64-bit Float Modbus Register

[alekseib]

How to read 64-bit Modbus register (4xIntegers) and convert it to understandable value in Control Expert?

[wdustman]

Assuming you know how to receive the information using the DTM browser (if not let me know). Use a move command to move the modbus information received to any free %MW location. Make a new variable (LREAL) starting at that same %MW. You can then use that variable and view the value.  [I'm on mobile, so let me know if that is enough to help, if not I can provide full details Monday].

[MichaelWisesa]

Problem is Control Expert/Unity Pro doesn't support 64 byte float or double float. We need to translate 64 byte float to 32 byte float manually (create logic) to workout the exponent and mantissa, and lose some precision.

These websites helps you breakdown the 2:

https://www.binaryconvert.com/convert_double.html

https://www.binaryconvert.com/convert_float.html

https://ryanstutorials.net/binary-tutorial/binary-floating-point.php

Essentially you need to remap the exponent from 11 bit to 8 bit and mantissa from 52 bits to 23 bits

[alekseib]

I wrote 4 INTEGERS by bits to 1 STRING. By using LEFT_INT, MID_INT and RIGHT_INT function blocks received sign, exponent and mantissa. How to convert them from STRING to numbers (int, dint, real)?

[MichaelWisesa · Accepted solution]

I did something like this:

1st is to read the data in words, or you can use byte as word instruction

2nd to split the sign, exponent, and mantissa from 64 bit format

3rd to rearrange the sign, exponent, and mantissa to 32 bit format

[riyat.danisman]

Hello,
I'm wondering about something. There should be 4 words in total for 64 bits, but you did it with "Double2_I, Double3_I, Double4_I". Where did you use "Double1_I"? I couldn't see it in the photo.
I'm trying to read a "wh" value from the energy analyzer from the modbus map in order;
801
802
803
804
with a total of 4 INT/Holding Registers.

Does it have to be used as;
Double1_I : 801
Double2_I : 802
Double3_I : 803
Double4_I : 804
?

[PetrBC]

If you're looking for a simple, elegant solution, it is there. At output, you connect WORD_AS_REAL.

(*Converting block - word of 64bit REAL to word of 32bit REAL*)
(*left are outputs word, right are inputs word*)
(*MW64 = most word of 64bit real*)
(*UW64 = upper word of 64bit real*)
(*BW64 = bottom word of 64bit real*)
(*LW64 = least word of 64bit real, not used by converting*)
(*HW32 = high word of 32bit real*)
(*LW32 = low word of 32bit real*)

HW32.15:=MW64.15;(*sign*)

HW32.14:=MW64.14;(*highest exponent*)
HW32.13:=MW64.10;(*copying of exponent*)
HW32.12:=MW64.9;
HW32.11:=MW64.8;
HW32.10:=MW64.7;
HW32.9:=MW64.6;
HW32.8:=MW64.5;
HW32.7:=MW64.4;

HW32.6:=MW64.3;(*copying mantissa*)
HW32.5:=MW64.2;
HW32.4:=MW64.1;
HW32.3:=MW64.0;
HW32.2:=UW64.15;
HW32.1:=UW64.14;
HW32.0:=UW64.13;

LW32.15:=UW64.12;(*copying mantissa*)
LW32.14:=UW64.11;
LW32.13:=UW64.10;
LW32.12:=UW64.9;
LW32.11:=UW64.8;
LW32.10:=UW64.7;
LW32.9:=UW64.6;
LW32.8:=UW64.5;
LW32.7:=UW64.4;
LW32.6:=UW64.3;
LW32.5:=UW64.2;
LW32.4:=UW64.1;
LW32.3:=UW64.0;
LW32.2:=BW64.15;
LW32.1:=BW64.14;
LW32.0:=BW64.13;

[ag56]

Hello alekseib, use attached DFB to convert the two most significant registers of a long real into 32 bit real.

The reduced precision of the result will be Negligible in most cases.