Bytes to Float not working

[Marcos_Eztronic]

Hello all.

I am trying to convert some bytes that I receive from CAN to a float variable.

However, the result is not correctly. I tried to change from little endian to big endian, not works.

For example, I am receiving the data 0x40333333, it`s suposed to be 2.8, but the result is totally different.

The images below will help to explain the situation (the variable that I am focusing in this example is cellParams.underVoltage

- variable initialiazion

cellParams.underVoltage = 2.80;

- initialization confirmation:

So after the initialization, I will send a CAN command to change the value (I will send the same value)

The new value is in the image below, from buffer position 3 until 6

When I receive the buffer, I try to do the following:

int dummyInt = DeserializeInt32(&cmdMsg->data[3]);
//cellParams.underVoltage = atof(&cmdMsg->data[3]); (same problem...)
cellParams.underVoltage = (float)(dummyInt);

static uint32_t DeserializeUint32(const uint8_t *buffer)

{

    int value = 0;

    value += (int)buffer[3] << 24U;

    value += (int)buffer[2] << 16U;

    value += (int)buffer[1] << 8U;

    value += (int)buffer[0];

    return value;

}

the returned value to dummyInt is OK:

The float value, before cellParams.underVoltage = (float)(dummyInt), is:

And after cellParams.underVoltage = (float)(dummyInt), is:

but if I check the variable (not in memory monitor), the hex value looks fine, but the float value lookw incorrect:

PS.: I tried atof, memcpy, not success.

Could someone help me? Thanks in advance!

[jiri_kral]

Hi, 

I tried it on my side and only working method for me is memcpy - like this: 

unsigned int counter = 0x40333333;
float f=0.0;
memcpy(&f,&counter,4); 

[Marcos_Eztronic · Accepted solution]

Hello, thank you so much for the answer.

Memcpy not worked for me, but I found another solution (thats similar to memcpy, but that`s the onlyone that worked...)

static float DeserializeFloat(const uint8_t *buffer)

{

uint32_t value = 0;

value |= buffer[3] << 24U;

value |= buffer[2] << 16U;

value |= buffer[1] << 8U;

value |= buffer[0];

float f = *((float*)&value);

return f;

}

Anyway, thank you very much for your time.