Feuchtraumabzweigdose

This PCB is meant to function as an experimental playground for the OpenBike project. It integrates WiFi and Bluetooth via the ESP32 MCU and LoRaWAN through the ubiquitous RFM95W modem. It also has a dedicated header for a gyro module as well as a switchable power supply for GNSS/GPS modules that can be switched off for deep sleep. It exposes a shared SPI bus, two serial lines and all unused pins of the ESP32, with several RTC interrupt methods available in hardware. A dedicated dynamo-to-Li+ charger as well as an efficient 3.3V LDO to supply the logic can be populated optionally.

All in all it is an attempt to migrate most of the functionality of a TTGO T-Beam while being able to run days or weeks on end on one 18650 cell while mostly being in deep sleep. The T-Beam will usually deplete the cell within 30 hours.

For full functional documentation see docs/feuchtraumabzweigdose_docs.pdf.

The project is named Feuchtraumabzweigdose because it is meant to fit snugly into a waterproof junction box – or Feuchtraumabzweigdose in German :D

Known issues in v0.1

• stk forgot the pullup resistors on the BOOT and RESET buttons respectively. This can be mitigated through botch wire.
• stk probably also forgot a pulldown resistor for the interrupt on the MPU9250 breakout header. Botch wire, again. When using the MPU9250 as a wakeup source, at least within the Arduino framework for quick and dirty first prototyping, apparently you have to actively wait for a bit between setting an external wakeup source and actually putting the device to sleep.
• as of now, the circuit has no undervoltage protection for the 18650. Use protected cells, if possible
• The voltage divider that is meant to be read by the ADC is 1:10, which brings down the measurable range of values quite a bit. Replace R3 with a 1M resistor to bring it down to a simple 1:1 divider, but this also increases the power draw during deep sleep.
• The 10k resistor at the MCP73832's PROG pin (R2) limits charging current to 100mA. Might want to replace this with maybe a 2k resistor (500 mA) or something in between This should not be increased too much
• in v0.1, the MOSFET switching the external power for the GPS (Q1) has pins 2 and 3 switched. A quick fix is to rotate it 180 degrees and switch the Source and Drain pins, then connect the Gate pin to one pad of the unpopulated C1 at the R1/R5 junction. See this image for reference.

Firmware

See stadtulm/feuchtraumabzweigdose-testing