Dear All,

I have a problem with converting real number into double integer. I used TRUNC to execute the conversion, but I couldn't get the answer which I want at the output of TRUNC function. For example, I want to convert real number 4294967295.3, and I want to get double integer 4294967295 after conversion. But at the output of TRUNC function, I was getting the value of 1333788671. 

I understood that it's because conversion of 4294967295.3 into Hex is 4F7F FFFF, and conversion of HEX number 4F7F FFFF into decimal is 1333788671. 

Does anybody know what function should I use to get the answer I want?

Thanks.

Best Regards.

Although your description is fascinating, it does not let to conclude the problem you might have.

Show us a screenshot of the TRUNC instruction in question while being online. One picture is worth a thousand words.

Edit:

Or maybe it's not necessary.

The value you expect (4294967295) is outside of the DINT datatype range. You should do error checking while executing conversion ops.

Dear Jacek,

Yes, you are right. I tried to use a smaller real number for the conversion, then I was able to get the correct answer. Previously the real number exceed the range of DINT, so when it was converted to DINT, it was first converted to HEX, then converted to decimal again, which caused the result much different. 

BTW, just another question, for DINT value in Siemens PLC, is it only can be the signed integer? Is there any way to define a double integer as unsigned one?

Thanks. 

kevinlin_5

BTW, just another question, for DINT value in Siemens PLC, is it only can be the signed integer? Is there any way to define a double integer as unsigned one?

The main problem is the REAL rounding. The rounding instructions can only round to DINT.

Dear Jacek,

Thank you for the reply. So does it mean that in Siemens PLC Dint can only be signed double integer?

kevinlin_5

Dear Jacek,

Thank you for the reply. So does it mean that in Siemens PLC Dint can only be signed double integer?

Out of the box, yes. 

S7-1200/1500 has all needed instructions for both signed and unsigned arithmetics and conversions. Additionaly S7-1500 has 64-bit floating point and integers.

The number 4,294,967,295.3 cannot be accurately represented by a 32-bit REAL. It can be presented by a 64-bit REAL.

The nearest 32-bit REAL will give is -
(16#4F7F_FFFF)    4,294,967,040.000
(16#4F80_0000)    4,294,967,296.000

The DINT representation of the hexadecimal coded number of 16#4F7F_FFFF is 1,333,788,671. It does not represent the DINT conversion of the REAL number. These are only the convention of the encoding of a binary code.

Per definition the DINT is a 32-bit signed number. The 32-bit unsigned number is called a UDINT. DWORD is a 32-bit coded entity without number type specification. It is perfectly legal to use DWORD as a data type when looking to work with 32-bit unsigned numbers.

The standard instruction set for the older Siemens PLCs only make it possible to convert from REAL to DINT. Even if converting to unsigned 32-bit number was possible, this will still not be good enough as the REAL number is already a problem.

Newer Siemens PLCs have more specific data types - mentioned by {jacek d}. 64-bit processing is also included.

64-bit processing would make sense for handling 4,294,967,295.3. {Piotr.M} made a user defined library available to process 64-bit REAL with older PLCs.

Using Double Float (64-Bit Real) in S7-300/400 - Library with functions

What is the origin of the mentioned number of [sic] 4,294,967,295.3 ?