espressif / esp-idf

Espressif IoT Development Framework. Official development framework for Espressif SoCs.
Apache License 2.0
13.63k stars 7.28k forks source link

Esp32 Reboot when GSM Module is connecting to network (IDFGH-4060) #5931

Closed dhirajp15 closed 3 years ago

dhirajp15 commented 4 years ago

Hi ,

I am facing the exact same issue. Using GSM module and esp32 together , when GSM module tries to connect to Network the ESP32 reboots with following logs:

rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT) configsip: 0, SPIWP:0xee clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 mode:DIO, clock div:2 load:0x3fff0018,len:4 load:0x3fff001c,len:6336 ho 0 tail 12 room 4 load:0x40078000,len:11308 load:0x40080400,len:6680 entry 0x40080760

Note: This voltage fluctuation is observed at EN pin , while the VCC is does not show any fluctuation.( I assume it is RF noise) . Also when I keep GSM antenna away from PCB , it does not reset.

IMG_20200306_151829

Using two 47uF resistors at EN pin to suppress the noise :

image

Please suggest if there is any better way to fix this issue

Originally posted by @dhirajp15 in https://github.com/espressif/arduino-esp32/issues/3520#issuecomment-701896350

nuclearcat commented 4 years ago

IMO, it is particular PCB layout issue. I have tested ESP32 with SIM800 and SIM7000 and they work perfectly.

dhirajp15 commented 4 years ago

for me testing with Quectel M66 module along with specific Airtel sim this issue is happening . Idea/Vodafone sim is working fine. I am trying to fix it but its difficult find out why there is significant drop of 1V at EN pin and no fluctuations at VCC pin.

gecko242 commented 4 years ago

As @nuclearcat said, this sounds like a layout issue. I have an ESP32 in close proximity with a SIM800 with no problems at all.

david-cermak commented 4 years ago

As said, this seems like a PCB/layout issue. In the posted schematic snippet, there's a net ESP... going somewhere, connected to the EN pin. Is it a reset button? (path leading though the entire board, where is it connected to...?) I'm just guessing here, but the pulses look almost like a "GSM buzz" and the reason why the Vcc net is clean might simply be the voltage regulator on the line..?

Also I don't see any problem adjusting your R/C values as you did to 47uF/1k.

Alvin1Zhang commented 3 years ago

Thanks for reporting, feel free to reopen.