How many tags can you realistically get from one PLC on s7 plus? I have a large project grabbing around 23k tags from a 1518 and I'm running into situations where I don't see updates for 10-12 seconds in the PLC after setting a value in OA. What is the best practice for grabbing this many tags from a single source?
I thought I remembered a way to offset polling groups from each other when they have the same cycle time, but it looks like I'm limited to the second range with that because the smallest sync time is 1s. That means all polling groups on a 1 second cycle would line up anyway. I also believed this to be too many tags for subscriptions, but I haven't tested that yet.
S7Plus limits
Search
Re: S7Plus limits
Can you please add the information which WinCC OA version you are using.
If you have around 23.000 tags and you poll them every second you will possibly get 23.000 VC/sec if all the items are not smoothed.
Are you sure that you want to process 23.000 VC/sec in your system all the time?
Can you observe a high CPU load for the Event and or S7 driver?
Is there a need to poll all the items every second or can you user longer poll cycles for some of the items?
Are the command processed
Best Regards
Leopold Knipp
Senior Support Specialist
If you have around 23.000 tags and you poll them every second you will possibly get 23.000 VC/sec if all the items are not smoothed.
Are you sure that you want to process 23.000 VC/sec in your system all the time?
Can you observe a high CPU load for the Event and or S7 driver?
Is there a need to poll all the items every second or can you user longer poll cycles for some of the items?
Are the command processed
Best Regards
Leopold Knipp
Senior Support Specialist
-
MattPaulissen
- Posts: 26
- Joined: Fri Feb 19, 2016 9:27 pm
Re: S7Plus limits
It is version 3.16. I'm in the process of figuring out exactly how many tags really need this update rate, but mostly I'm wanting to know what the best practice is for grabbing large numbers of tags from a single source? Is there anything I should try to do inside OA to optimize this?
My understanding was that smoothing only helps for the Event manager, but the S7 driver would still be doing a lot of work even it doesn't forward the messages up. How many VC/sec can the S7 Driver realistically manage?
My understanding was that smoothing only helps for the Event manager, but the S7 driver would still be doing a lot of work even it doesn't forward the messages up. How many VC/sec can the S7 Driver realistically manage?
Re: S7Plus limits
If you are using a _smooth config smootihng is processed in the driver and only data after smoothing is send to the Event Manager.
When you observe performance issues you have to look in detail where the bottleneck is. In a communication with a PLC possibly also the PLC can be the bottleneck.
In your test where you have observed that a command is processed delayed you can possibly activate the debug flag "-rcv 2 -dbg 2 " for the S7+ driver to see when data is received from the Event Manager and when the driver sends it to the PLC.
When there is a delay between the log message and the signal in the PLC it looks like the PLC causes the delay.
Information for the possible number of VC/sec in a system is a number which cannot be defined for every system as there are too many conditions which have an influence:
-- Hardware configuration
-- Operating system
-- Project configuration (single, redundant)
-- Size of the project
-- Complexity of the project (running scripts, API managers, alert handling, ....)
Best Regards
Leopold Knipp
Senior Support Specialist
When you observe performance issues you have to look in detail where the bottleneck is. In a communication with a PLC possibly also the PLC can be the bottleneck.
In your test where you have observed that a command is processed delayed you can possibly activate the debug flag "-rcv 2 -dbg 2 " for the S7+ driver to see when data is received from the Event Manager and when the driver sends it to the PLC.
When there is a delay between the log message and the signal in the PLC it looks like the PLC causes the delay.
Information for the possible number of VC/sec in a system is a number which cannot be defined for every system as there are too many conditions which have an influence:
-- Hardware configuration
-- Operating system
-- Project configuration (single, redundant)
-- Size of the project
-- Complexity of the project (running scripts, API managers, alert handling, ....)
Best Regards
Leopold Knipp
Senior Support Specialist