Hi,

first of all, Mr.V is correct.

Please compare Real with Real, in newer versions of TIA (like V18) it would not be possible to compile this code due to this different data type.

But this would not work with Real neither, because the binary interpretation of 16#0000_0001 is way to small for data type Real.

The binary representation of 16#0000_0001 is "0000_0000_0000_0000_0000_0000_0000_0001", which is way outside of the range of Real can handle.
Furthermore, Real is/will be rounded.

So, first - use seperate/own tags for datatype "Real" and not markers.
second - use values which can be represented in Real.

Following FAQ may be helpful:

How accurately can I calculate with REAL numbers that are used in extensive formulas?

Why is a difference detected in the "Compare real numbers for equality" operation although the floating-point numbers are equal in the status?

How do you check real numbers for parity?

*****EDIT******

The smallest *positive* value represented in Real is "+1.175495e-38" -> 16#0080_0005

Up from this value the COMP should work

REAL

-> +1.175495e-38 to +3.402823e+38

regards

sancruz