(ST) Embedded function calls

Is this possible? I'll use this scenario as an example.

I have a function block that represents the behavior of an I/O card, say a 2 channel thermocouple card, and it has read, write, and configure behavior. so I have a separate read function block and a write function block, etc. Almost any card I use Analog Input cards, outputs cards, etc will use the read, write, configure fb's. So my question is, I want to 'read a value from the thermocouple 2 card',

I would call
<code>
thermo2.read(word);
</code>

is this possible?

I don't think the IEC61131 has this, but Beckhoff TwinCAT has a feature called "actions". Each action is a subroutine associated with a POU or an FB and shares the local variables.

So if you have an FB 'Thermo2', you can put all your code that handles reading data inot an action called 'Read' and later call 'Thermo2.Read' elsewhere in the program.

PS Compare this to 'methods' in Visual Basic.

LadderLogic said:

PS Compare this to 'methods' in Visual Basic.

Click to expand...

But VB is a object oriented language and ST is not.
In VB 'thermo2' would be an instance of an object and 'read()' would be a function or 'method' inside of that instance.

yes you can, but syntax is different.
or use value:=thermoread(2)
or use thermo(read,2)

The IEC standards are vague and loose, so every platform's implementation of ST is different in capability. However, it is pretty standard to allow function block instances to be part of a structure, so you could have a structure named 'thermo2' and that structure could contain a function block instance named 'read' and you would call it exactly as you have it in your first post.

Also you can look at codesys v3 and twincat 3 controllers to get more relaxed ST enviroment.

Automation Studio 3.X and above also provides a great ST implementation that I much prefer to Codesys 3 (I've never tried TwinCat 3, but hear good things).

Rockwell's Studio 5000 is the worst I've ever encountered, but there are probably even worse out there. I'd be surprised if you could do what you are trying to do with a Rockwell system.

actions are part of the SFC language but they can be implemented in ST.
Shooter and Nettogisen sum things up pretty well.

capinwinky said:

The IEC standards are vague and loose, so every platform's implementation of ST is different in capability. However, it is pretty standard to allow function block instances to be part of a structure, so you could have a structure named 'thermo2' and that structure could contain a function block instance named 'read' and you would call it exactly as you have it in your first post.

Click to expand...

errr.
IEC 61161-3 is not a standard, it is a specification.
The IEC specification is pretty specific on how things must work. It isn't specific on how the compiler writer implements the specification.

Allow FB instances inside a structure is something that Siemens does because all the data is in a data block some place. The IEC specification doesn't mention data blocks.

There are a couple extensions that make sense. It is easy to implement a UNION. It wouldn't be too hard to allow one to declare an array of FB instances. True object oriented features would be a lot of work to implement and probably wouldn't get used by most programmers.

I am still working on my ST compiler off and on. I am actually working on function block instances. I am allocating the static memory for each instance within the static data are of the parent function block or program. Functions cannot have function blocks as they have no static data.

actions are part of the SFC language but they can be implemented in ST.
Shooter and Nettogisen sum things up pretty well.

capinwinky said:

The IEC standards are vague and loose, so every platform's implementation of ST is different in capability. However, it is pretty standard to allow function block instances to be part of a structure, so you could have a structure named 'thermo2' and that structure could contain a function block instance named 'read' and you would call it exactly as you have it in your first post.

Click to expand...

errr.
IEC 61161-3 is not a standard, it is a specification.
The IEC specification is pretty specific on how things must work. It isn't specific on how the compiler writer implements the specification.

Allow FB instances inside a structure is something that Siemens does because all the data is in a data block some place. The IEC specification doesn't mention data blocks.

There are a couple extensions that make sense. It is easy to implement a UNION. It wouldn't be too hard to allow one to declare an array of FB instances. True object oriented features would be a lot of work to implement and probably wouldn't get used by most programmers.

I am still working on my ST compiler off and on. I am actually working on function block instances. I am allocating the static memory for each instance within the static data are of the parent function block or program. Functions cannot have function blocks as they have no static data.