Ignition 8.1.10 Vision HMI Performance/PLC Overloading

Ignition
,
1

Good Day,

I have read a few posts now regarding an issue we seem to be having with our Ignition Vision set-up.

Essentially, I believe the issue is we are overloading our PLC and we are not getting the response times we are hoping for on our HMI application. We are using a ControlLogix L83E V33 and this is the current statistic page for the device:

image
image961×884 52 KB

CIP Connection Size set to 2000, Concurrent requests set to 4

As you can see we are receiving a response time (I am assuming this is the actual sampling interval) of roughly 3.5seconds on the 1000ms sampled data. Right now I am trying to optimize the HMI application since we would love to use ignition for more HMI applications, but for push buttons a status return of anything greater than 1-1.5 second would not work for our indicators.

What I am wondering is apart from changing the CIP size or the concurrent request count, is there a guide of sorts that shows how to optimize tag usage on the gateway in order to improve sampling intervals? I searched the forms for awhile and didn’t see anything.

It is a point to point connection from the PLC to the Ignition Gateway computer so this is not going through the cloud nor is there any other major traffic on the network.

I should also mention that at the start, we brought over most tags from the PLC and created the UDTs in our ignition gateway like the ones we have in the PLC. I am now seeing that this is a poor decision as we ended up with thousands of tags that are not being used in our HMI application that we use in the PLC, so right now I am deleting the unused tags. I am seeing sampling improvements by doing this, but the 1-1.5 sec actual sampling interval goal for the tags sampled at 1sec seems pretty far away.

2

It’s probably not a good idea to pull ALL tags into your HMI project as this just adds to the overhead. However, the issue isn’t really on the gateway side per say, but rather on the PLC side of things. We have systems with much higher tag counts, and ran into issues with the PLC CPU being overloaded, as well as showing in Ignition as it being overloaded.

What we ended up doing was adjusting the period rate on the PLC from 20ms (I think this is default) to something like 100ms, and this gives the CPU more time to process communications traffic (5x roughly). That might be enough for you to reduce the load seen by Ignition to about 50%. Based on the Mean Response Time (99.1ms), I’m guessing you have a periodic rate set to 100ms. To get it under 100%, you will have to bump it to 300ms to give you some margin. You might also have to adjust the Request Packet Interval setting as well, but also look into the System Overhead Time Slice, as increasing this will allow more throughput.

Some of this is discussed in this thread.

1 Like