Hi everyone, case packer, 1769-L33ER with a PanelView Plus 7 on the same controller. Operators say the screens have gone sluggish, button presses take a second to show, and only when the machine is running at full rate. No faults anywhere.
T_Fast, the 5ms periodic task that does the flight tracking, has Overlap Count climbing on the Monitor tab, about 3 a second at full rate and zero when idle. We added a second PowerFlex 525 and tracking for a second lane about a month ago, which is roughly when it started. I doubled the RPI on the two drives from 20ms to 40ms and restarted the PanelView, no change.
Is the built in Ethernet port on the L33ER the bottleneck with the extra drive, or any thoughts?
Not the port. Overlap count going up means the 5ms trigger is arriving before the previous pass of T_Fast is done. What's the task's watchdog, what priority, and what's in it, ladder or ST? And which task does the second lane's tracking run in, the same 5ms one?
Watchdog 500ms, priority 5, ladder, about 30 rungs of flight tracking per lane, both lanes in T_Fast. So doubling the lanes doubled what's in a 5ms task. Max 6.8ms is over 5. I see it now, I think.
Periodic tasks always win over the continuous task, and comms live off the continuous task's time slice. Starve that and the PanelView starves.
Two things stacked. T_Fast running up to 6.8ms in a 5ms period runs past the next trigger whenever a scan goes long, overlap, and it takes priority every time. The continuous task gets what's left, and the System Overhead Time Slice that serves the PanelView comes out of that. That's your sluggish buttons, and no fault because nothing fails.
Fit the period to the work, not the other way round. Max 6.8ms, so 10ms with margin, and check the tracking still resolves at 10ms, it should. Or split lane 2 into its own 5ms task offset from the first, if 5ms is really needed. A CODESYS or TIA cyclic task does the same. Writeup on period versus execution time: https://plctr.com/understanding-plc-scan-time-and-cycle-time/