcontroller should be powered via a low-dropout voltage
regulator (LDO) with very low quiescent current and high power supply
rejection ratio. DC-DC conversion is not usually recommended due to their
relatively high quiescent current compared to the microcontrollers’ sleep
mode current.
- Sensors with digital outputs should be powered by the
microcontroller’s I/O pins or via LDOs. The sensors have to be tristated from communication buses to prevent vampire
powered.
- Sensors requiring signal conditioning should be powered via LDOs.
DC-DC conversion may be necessary for sensors with excitation voltages
higher than the source. An LDO should be connected in series with the
DC-DCs’ output to reduce power line ripple.
- The communication module should be powered via LDO, and adding
DC-DC conversion where necessary.
The added benefit of power partitioning is the reduction of noise when
reading sensitive analogue sensors.
Read the rest of the article here.
Dr Royan Ong
DreamCatcher
Technical Training
April 2010