SLOW, VERY SLOW RSView32

We have an application with 3 PLC5 with about 4000 tags in each of the 9 PC with RSView32.

They com over Ethernet to the 3 PLC (PLC5/40C + Enet module) and exchange messages between each other.

Whe we have all 9 PC online, they pass throuhg a RSLinx Gateway equiped PC to have data from the 3 PLC.

When more than 5 PC are online... the comunication is VERY slow... Tag need more then 10 seconds to update.

Is there a way to make this better... is it normal now like it is?.

Is there a formula to figure what speed we should be getting out of such loaded system...

Perhaps they should have stuck to rebadged Citect Pierre!

That sounds like a huge amount of traffic to expect the controllers to serve up. Are the tags in scan classes ? What kind of update rates are they configured for ?

Check the inter-controller messaging to be sure it's not written to continuously trigger the messages.

Is the RSLinx Gateway actually performing as a Gateway (linking TCP connections to ControlNet or another TCP connection) or is it acting as a Remote OPC Server ?

Any time the number of display terminals exceeds three or four, I try to use Remote OPC or a Client/Server system.

Ot.......

OT

Hi Ken...you are up early this morning!!!! Happy New Year!!!(a little late)

end OT

David

David_Emmerich said:

OT

Hi Ken...you are up early this morning!!!! Happy New Year!!!(a little late)

end OT

David

Click to expand...

I would guess he's actually up late

I am having some of the same issues. With 5 SLC's 1 RMC and a VersaView. It helped to put a SVC in logic when you need to update tags.

Another thing done was instead of having the message control resend the message we changed the msg trigger to a time base S:3 or S:5. I have one message that has to do with motion so that message is transmitting continuously but data we slow down the trigger and it made the whole thing faster.

With the SLC, with a single client, the situation is a little bit simpler and you took the right approaches to improving performance. The SLC's main CPU handles messaging during housekeeping only, whereas the PLC-5E and 1785-ENET sidecar run a dedicated subsystem.

In Pierre's application, the question of whether RSLinx Gateway is literally running in Gateway mode or running as a Remote OPC Server is the big question. Until that is determined, everything else is guesswork.

It is a misconception that talking to a device over ethernet allows the device to handle unlimited communications. There is actually a inverse relation to the speed of communications vs. the number of connections you have to a device. The more connections that you have to the device the less time it has to servcice each connection.

A good example is talking in a chat room. If you you are talking to one person you can hold a very quick converstation but the more people you add to the conversation the longer it takes to respond to each one because you have to devide you time between all of them.

So back to the PLC's. A PLC processes in a cycle x msec to recieve requests, x msec to process ladder, and xmsec to process responses. This is a very simplistic explanation of a much mor complex process. So, if each RSLink is connected and requesting the same amount of data then the PLC's have to process the same requests multiple times and since they have to process requests for more than one connection they cannot process as much.

We have seen in our testing that regardless of the server you are using there is usually a ver specific line at which communications speed is impacted. It is very likely that 5 connections is the max that can be handled. We typically recommend that customers make one connection and then connect multiple clients to the same OPC Server. The OPC packet can be a larger bandwidth and almost all OPC Cleints defautl to a report by exception mode which means they are updated when the data has changed in the server form one poll cycle to another.

Ken Roach said:

...the question of whether RSLinx Gateway is literally running in Gateway mode or running as a Remote OPC...

Click to expand...

Thanks guys for your comments.

The application is running through The RSL-G acting as an OPC Server.

Furthermore, we have on this Server, the Gateway/OPC program AND also TrendX which log about 500 Tags to the second.

We have a manage switch
The network is at 100MB

When all PCs(9) are ON-LINE, the most traffic we get out of a pipe is 17%.

Dumb but simple observation is...

When I look into a local RSLinx and I click the Gateway link, It takes 9 second to SEE the status of the 3 connected PLCs.

When I click the direct Ethernet driver to see the same thing... it takes only 4 seconds.

So the direct link IS way faster... unless we shutdown the OPC server and then the RSViews fall unto the direct driver link... then its very slow... in the range of 10 seconds.

When I monitor some Encoders connected to a VHSC modules on the main rack, in RSView Tag monitors, I see there value change every second. It look very fast to me.

The problem or should I say challenges is that when we are running the RSView32 application, the screens just dont show up when they are commanded to do so... before 1 second, and then the values are shown after another 3 to 5 seconds.

Also, the logs are on the Server, so alarms are somewhat slow to appear.

For the MSG, they are continuous. 4 WORDS READ from the 3 PLCs to there 2 other conterparts. And this is VERY slow to show on the RSView but VERY fast between the PLCs.

I have started to cleanup the SCADA applicatio.

First I installed 4 new Quad Core PCs in the main control room with XP Pro and 4 Gig of RAM with FAST HD.

They all have Dual Screens. So 4 PC + 8 Screens

I have leaned up the TrendX setup. Removing ALL but 1 Trend in each PCs AND leaving all the other trend for a PC dedicated to the Managers.

Also I have remove a lot of unused Tags.
------------------------------------------------
My main challenge is finding what is slowing us down.

I see the com between the PC and the PLC is FAST when ONLY one PC is live.

Then , passing through the OPC server must be (for the PLC) no more traffic if 1 PC or 100 PC.

Now when I have 10 PC through the OPC Server... why would there screen appear slow to update?

Again... the ethernet is only up to 17%...