1 Reply Latest reply on Feb 25, 2020 11:28 AM by KyTr_1955226

    PSoC5LP I2C Bootloader NACK on read command (bootloader resetting slave address)?


      Hello all,


      I'm running into a strange problem with bootloading a PSoC5LP device over I2C interface.

      My bootloader host is a PC app using the cyusbserial.dll library.  During testing I was able to get this working, but there seems to be some funnies going on and I'd like to confirm a couple things.


      What I'm finding is that I can jump to the bootloader from the application successfully, and then the "Enter Bootloader" packet is received and ACKed by the PSoC:



      Then about 13ms later, when the host sends the address(R), the PSoC NACKs:



      My question is: does entering the bootloader reset the I2C Slave Address?


      The reason I think this might be the case is that the design I am using has multiple PSoC devices on the I2C bus, and the slave addresses of each are determined at runtime via a switch on the PCB.  The switch controls a set address lines going to the PSoC devices, and the slave address of each is based on the reading from these pins, allowing each processor in the chain to have a unique address.  This is set on startup in both the application and bootloader projects like so:


      #define I2C_BASE_ADDR = 0x40   //0x80 when left-justified
      int main(void){
          SI2C_SlaveSetAddress(I2C_BASE_ADDR + ADR_Read());


      The only reason I can think of where the enter_bootloader command would be ACKed, and the first read at the same address would be NACKed is that the entering the bootloader is changing the I2C Address after I'm setting it.  I'm having trouble figuring out where in code this might be happening.

      SI2C_Start() definitely changes the address to the default set in the component as part of SI2C_Init(), but that's why I'm calling SI2C_SlaveSetAddress() afterward.  From what I can tell, SI2C_Init() should never be called again unless SI2C_Stop() is called first (SI2C_initVar gets reset to 0).


      Can anyone confirm if this is the case?

      Of course, if anyone has an idea as to what else the issue might be, I'm all ears.