HC-SR04 with STM32 gives unexpected values

Question

On an Arduino it is easy to get an HC-SR04 ultrasonic sensor getting to work. However, on my STM32(F103C8T6) I get strange values.

I assumed I connect it correctly, since I get values on the trigger pin. I checked the voltage on the 5V pin going to VCC of the HC-SR04 is 4.87 V.

The following program I used (only most important parts and with generated comments removed):

// PA10 US1 Trigger
#define US1_Trigger_Port GPIOA
#define US1_Trigger_Pin  GPIO_PIN_10
// PA9 US1 Echo
#define US1_Echo_Port GPIOA
#define US1_Echo_Pin  GPIO_PIN_9

uint32_t us1_duration;
uint32_t us1_distance;

uint32_t raw_values[1024];
uint32_t n = 0;

And for the loop (default part, auto generated by CubeMX):

int main(void)
{
  HAL_Init();
  SystemClock_Config();
  MX_GPIO_Init();
  MX_SPI1_Init();
  MX_USART1_UART_Init();
  MX_NVIC_Init();

  while (1)
  {
    HAL_GPIO_WritePin(US1_Trigger_Port, US1_Trigger_Pin, SET);
    HAL_Delay(10);
    HAL_GPIO_WritePin(US1_Trigger_Port, US1_Trigger_Pin, RESET);

    uint8_t state;
    do
    {
      state = HAL_GPIO_ReadPin(US1_Echo_Port, US1_Echo_Pin);
    } while (state == RESET);
    uint32_t us1_start = HAL_GetTick();

    do
    {
      state = HAL_GPIO_ReadPin(US1_Echo_Port, US1_Echo_Pin);
    } while (state == SET);

    uint32_t us1_end = HAL_GetTick();
    us1_duration = us1_end - us1_start;

    us1_distance = (us1_duration / 2) / 29.1; // cm (?)
    raw_values[n] = us1_duration;

    n++;
    if (n == 1024)
    {
      break;
    }

    HAL_Delay(1000);
  }
}

I put a breakpoint in the break statement to stop the program to check values. I moved my hand between 10 and 20 cm (4-8 inches) away from the sensor but the raw values I get are:

Name : raw_values
Details:{195, 196, 1, 1, 196, 1, 4, 1, 1, 1, 197, 0 <repeats 1013 times>}
Default:0x20001070 <raw_values>
Decimal:536875120
Hex:0x20001070
Binary:100000000000000001000001110000
Octal:04000010160

The raw values I expect would be in 10-20cm range (at least most of them), which would result according to the datasheet in raw values of 10 * 58 to 20 * 58 which is 580 to 1060. Especially not around 0's and not all same values around a certain value (this time 200).

The data sheet info I used can be found here: datasheet

Update I checked it with a logic analyzer, and noticed the high parts of the echo durations are less than 1 ms. I don't know why I still get values of 200 sometimes, maybe when there is no echo.

Answer 1

I found out the HAL_getTick returns milliseconds, and the echos are less than 1 ms (sometimes).

However, I changed it just by counting the loop iterations and I get much better values:

Name : raw_values
    Details:{118, 233, 797, 971, 268, 15, 290, 854, 455, 62, 258, 525, 867, 865, 866, 581, 576, 576, 587, 588, 581, 0 <repeats 1003 times>}

The last 8 values (except the 0) I got when I placed a box steadily in front of the sensor.

I think if I calculate the exact duration (in ms) of one loop it should be ok and I forget about the bad resolution of the HAL ticks. Or I should use some hardware timer that I start and check.