Very low power with alkaline battery

[Artix]

Hello,

I'm building my own pcb for mysensors and i'm doing a lot of tests since 1 year now.

My parts are all selected to work between 1.9 and 3.3V :

• Atmega328p (100na interrupt powerdown, 4.3uA watchdog powerdown)
• rfm69 (100na powerdown)
• SI7021 (powered by pin)
• OPT3001 (powered by pin)
  etc...

To power my pcb i have done two tests :

Something like https://hallard.me/category/ulpnode/ where i'm able to get less than 1uA (between 400nA and 1uA) of current when sleeping using only 1xAAA battery
Used/tested/documented by @scalz too.

Directly powered by 2xAAA battery where i get ~4.3uA of current when sleeping (due to 328p watchdog)

Then i checked about alkaline discharge rate... that's < 0.3%/month (wikipedia) I take an optimistic value of 0.15%

For a 1200mAh AAA battery that's : (0.15/100)1200 / (3024) = 0.0025mAh per battery.

For the first test it's 2.5uA self-discharge for <1uA consumption
For the second test that's 5uA self-discharge for 4.3uA consumption

Now i'm asking myself if it's worth trying to go under battery self-discharge rate for sleeping node ?

[gohan]

If you go around 6 or 7 uA you are still going for a long battery life (depending how often you are sending data)