Power rail transience when using SPI on battery
Posted: Wed Mar 11, 2020 12:31 pm
Hi everyone,
Working on a small datalogger project using an ESP32-WROOOM that involves an analog sensor signal being recorded and saved to SD card. Having a big power rail spike issue that has been isolated down to the SPI. Setup is as follows:
1. Core → ESP32-WROOM
2. Power circuitry for USB and battery
3. Peripherals:
a. Sensor (Analog input)
b. ePaper screen (SPI)
c. SD card (SPI)
Note:
This issue only occurs when the setup is powered on a 3.7V 1200mAh LiPo battery.
There is no WiFi or Bluetooth in use.
Circuit has been decoupled with bulk caps.
USB: Stable 3v3 and no power dips/ spikes appear during the record procedure.
BATTERY: When the code enters record mode there are significant power dips on the 3v3 rail that appear to be caused by the SPI (SCLK & MOSI).
The only difference between the power input circuitry is that the battery circuit utilises a power trench mosfet whereas the USB goes through a shottky diode.
The power rail spikes are affecting the 3v3 rail to the extent that the sensor readings are getting completely messed up.
Wondering if anyone has experienced something similar? Sense is that it is a battery/ current limiting issue or something to do with the ESP32.
Working on a small datalogger project using an ESP32-WROOOM that involves an analog sensor signal being recorded and saved to SD card. Having a big power rail spike issue that has been isolated down to the SPI. Setup is as follows:
1. Core → ESP32-WROOM
2. Power circuitry for USB and battery
3. Peripherals:
a. Sensor (Analog input)
b. ePaper screen (SPI)
c. SD card (SPI)
- Channel 1 (yellow) is the 3.3V rail, Channel 2 (Blue) is the SPI clock.
- bbfe0a54-e258-4dc7-a4c3-913c2b6b895c.png (28.85 KiB) Viewed 5055 times
Note:
This issue only occurs when the setup is powered on a 3.7V 1200mAh LiPo battery.
There is no WiFi or Bluetooth in use.
Circuit has been decoupled with bulk caps.
USB: Stable 3v3 and no power dips/ spikes appear during the record procedure.
BATTERY: When the code enters record mode there are significant power dips on the 3v3 rail that appear to be caused by the SPI (SCLK & MOSI).
The only difference between the power input circuitry is that the battery circuit utilises a power trench mosfet whereas the USB goes through a shottky diode.
The power rail spikes are affecting the 3v3 rail to the extent that the sensor readings are getting completely messed up.
Wondering if anyone has experienced something similar? Sense is that it is a battery/ current limiting issue or something to do with the ESP32.