I2C error codes F767ZI

[RPatt.11]

I am working to get one F767ZI board (master) to send 1 byte to a second F767ZI (slave).  I'm using 4k external pull-up resistors on SDA and SCL.  I am trapping errors in the master and they are:

After the following code on the master I get "Error 32".

ret = HAL_I2C_IsDeviceReady(&hi2c1, 30,1, 1000);
if ( ret != HAL_OK) {
  uint32_t error_code = HAL_I2C_GetError(&hi2c1); 
  printf("Error %d\r\n", error_code);
  }  

Then, after the following code on the master I get "Error 4".

ret = HAL_I2C_Master_Transmit(&hi2c1, 30, TX_Buffer, 1, 1000); //up to 1 second to try and send
if ( ret != HAL_OK) {
 uint32_t error_code = HAL_I2C_GetError(&hi2c1); 
 printf("Error %d\r\n", error_code);
 }

Since this is my first time ever trying I2C, I'm sure I'm addressing something wrong or some other simple error.

I can't find anything in the HAL/LL reference to tell me what those error codes mean.

thank you, russ

p.s.  My master F767 spits printf statements to UART4 so I can watch it.  My slave F767 spits them to the terminal window in Ozone using SEGGER_RTT.

[RhSilicon]

I don't think it's a good idea to try to read an error from the I2C port only when receiving a value other than HAL_OK.

typedef enum
{
    HAL_OK = 0x00U,
    HAL_ERROR = 0x01U,
    HAL_BUSY = 0x02U,
    HAL_TIMEOUT = 0x03U
} HAL_StatusTypeDef;

When I check a device (STM32F4), I have to shift a bit in the address.

For example, the TMP117 address, which is 0x48, after shifting a bit, becomes 0x90; // 0x48 << 1

[RPatt.11]

I thought the HAL I2C functions  handle the necessary address adjustments internally, including left-shifting the address by 1 and setting the LSB based on the read/write operation. So, you don't need to left-shift the address manually before passing it to the HAL I2C functions. 

[MM..1 · Accepted solution]

Simply your slave dont reply and first error 32 timeout

#define 	HAL_I2C_ERROR_NONE   0x00000000U
#define 	HAL_I2C_ERROR_BERR   0x00000001U
#define 	HAL_I2C_ERROR_ARLO   0x00000002U
#define 	HAL_I2C_ERROR_AF   0x00000004U
#define 	HAL_I2C_ERROR_OVR   0x00000008U
#define 	HAL_I2C_ERROR_DMA   0x00000010U
#define 	HAL_I2C_ERROR_TIMEOUT   0x00000020U

[RPatt.11]

My assumption here was wrong (that I should not use the left-bit-shifted address in the HAL function call).  It requires it.  In my case I assigned a "Primary slave address" to my slave F767 of 30 decimal.  So, when my master F767 send a byte it should have used an address of 60 (that is 30 left-shifted by 1 bit).

HAL_I2C_Master_Transmit(&hi2c1, 60, TX_Buffer, 1, 1000);

 Then my slave could receive it with 

HAL_I2C_Slave_Receive(&hi2c1 ,(uint8_t *)RX_Buffer, 1,3000); //Receiving in Blocking mode, wait 3 seconds before giving up

 Of course, timing is important as the approach above the slave only waits 3 seconds for the master to send.  This can obviously be increased up to HAL_MAX_DELAY which is 49.7 days.  😉

[RhSilicon]

---

@RPatt.11 wrote:

My assumption here was wrong (that I should not use the left-bit-shifted address in the HAL function call).  It requires it.  In my case I assigned a "Primary slave address" to my slave F767 of 30 decimal.  So, when my master F767 send a byte it should have used an address of 60 (that is 30 left-shifted by 1 bit).

HAL_I2C_Master_Transmit(&hi2c1, 60, TX_Buffer, 1, 1000);

 Then my slave could receive it with 

HAL_I2C_Slave_Receive(&hi2c1 ,(uint8_t *)RX_Buffer, 1,3000); //Receiving in Blocking mode, wait 3 seconds before giving up

 Of course, timing is important as the approach above the slave only waits 3 seconds for the master to send.  This can obviously be increased up to HAL_MAX_DELAY which is 49.7 days.  😉

---

I use only 5 milliseconds of timeout.

Line 88:

if(MPU6050_send(&hi2c3, MPU6050_ADDR, data, 2, 5) != HAL_OK)

HAL_StatusTypeDef MPU6050_send(I2C_HandleTypeDef *hi2c, uint16_t DevAddress, uint8_t *pData, uint16_t Size, uint32_t Timeout)
{
HAL_StatusTypeDef result = HAL_I2C_Master_Transmit(hi2c, DevAddress, pData, Size, Timeout);

return result;
}