wildswing
Member
Hey fellas,
This one's more of a curiosity than "gotta fix it now" kinda thing. I've just discovered some PID loops in one of our PLC5s that are triggered by pulse bit (ON for 100 and OFF for 100), not a single scan timer.dn XIC. The PID update times are set to 100 msec. The control is a PD file (not Integer). The PLC scan time varies from 20 - 50 msec.
I've read all the help files and a bunch of posts here, so I think I'm on the right track here. Just asking for confirmation if my thinking is right or not.
To me that's programmed wrong. With the control being a PD file (not N), the PID instruction will execute 2 to 5 times when the trigger pulse is ON, then lie quite while it's trigger is OFF. With the update time set to 100 msec, the Ki (no Kd) will act 2 - 5 times more often than expected, thus to achive acceptable tuning, it's value is most likely 2 - 5 times lower than it would be if the PID were being triggered properly.
To be correct, these PIDs should at least be triggered by a timer.dn xic that's true for only one scan. Better yet, they should be in a 100 msec STI file.
Looks to me that the original programmer thought the PIDs would only execute on the false to true rung transition, not every scan, or he thought the PD worked like an N file control with a .DN bit. This has most likely been this way since day one of the OEM setup many moons ago so I'm not in a rush to fix it.
This one's more of a curiosity than "gotta fix it now" kinda thing. I've just discovered some PID loops in one of our PLC5s that are triggered by pulse bit (ON for 100 and OFF for 100), not a single scan timer.dn XIC. The PID update times are set to 100 msec. The control is a PD file (not Integer). The PLC scan time varies from 20 - 50 msec.
I've read all the help files and a bunch of posts here, so I think I'm on the right track here. Just asking for confirmation if my thinking is right or not.
To me that's programmed wrong. With the control being a PD file (not N), the PID instruction will execute 2 to 5 times when the trigger pulse is ON, then lie quite while it's trigger is OFF. With the update time set to 100 msec, the Ki (no Kd) will act 2 - 5 times more often than expected, thus to achive acceptable tuning, it's value is most likely 2 - 5 times lower than it would be if the PID were being triggered properly.
To be correct, these PIDs should at least be triggered by a timer.dn xic that's true for only one scan. Better yet, they should be in a 100 msec STI file.
Looks to me that the original programmer thought the PIDs would only execute on the false to true rung transition, not every scan, or he thought the PD worked like an N file control with a .DN bit. This has most likely been this way since day one of the OEM setup many moons ago so I'm not in a rush to fix it.