Convert uint16 to float32

Using Arduino Networking, Protocols, and Devices
1

Hi all,

i'm a newbie to arduino.

How to convert 2 bytes of uint16 to float32 value.

These values i'll be receiving from my modbus device.

thanks

2

The compiler does that for you.

unsigned int x;
float f;

x = 432;
f = x;

// Sometimes you want to force the compiler to convert the types.
// But that is not needed in this example.
f = (float) x;
3

if it is used e.g. as a parameter for a function you can also multiply * 1.0

unsigned int x = 355;

float f = sqrt(x*1.0);

The other way around is not always possible as

  • the range of float exceeds that from ints
  • floats can have a decimal part that cannot be represented with an int.
4

Hello,

From my meter i'm receiving two uint16 values which combine to make a float32 value.

a=13667;
b=17992;

in python i used struct.pack and struct.unpack to get float values.

First i packed the int values to Hex and then unpacked it to float. But i used to specify endian also.

How to do this operation in arduino Leonardo.

5 
// untested example

union
{
  uint16_t u[2];
  float f;
} meterdata;

meterdata.u[0] = a;
meterdata.u[1] = b;
Serial.println( meterdata.f);
6

Hello Erdin,

Thanks for the reply.

The data converted is for little endian.

How to do it for big endian.

The data converted for little endian is correct.

7

Do you know what a union is ?
Every item occupies the same memory location. So the 4 bytes of the array are at the same memory location as the 4 bytes of the float.

To change them, you could do this

...
meterdata.u[0] = b;
meterdata.u[1] = a;
...

This was an example with a union.
It is also possible to do it directly in the memory location of a float variable, or with a unsigned long in between the conversion.

I'm not 100% sure about your data. Are the uint16 variables the 16 upper and 16 lower bits of a floating point variable ? So the actual integer value of those uint16 has no meaning ?