Since adding some new tags to my tree it seems my tag response time has worsened. Simply changing a boolean value through a push button tag binding takes more than a few seconds.
Screenshot shows stats for connection to a AB L33ER PLC over Ethernet.
How do I resolve?
What firmware version is the PLC running? Periodic or continuous task? Can you or have you increased the overhead timeslice setting in the PLC?
v32. Continuous. Ive never changed it. I think I can change.
Try changing it.
On the Ignition side you can maybe tweak the Concurrency and CIP Connection Size settings on the driver but you’d probably need to adjust the overhead timeslice before doing this. If a connection size other than the default is going to work it will likely be <= 2000. I’d try 1000 to start and see if that helps at all.
OK. What does the overload and monitored item count mean in real terms? I don’t believe this was an issue until I added some new tags to my tree.
I changed the scan rate in Ignition. This has reduced the overload but doesn’t seem to have sped up the response time in my client?
Look at the mean response time versus the number of requests. Multiply. Divide by # of concurrent requests. The resulting # of milliseconds is the fastest pace your connection supports.
You might gain a bit by increasing concurrency, but if the bottleneck is in the PLC (likely), the response times will go up and you won’t really be going faster.
PLC bottlenecks can be mitigated by changing a continuous task to a periodic. In your case, I’d recommend a 10ms or 20ms period. Logix processors also have a setting for communications time slice. Increase that.
It didn’t change anything because when you were requesting 1000ms polling you were only getting ~3200ms anyway.
Overload is how far over the requested/ideal sampling you are. Monitored Item count is basically just subscribed tag count for that device.
You might also want to review this topic if you are reading much from AOIs:
Also, avoid reading alias tags. (I recommend avoiding aliases, period.)
