Wonderware Screen slowdown

Hi...
Our system using Controllogix Redundant SYstem with 1757-SRM, 1756-CNB on local chassis (Both primary and secondary with same slot for each module) and remote chassis using 1756-CNB, I/O and 1756-ENBT ethernet module. Distance from Controllogix to HMI workstation is 5 meters. The connection from controllogix chassis to computer via 8 ports HUB and three HMI workstation are connected to HUB.
The HMI using Wonderware VIew version 5.00 (three workstation with windows 2000 professional OS) to access to same controllogix, with diffrence address.
We have prom with slowdown on network and cause Wonderware opening screen is slow too. If we want to change the parameter on Controllogix, say for PID setpoint and it takes about 3-5 minutes to change.
Could anyone help us to solve that problem?
Thank you for help.
James

... Hrm.

First, use a switch, not a hub. What is your network speed? 10MBps? 100?

Second, on the topic configuration, select 'Advise only active items'.

Do you really mean version 5.00? isn't that, well, very very old?

Wonderware screen slowdown

I use latest version. The speed is 100 MBPS (5 meters distance between Controllogix and HMI).
Any idea?

What I/O server are you using?
How many tags are in your project?

Something is seriously wrong with what you've described...

Is performance better if only one HMI is running and the other two are unplugged?

is about 1700 tags and use OPC server. If I use only one HMI the network speed will come up. The problem is if I use two or three workstation is same time.
James

Am I understanding correctly? You have more than 1 HMI polling the ControlLogix with 1700 tags? What is your scan rates? Are you using I/O servers on all machines? Instead of one machine acting as an I/O Server with the other workstations pointed to it? I am not sure what Wonderware calls it exactly, but I think they use terminal services to allow one machine to act as an I/O server with the other workstations attaching to it for tag data. You are seriously overloading the CL communication ports if you are setting up seperate HMI's to access it directly. This will kill network performance(not to even mention using a hub instead of a switch).


David

Sounds like a network congestion issue. I have 3 View applications with over 3000 tags, connected to 3 PLCs and there is no delay to speak of. Try these things to free up the congestion.

Make sure your remote chassis is not on the same network that you HMI is on. You should be able to use a crossover cable to connect the local rack directly to the remote I/O.

Use only one OPC server to connect to the PLC. Also make this a direct connection to the PLC. So in your (master) computer, the one running the OPC server, install 2 NICs, one to talk to the PLC and one to network with the other View applications.

You have two options for getting the data to the View applications. You can have all the View applications talk to the OPC server. In your Access Names you set Node Name to the name of the computer that is running the OPC server. The other option is to have the (master) View application obtain the data from the OPC server and have the other two View applications obtain their data from the (master) View application. In this case you set the Node Name to the (master) View application's name, Application Name to VIEW, and Topic Name to TAGNAME. The first option is the easiest to work with.

You can set-up Topics in you OPC server for different update times, this way your OPC server isn't trying to get all 1700 tags at once.

Just for info - When View applications talk to each other they use SuiteLink protocal. Terminal Services is a Thin Client architecture in which the application is run on a Server and Clients access the application running on the Server. Since the Clients are not running the application they don't need the latest hardware, they are thin.

Thanks Tark, I was unsure about that. I had not done anything like that for a while now and could not remember. I will recommit that to memory. Your explanation started to jog my memory. The local rep walked me through it (Lee Seneker, excellent guy btw) and I have not done anything with it since.


David

W/Ware

Tark is right.

We had same 'issue' happen with 8 W/Ware PC's trying to get updated from 40 RTU's. All com's were over fiber, so we made 1 PC the master to the RTU's and had the other 7 PC's only communicate with the master PC.

Works great now.

I too use multiple I/O servers with Wonderware....Each of my control system PLC's has a single HMI node that provides the I/O server to the rest of the system. I use this approach so that a network failure does not totally shut down the system as it would with a centralized tag server.

Use a switch..........not a hub.

How are you handling the failover in case of a processor failure?
What I/O server are you using? RSLinx Pro? DASTcp?

Dave

Our application is need to use three computer with the same screen and level. After we count the tags, we have 3500 tags that need to read from controllogix. We use RSLinx Professional ver 2.42 to create OPC/DDE Topics. Each computer has its own RSLinx. We set polled rate for 500 ms and maximum packet limit for PC#1 is 10000, PC#2 is 20000 and PC#3 is 100000. In RSLinx communication event, we found error message:
"ControLogix connection X got disconnected" and "ControlLogix optimized packet being reinitialized"
Othet message is "Item does not exist in processor..." This error message always appear on error log.
In the topics monitor, we found message panding-respond.
In COntrolLogix, the time slice setting is 12%. No arrays or user define tags (UDT).
We try to make different scedule and we plant to increase time slice, redefine tags and using short tags name and using arrays.
Could this solution will improve network performance? Any idea?
Thank you.
James

You only need RSLinx on one computer. Using 3 RSLinx applications is (more than likley) flooding your network. Use only one RSLinx application and have all the Wonderware applications talk to it.