vanishing trend

[unsaint32]

With Controllogix (V20 & V21), when I trend, if I set the time to more than 3 minutes, the trend lines disappear. Is there a way to prevent this? I am attaching a screenshot of this issue.

 

[rdrast]

Is there a reason you have a period of 3ms? Just a guess, but you may just be overloading it. Try a longer sample period.

 

[Ken Roach]

Right-click on the Chart (onscreen or in the Project Organizer) to go to Chart Properties, then select the Sampling tab.

There's a section where you can select the trend's limit on the number of data points.

You can select an absolute number of samples, or a time period, or just choose to fill up your hard drive.

That's probably what's going on; four Pens at 3 ms sample rate would accumulate 1333 records per second, so it would exceed the default 200,000 record limit in 2 1/2 minutes.

 

[unsaint32]

Right-click on the Chart (onscreen or in the Project Organizer) to go to Chart Properties, then select the Sampling tab.

There's a section where you can select the trend's limit on the number of data points.

You can select an absolute number of samples, or a time period, or just choose to fill up your hard drive.

That's probably what's going on; four Pens at 3 ms sample rate would accumulate 1333 records per second, so it would exceed the default 200,000 record limit in 2 1/2 minutes.

I selected no limit but the vanishing trend issue is still happening as you can see in the picture I am attaching.

 

[Gil47]

At 10 ms sample time, you appear to get a six minute trend,
which is probably still faster than your scan time,
give 100 ms a try, its still fairly fast out in real world time

 

[Ken Roach]

Instead of attempting to poll individual data points extremely fast, RSLogix 5000 and the ControlLogix controller use some controller memory to store the data fast and upload it to the TrendX object in big chunks.

It's possible that you're running into a limitation of this function that isn't obvious in the configuration screens.

 

[Scott_Godin]

Has there been a reason identified as to why this happens. I experience it a lot also.

 

[dirtyc]

On a related note, if your trend sample period is faster than your task scan time, is the trend data still valid? Or are they even related at all? I have a 15mSec scan time for my task (the entire processor logic is in the single continuous task) and I have trend sample period at 1mSec. The trend *seems* to be showing tag data changing at a rate even faster than the scan time, but I wanted to see if the data was actually valid.