I would like to turn the IMP on without turning the wifi part on so I can save battery power. Ideally I want this to last a year to 2 years.
Here is how I envision this to work:
-Check once an hour online to see if there is any updated code (regardless of any state changes)
-Shut wifi off
-Every minute check a sensor and then processor goes to sleep.
-If sensor hits threshhold, turn WIFI on, connect and post new data online.
-disconnect WIFI go to full sleep
-Rinse and Repeat
Yes - when the imp was designed, power consumption was an important feature, as most WiFi products are also likely battery products.
We expose a number of APIs in the imp that allow you to go into a deep sleep (the imp draws 4uA while in deep sleep, so your batteries will naturally discharge quicker than the imp draws power), set lower power modes for WiFi, and control when it is connected disconnected.
The Nora reference design is designed to work exactly how you describe. Wake up every 15 minutes, sample data, and go back to sleep.
The Nora can get ~60K wake and connect cycles using 2 AA Alkaline batteries (that’s a wake every 8 minutes or so for a year).
If you’re only connecting when sensor values reach certain thresholds you can get even better battery life. A lot of sensor ICs allow you to set “interrupts” that can toggle pins when certain conditions are met. If you select the proper ICs, you could potentially have your imp ONLY wakeup a send information when those conditions are met.
I want the IMP to be the IC that manages the sensor data rather than lets say Arduino sending an IRQ to wake up the Imp when the value is hit.
The IMP wakes from deepsleep with the wifi off unless your code needs to talk to the server. This means you can get a million wake/check/deepsleep cycles if wifi isn’t required. You can batch data using nvtables during deepsleeps. warning:I’m a newbie to all this so double check what I say.