Problem With For Loop and Trig bug

Engineering 🇬🇧
omeo
2018-02-01
2023-08-17

  • omeo - 2018-02-01

    Hi
    I recently modified a project, but I found that when I changed the trig array value and the for loop did not change, it caused an error in the data:
    exp:

    PROGRAM PLC_PRG
VAR
   ft_Ctrl: ARRAY [0..2] OF F_TRIG;
   rt_Ctrl: R_TRIG;
   
   index: INT;
   iState: INT;
   xCmd: BOOL;
   xAutoMode: BOOL;
   xDone: BOOL;
END_VAR
--------------------------------------------
FOR index :=0 TO 3 DO
   IF ft_Ctrl[index].Q THEN
      iState:=90;
   END_IF
END_FOR
rt_ctrl(CLK:=xCmd);
xAutoMode S= rt_Ctrl.Q;
CASE iState OF
   0:
      IF xAutoMode THEN
         iState:=10;
      END_IF
   10:
      iState:=90;
      xAutoMode:=FALSE;
      xDone:=TRUE;
   90:
      iState:=0;
      xAutoMode:=FALSE;
END_CASE

    Notice that in the definition I give f_trig [0..2] and call f_trig [3] in the for loop without error。

    In this program, I hope that when xCmd is the rising edge, iState followed by 10,90 and then back to 0.
    But in fact, I use a single loop to monitor, the first loop iState set to 10. in the second loop, ft_Ctrl [3]. Q actually call the value of rt_Ctrl.Q and iState set to 90,and then back to 0.

    This is very hard to find, because the actual project is large, I forgot to change the for statement after changing the definition of the variable, resulting in the emergence of this problem. I tested this on both somachine and twincat. So if this is the case, can the alarm be exceeded?

    sorry for google translate..and thanks for attention

     

  • omeo - 2018-02-01

    Make it simpler:

    PROGRAM PLC_PRG
VAR
   ft_Ctrl: ARRAY[0..1] OF F_TRIG;
   rt_Ctrl: R_TRIG;
   
   index: INT;
   xCmd: BOOL;
   xDone:BOOL;
END_VAR
--------------------------
FOR index := 0 TO 3 DO
   IF ft_Ctrl[index].Q THEN
      xDone:=TRUE;
   END_IF
END_FOR
rt_ctrl(CLK:=xCmd);

    At this program,when xCmd R_TRIG,xDone will be true.
    I think it may be coherent after compilation so the address is continuously pointed, but it is somewhat counterproductive..

     

  • eschwellinger

    eschwellinger - 2018-02-01

    Hi,
    So your question is basically :
    How to find in huge projects these kind of application errors?
    You need to add the implcit check function: check bounds.
    Then set brakepoints for upper and lower - then you will find the code in your application where you write below the array...

    Note: later you need to remove this implicit functions again... ( just for the application development they should be in the application due Performance issues)

    BR
    Edwin

    IMG: Setbrakepoints.jpg

    IMG: checkarrayboungs.jpg

    IMG: Pouforimplicitchecks.jpg

     

  • omeo - 2018-02-01

    Yes, this is perfect
    Thank you very much

     

  • eschwellinger

    eschwellinger - 2018-02-01

    Hi,

    Zitat:
    the future version can join this check function?

    every CODESYS Version has this checkfunctions build in!
    BR
    Edwin

     

  • omeo - 2018-02-01

    I think for some time, if you write the program should be able to avoid this problem:

    PROGRAM PLC_PRG
VAR
   ft_Ctrl: ARRAY[0..iMax_ft_Number] OF F_TRIG;
   rt_Ctrl: R_TRIG;
   index: INT;
   xCmd: BOOL;
   xDone:BOOL;
END_VAR
VAR CONSTANT
   iMax_ft_Number:INT:=2;
END_VAR
---------------
FOR index := 0 TO 5 DO
   IF ft_Ctrl[LIMIT(0,index,iMax_ft_Number)].Q THEN
      xDone:=TRUE;
   END_IF
END_FOR
rt_ctrl(CLK:=xCmd);

    Uh, but writing like this adds to the complexity of the program. . I think I need to be cautious when I change the definition of an array.

     

  • Anonymous - 2018-02-01

    Originally created by: rickj

    Or you could just do this ...

    FOR index := 0 TO iMax_ft_Number DO
   IF ft_Ctrl[index].Q THEN
      xDone:=TRUE;
   END_IF
END_FOR
     

  • josepmariarams - 2018-02-04

    Hi.

    For me is not a bad solution to wrapp an array with an FB with get, add, set ... Methods

    Enviat des del meu Aquaris M5.5 usant Tapatalk

     

  • Anonymous - 2018-02-05

    Originally created by: rickj

    Josep M. Rams hat geschrieben:
    Hi.
    For me is not a bad solution to wrapp an array with an FB with get, add, set ... Methods
    Enviat des del meu Aquaris M5.5 usant Tapatalk

    Agree Joseph.

     

Log in to post a comment.