BHI160 sends no data

[lkamp]

Hello,

I am trying to get a BHI160 working with a STM32 microcontroller. My code is based on the sample code in the GitHub repository. I have attached my code at the bottom of this post.

Communicating with the chip seems to work fine. bhy_driver_init() returns 0. However, the INT pin never changes, so my code is stuck in the loops that wait for an edge on INT. If I remove the loops and start reading FIFO data right away, no data is received. I expect an initialization meta message and the sensor data from the accelerometer, but data_read is always 0.

int ret;

uint8_t fifo[100];
uint8_t                    *fifoptr           = NULL;
uint8_t                    bytes_left_in_fifo = 0;
uint16_t                   bytes_remaining    = 0;
uint16_t                   bytes_read         = 0;
bhy_data_generic_t         fifo_packet;
bhy_data_type_t            packet_type;

if(bhy_driver_init(bhy1_fw)) {
    return EB_ERROR;
}

while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_SET) ;
while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_RESET) ;

ret = bhy_enable_virtual_sensor(VS_TYPE_ACCELEROMETER, VS_WAKEUP, 10, 0, VS_FLUSH_NONE, 0, 0);

while(1) {
    ret = bhy_read_fifo(fifo + bytes_left_in_fifo, 100 - bytes_left_in_fifo, &bytes_read, &bytes_remaining);

    bytes_read           += bytes_left_in_fifo;
    fifoptr              = fifo;
    packet_type          = BHY_DATA_TYPE_PADDING;

    do {
        /* this function will call callbacks that are registered */
        ret = bhy_parse_next_fifo_packet(&fifoptr, &bytes_read, &fifo_packet, &packet_type);

        /* the logic here is that if doing a partial parsing of the fifo, then we should not parse  */
        /* the last 18 bytes (max length of a packet) so that we don't try to parse an incomplete   */
        /* packet */
    } while ((ret == BHY_SUCCESS) && (bytes_read > (bytes_remaining ? 18 : 0)));

    bytes_left_in_fifo = 0;

    if (bytes_remaining) {
        /* shifts the remaining bytes to the beginning of the buffer */
        while (bytes_left_in_fifo < bytes_read) {
            fifo[bytes_left_in_fifo++] = *(fifoptr++);
        }
    }
}

[o_o]

When the BHI bootloader is ready to receive code, the INT pin goes high, so you should add the following before the bhy_driver_init:

while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_RESET) ;

After the CPU run request (at the end of the init function) the INT pin goes low, then back high with the first interrupt. On a slow microcontroller, it is possibly that the INT pin is already low before you reach this line:

while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_SET) ;

In which case, I recommend to switch to a delay instead, before waiting for the INT pin to go back high.

**Note: On BHI160 there is a bug in the ROM which causes a crash when if the hardware configuration does not match the firmware (e.g. if expecting a magnetometer that is not present or broken). In that case, there will be no interrupt. So to be safe, always use the Bosch_PCB_7183_di03_BMI160-7183 firmware for initial testing on BHI160B and Bosch_PCB_7183_di01_BMI160-7183 on BHI160.

[lkamp]

I now have this code:

	while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_RESET) ;

	if(bhy_driver_init(bhy1_fw)) {
		return EB_ERROR;
	}

	DelayMs(100);
	while(HAL_GPIO_ReadPin(INT_BHI_GPIO_Port, INT_BHI_Pin) == GPIO_PIN_RESET) ;

The first loop finishes and bhy_driver_init() returns 0. But then the second loop runs forever, indicating I never receive anything in the FIFO, not even a message that initialization has finished.

[o_o]

1- Which board are you using ?
2- Which part number are you using?
3- Which firmware are you using ?
4- How did you convert the firmware file to C code ?

The #1 source of issues in that you are programming the incorrect firmware and that it fails to boot, because the initial self-test fails.

[lkamp]

1: I am using a custom board, the interface between the BHI160 and the microcontroller (STM32) is based on the schematic in the datasheet.

2: The number on the IC is 020 QP, I use BHI160, not the B version.

3: I use the firmware that is in the file "Bosch_PCB_7183_di01_BMI160-7183_di01.2.1.10836_170103.h" from the GitHub repository. Flashing the firmware seems to work, and the CRC checks work fine.

4: I did not convert the firmware myself, because I use the .h file from GitHub.

I attached a signal capture during the init process. The blue trace is SDA between STM32 and BHI160, the yellow trace is INT. I wait 1 second before starting the init process. As you can see, INT goes low for a short time, and then goes back to high. But that is still during the transfer of the firmware. After the transfer is complete, INT stays low.