Monthly Archives: February 2016
We have a convenience store within our community. It has a freezer which is used to store stuff such as frozen vegetables, ice cream, chicken, meat etc. Freezer temperature has to be below 18 Degree Celsius to prevent growth of harmful bacteria. This can be a problem in India because of frequent power cuts especially during summers. Monitoring freezers and refrigerators can help in detecting spoiled food and thereby avoiding health risks such as food poisoning.
We built a dual temperature monitoring system using a WiSense WSN1101L low power wireless mesh network node operating in the license free band (865 – 867 MHz) in India. This node has two thermistors. One measures the ambient temperature in the store and the other one measures the temperature inside one of the freezers in the store. The temperature data is periodically measured and sent to WiSight (Our cloud storage and visualization app running on Amazon AWS). The node is configured to run as an RFD (reduced functionality device). It is powered by two AAA batteries. Reduced functionality nodes spend most of the time in deep sleep (consuming a couple of micro-amps of current) waking up periodically (say once in 5 minutes) to sense and (if required) report data to the external world via the coordinator node.
The sensor node (located inside the convenience store) is 3 hops away from the coordinator node. The latter is connected to a Raspberry Pi which is forwarding sensor data to WiSight running on AWS.
Here are a few pics of the prototype in action.
You can see the thermistor inside the freezer in the pic below.
A snapshot of temperature data from the freezer over the past couple of days. The freezer has mostly been at around -23 deg C (set point). You can see two spikes followed by a slow return to the set point. Need to analyze the spikes.
Wireless mesh coverage requires many FFDs (routing nodes) spread throughout a “smart” community especially when there is a high density of sensors / actuators generating considerable traffic. Street lights (mains or solar powered) and power points on terraces can be utilized to provide coverage throughout a community. There are a couple of issues though.
- Mains powered street lights are switched on only when it gets dark so the nodes need to have an additional power source to run during day time.
- In Bangalore, we have power cuts and the situation gets worse in the summers. When utility supply is off, generators take over to supply power to houses but not necessarily to street lights.
This means that nodes powered by mains supply need to have battery backup and the battery must be rechargeable. We are currently testing a separate PSU board which takes in power from two sources. One is a 5 Volts supply derived from mains. The other source is a rechargeable single cell lithium battery with a nominal output voltage of 3.7 Volts. The PSU has automatic switch over functionality. When mains supply is present, it charges the lithium battery. It also has an LDO which outputs 3.1 Volts to power the wireless node.
When mains supply is present or restored, the PSU charges the lithium battery as well as provides power to the wireless node. When mains goes off, the PSU automatically switches over to the Lithium battery. The switch over (either way) is very fast. In our tests with this PSU powering an WSN1101L FFD (average current consumption of 22 mA @ 3.1 V), the FFD does not reboot during switch over. The battery charging algorithm depends on the temperature of the battery. The temperature is measured using a thermistor glued to the battery. We noticed that if outdoor day time temperature is high, charging is suspended till temperature goes down (towards the evening). This is not an issue when using high capacity (2000 mAH) Lithium batteries as there will always be enough juice to keep the node powered till the next recharge window.
Here are some pics of an WSN1101L node powered by the PSU board. The setup is being tested outdoors. The enclosure is weather proof.
Snapshot of Lithium battery voltage over the past few days. Initially, mains was not supplied and you can see the gradual fall in the battery voltage. After mains supply was switched on, the battery voltage rose to around 4.16 V (max charging voltage) and has remained there since there is not much power disruption these days.