We are using Purple Air sensors to collect data at one minute intervals. We designed a solar charging system that incorporates a 12V 30 W solar panel with a an UltraSmart MPPT Charge Controller optimized for LiFePO4 battery charging. The battery is 12.8 volt 18Ah battery. We regulate output to deliver 5 volt for to power the Purple Air sensor. We are near Cleveland Ohio (41.2795° N)
We wanted to get higher temporal resolution air quality data than are available through the Purple Air API. So we pull data directly from the sensor each minute. We first deployed sensors with this system in June of 2025 and the power setup was sufficient for the system to operate until the start of November when our batteries went dead. My understanding is that the Purple Air sensors are powered continuously including the particle sensors, MCU, and Wi-Fi remain energized continuously. What I want to understand is whether sampling data every minute, as we are doing, is increasing power consumption. And if so, how much power we would save if we change our sampling interval.
We sent a .json request to the IP address of the sensor. We receive a response and then push data from the sensor to our own database. We are on the same local network. We prefer this to using the purple air API because it allows us to choose how often we want to receive data. We want to know, does each .json request actually actually cause meaningful power consumption?
As stated, our problem is that the power demands of the purple air sensor are greater than we can supply with the particular battery and solar system we have created. One thing we just implemented is a low battery voltage cutoff circuit – we shut power to the sensor at a fixed voltage (e.g. 12.0V) and then repower the 12.4 volts. Without this circuit we have been completely depleting the battery. It appears to us that even if the purple air sensor is not receiving enough power to actually full start (i.e. blinking blue lights and capacity to connect on the internet), it seems to continue to consume power. Are we right in thinking that it consumes power even if it is not receiving enough power to fully turn on?
Is there any reason why we should not be able cycle the sensors on and off as a means of conserve power? If we WERE to cycle the sensor on and off, how much time would we need to allow for a) internet connectivity to come back up and b) the sensor to stabilize so that the reading taken will be good?