Codesys Retentive Timer? (RTO/RTON)

drbitboy said:

Thanks!


So is there any difference between an argument in the VAR_INPUT block and an argument in the VAR_OUPUT block?

Click to expand...

Yes. In the latest version at least, you cannot write to outputs from outside the Function Block


They also appear on the right hand side of the function block for the graphical languages.


In ST, your call must use the '=>' operator to assign the outputs to variables, so:
instance_name(in:= myin, out => myout);

drbitboy said:

Working the other side i.e. ET, just four statements, plus one for TT. Untested.

(*    rtovar(IN:=RTOin, PT:=RTOpt, RST:= RTOrst      // Inputs  *)

(*          , Q=>RTOq, TT=>RTOtt);                // Outputs *)

(****************************************************************)

​

Click to expand...

That is why there is a Var_In_Out, local variables are not guaranteed to hold their last status on return to the block (although it appears they do )and if there is a variable that needs to be exposed externally from/to the function it needs to be passed both ways i.e. Var_In_Out.

drbitboy said:

Thanks!


So is there any difference between an argument in the VAR_INPUT block and an argument in the VAR_OUPUT block?

Click to expand...

AustralIan said:

Yes. In the latest version at least, you cannot write to outputs from outside the Function Block


They also appear on the right hand side of the function block for the graphical languages.


In ST, your call must use the '=>' operator to assign the outputs to variables, so:
instance_name(in:= myin, out => myout);

Click to expand...

Yes, I just tried writing to VAR_Outs this morning in 3.5.18 and it said "No! DUMB@$$! You can't do that"



As for the => operator, I guess I never noticed as I don't assign outputs in the FB call. I just reference them as members of the instance some where else.

parky said:

That is why there is a Var_In_Out, local variables are not guaranteed to hold their last status on return to the block (although it appears they do )and if there is a variable that needs to be exposed externally from/to the function it needs to be passed both ways i.e. Var_In_Out.

Click to expand...

Where in the documentation does it say that Var_Inputs won't be guaranteed to keep there value? if that is so... I have some code I need to rethink.

I have been using FB sometimes as a Data Only class instead of Structs because in some cases I might want to add Methods or Properties later. SO I just use Inputs to hold all my data.


VAR_In_Out is just a reference so then your real data has to be stored some where else, which gets messy.

In Rockwell Land you can't even reference a IN_OUT using the AOI instance, which makes me want to strangle the Studio 5000 Product Manager, though she did seemed like a really nice person when I talked to her the one time, so maybe just strange her Boss instead LOLOLOLOlz

Edit:

Can you do that in Codesys? I need to find out. Usually I want to do it Rockwell as a hack because you can't pass structs/AOI Types in by value with an Input.

It's not variable inputs or outputs, it's the local variables, again I have never found an issue but I read somewhere in the IEC notes on this that the data is only valid within the block & cannot be relied on in returning to the block as it is supposed to be initialised on entering the block.
As for the IN_OUT this is a standard part of the IEC it does seem that this supposed transportable code idea is not that transportable from one system to another.
It does appear that there are many discrepancies in the interpretation of what this universal code is supposed to support.

Here's my attempt, lets you time past 49 days. You can add some line breaks and tabs to make the ET look a bit more readable.

Late to the Game

I'm late to the game here. I usually check the Oscat Basic library (source code available here: http://www.oscat.de/component/jdownloads/summary/2-oscat-basic/6-oscat-basic-333.html )

when looking for different implementations.

There is a function block in there called ONTIME. It is similar to RTO. The version there only measures to the second resolution but it is easy to modify it to read milliseconds.
There is a dependency in the ONTIME function block to a function called T_PLC_MS(). But this can be replaced with TIME_TO_DWORD(TIME()).