Some questions about using float, snprintf, sprtinf on STM32

Question

I would like to advise you, I create software for the STM32L151C8 microcontroller that will control the LCD segment display via the I2C PCF85176 driver

Design:

battery powered
configurable by the attached bluetooth module and keeping configuration values in EEPROM / FLASH
pulse ounter on the interrupt and holding the value in the flash, the device does not reset after inserting the battery
displaying on LCD pulse values after simple mathematical operations (addition, multiplication)

My questions:

using the snprintf / sprintf function to convert floats to char array - it uses a lot of memory? using float (this MCU don't have FPU i should using uint and only add dot on LCD?) I choose good MCU for this application?

void LCD_Update(void) {
    uint8_t msg[2+20];
    uint8_t *display_mem=msg+2;
    msg[0]=CMD_OPCODE_DEVICE_SELECT | 0 | CMD_CONTINUE;
    msg[1]=CMD_OPCODE_LOAD_DATA_POINTER | 0;
    memset(display_mem, 0, 20);

    int i,j;
    for(i=0;i<8;i++) {
        for(j=0;j<4;j++) {
            uint8_t nibble=(display_digits[i]>>(4*j))&0xf;
            uint8_t nibbleaddr=digit_addrs[i][j];
            uint8_t byteaddr=nibbleaddr>>1;
            nibbleaddr&=1;
            display_mem[byteaddr]|=nibble<<(4*nibbleaddr);
        }
    }
    HAL_I2C_Master_Transmit(&hi2c1, PCF8576_ADDR, msg, sizeof(msg), 100);
}
 
void LCD_Clear(void) {
    for(int i=0;i<=8;i++) {
        display_digits[i]=0;
    }
    LCD_Update();
}
 
void LCD_Print(char* str) {
    int idx=0;
    for(int i=0;i<=8;i++) {
        char c=str[i];
        if(c>='A' && c<='Z') {
            display_digits[idx]=alpha[c-'A'];
        }
        else if (c=='.') {
            display_digits[idx-1]=numsDot[str[i-1]-'0'];
            idx--;
        }
        else if(c>='0' && c<='9') {
            display_digits[idx]=nums[c-'0'];
        }
        idx++;
    }
    LCD_Update();
}
 
void LCD_PrintInt(int value) {
    char str[8];
    sprintf(str, "%d", value);
    LCD_Print(str);
}
 
void LCD_PrintFloat(float value, uint8_t length) {
    char str[length];
    snprintf(str, length + 1, "%f", value);
    LCD_Print(str);
}

Above is fragment of my program to control LCD

You just have to look to the map file, ie. the file giving the size and location of every section of your executable. — Guillaume Petitjean, Mar 24 '21 at 13:27

score 0 · Answer 1 · answered Mar 24 '21 at 16:37

To address you two questions:

1/ Float point calcuations are achieved by software if the MCU has no FPU. And yes, there will be more code so more memory used. But I believe that would be enough for your application. You can use you current MCU.

2/ You can print directly float using sprintf family functions and print on your LCD. By default this is disabled on STM32 compiler. you need to add the -u _printf_float flag to you linker to enable this feature. You can refer to this forum. This was also answered in another topic

Some questions about using float, snprintf, sprtinf on STM32

1 Answers1