[General] Type casting operation integer to 32 bit floating point

[nayakajit87]

I am using arduino Nodemcu and Modbus device to communicate. I have connected the nodemcu tx and rx line to Modbus device TX and rx line . Here i am sending modbus request in hex and could able to read data .

I have attached image for reference. i expected to display the value as per modbus display. The value is is displaying matching with unsigned decimal in modscan32. How could i convert to 32bit float to display actual reading.

method1:

#include <ModbusMaster232.h>
#include <SoftwareSerial.h>
float  data[100];
ModbusMaster232 node(1);

// Define one address for reading
#define address 1
// Define the number of bits to read
#define bitQty 70

void setup()
{
  Serial.begin(9600);
  // Initialize Modbus communication baud rate
  node.begin(9600);


}

void loop()
{

  int result =  node.readHoldingRegisters(address, bitQty);
  data[0] = (float) node.getResponseBuffer(0);
  data[1] = (float)node.getResponseBuffer(1);
  data[2] = (float)node.getResponseBuffer(2);
  data[3] = (float) node.getResponseBuffer(3);
  data[4] = (float) node.getResponseBuffer(4);
  data[5] = (float)node.getResponseBuffer(5);
  for (int i = 0; i < 100; i++)
  {
    //data[i] = node.getResponseBuffer(i);
    Serial.println(data[i]);
  }
  Serial.println("............");



}

type casting tried
method2:

void loop()
{

  int result =  node.readHoldingRegisters(address, bitQty);
  data[0] =float( node.getResponseBuffer(0));
 // long val =(data[3] << 24) | (data[2] << 16) | (data[1] << 8) | data[0];
//  Serial.print("val:");
//  Serial.println(val);
  data[1] =float(node.getResponseBuffer(1));
  data[2] = float(node.getResponseBuffer(2));
  data[3] = float( node.getResponseBuffer(3));
  data[4] =float( node.getResponseBuffer(4));
  data[5] = float(node.getResponseBuffer(5));
}

method 3: it almost matched the reading of modscan32 but not accurate enough to match software reading

void loop()
{

  int result =  node.readHoldingRegisters(address, bitQty);
  data[0] =float( node.getResponseBuffer(0));
 long val =(data[3] << 24) | (data[2] << 16) | (data[1] << 8) | data[0];
 Serial.print("val:");
 Serial.println(val);
}

Only one thing is controller is 8 bit controller and modbus device uses 32 bit precision. I m suspecting the result may not match because of it.
kindly requested to give some suggestion to display the data

 

[FvM]

The MODBUS protocol specification doesn't have a definition how to represent larger data entities than 16 bit. 32 bit integer and float data are however used by custom implementations, but there's no generally agreed format you can rely on. You first need to find out which float representation is used by the MODBUS server you are communicating with.

 

[Easyrider83]

    float f = 3.1415926;
    uint32_t u = *(uint32_t*)&f;