Emerson Exchange 365

Division not giving the same value as multiplication

Hi Everyone,

In some of our modbus landing modules we are using DIV blocks to convert floating point values to 16-bit UINT. In other modules, we are using MLTY.

I noticed that they do not give the same results, see 16-bit UINT OUT parameters on screenshot below:

It seems that MLTY works better, but I need to explain to the customer why we should change from (x / 0.001) to (x * 1000). Any ideas?

As you see, the result is the same also with CALC blocks: division gives an incorrect result.

Istvan

• floating point
• Precision

István Orbán

• 22 Jan 2015 5:42 AM
• Cancel

4 Replies

• AdrianOffield

  

  • 22 Jan 2015 6:06 AM

  Floating point precision is always going to be greater than integers, therefore, dividing a floating point number by another will give you a fraction (then rounded by DeltaV data type conversion), whereas using floating point and multiplication with integers will result in an integer.
  
  Implying decimal places by converting to integers, I would always use multiply.

  • Cancel
• Niklas Flykt

  

  • 22 Jan 2015 6:09 AM

  In reply to AdrianOffield:

  Or the machine is a really, really old Pentium processor machine that had this problem [;)]

  Niklas Flykt 

  Klinkmann Oy
  

  Key Account Manager safety products

  nikfly@gmail.com

  • Cancel
• István Orbán

  

  • 22 Jan 2015 6:11 AM

  In reply to Niklas Flykt:

  Nope, this is an SX controller flashed to v12.3.1 :)
  Adrian, the thing that caught my eye is that the output of the MLTY and the DIV blocks look to be the same (and correct). Only the conversion distorts the value.

  • Cancel
• AdrianOffield

  

  • 22 Jan 2015 8:29 AM

  In reply to István Orbán:

  True, but the OUT parameter is floating point so it is showing you a FP representation. Try performing the division in a calc block and writing to the UINT instead of wiring the OUT to the UINT parameter.

  • Cancel