RFM69HW higher current consumption after battery shortage

[Ivano]

Hi all and many thanks Felix for the library.
I made a simple project like moteino using atmega328p and rfm69hw transceiver. Used fuse L:E2 H:D2 E:FF so 8MHz internal clock and no BOD.
The device sends a datagram every 3 minutes and sleep in the meantime; the consumption in sleep period is about 6uA
I experimented an issue linked to the battery shortage (3x AAA 1,5V) after about a year of use: after changed the batteries, the consumption during the sleep is increased to 50uA.
I've reprogrammed the micro (sometimes I've experimented EEPROM corruption caused by the low battery), but the device still consumes 50uA so needed to change RFM69HW module to return to 6 uA.
It seems the RFM remains corrupted. Does anyone experimented something like this? There is a way to return to the "factory settings" of the RFM?

[Felix]

Hi Ivano,

First, I no longer use HW modules, only HCW. But they should be equivalent in functionality and behavior. These modules do not have a mind of their own, or any kind of memory, non-volatile storage, to store settings or change from any "factory" settings. Once you power them off or reset via RST pin they return all registers and behavior to POR state.

If this is not an OEM Moteino or LowPowerLab custom made board, then I obviously cannot guess or speak to it because I don't know what components were used.

The LDO I commonly use for all Moteinos and most private projects where LDO is needed is the MCP1703, this is about 2uA quiescent current in normal use. If your battery drops below ~3.5v (3.3+dropout) it will start draining a lot more current.
For most low power projects you can do without the LDO if you can use 3V batteries and run the AVR at 8mhz. Like the MotionMote, this uses around 2uA when not actively detecting motion, and it can run on a CR123A for many years.