RPI and network communications

Join Date
May 2011
Location
South Carolina
Posts
87
This is a very broad subject but I am really unsure how to ask it.
We have a fairly large network with several hundred IP addresses (600 +). We have mainly Micrologix and Compact Logix PLCs on our plant machinery but 4 machines each have 5 Siemens PLCs. A lot of the machines are exactly the same but with different IP addresses. For instance we have 14 waterjets the same, 24 hydraulic presses the same, etc. I have verified that there are no duplicate IP addresses and I have changed a lot of things from multicast to unicast. The problem is, if I connect machines to the network, we start having communication issues, meaning that machines stop responding and you have to disconnect them from the network and power down/power up to reset communications. It may run for a couple of hours before we have an issue or it may run for 3 days. I have been racking my brain trying to figure out why this is happening and I am having no luck. The only thing that comes to mind is that all of the RPIs that I have looked at on the various machines are set to the default of 20 ms. I am wondering if this could be the root cause. Is there a guide on what the RPIs should be set to based on what the machine function is? So as an example, a hydraulic press would have a maximum allowed RPI of 250 ms and a controller with waterjet robots would have a maximum allowed RPI of 50 ms. Any guidance would be greatly appreciated. My next step is to find someone with the experience and credentials that could come in and help me find the reason, and I am sure that would be quite expensive.
 
I wish I was nearby. This is question similar to "My car is making a clanking noise. What is wrong?" People will then start listing out every part in the car that can be making a clanking noise without doing any real diagnostics.


The first thing you need to do is map out the hierarchy of the network. This is a pretty huge network, and your problem is very likely to be congestion from all of the devices.
Do you just have all of these devices going to a large switch? Is this a 1GBPS network? Is it actually operating at that speed?

This is one of those instances where you may really need an evaluation by someone who is well-versed in networking to check into this. I'm not sure of anyone local, but I would be willing to assist with this issue if you aren't able to find someone closer.


And to be more specific, the first thing that needs to be done is to take these 600 IP's and work up a map to visualize the segmentation. It needs to be grouped by device -- > machine --> any switch / etc in machine --> primary switch (across plant floor) --> anything that needs to go to the outside world. That is just a ton of deterministic stuff for an improperly segmented network.
 
Last edited:
As Sparkie mentioned, first thing is document what you have. Do you have a network backbone, with managed switches, or is it just a bunch of "islands" connected together? Are there any V-Lans? Was this network designed or did it just develop over time?
 
If I were in your situation, first follow sparkie's advice and map out your network. Start by connecting small segments and use Wireshark to observe the traffic. The port statistics and logging functions of your switches (assuming managed switches) will be helpful too.

Shawn
 
Good mapping,managed switches, and vlans are going to be the key to getting this to work correctly. If machines 7 thru 15, 24, and 65 need to talk to each other, then they need to be separated by a vlan so that their traffic doesn't cause issues with the rest of the machines. If all the machines need to, at some point, deliver information back to a central location, then update time will become an issue. This kind of networking is chock full of pitfalls, and while hiring an expert sounds expensive, it is usually money well spent if it eliminates the downtime.

Bubba.
 
Impossible to diagnose with the information provided, but saying that "Connecting machines to the network, you start having communication issues" sounds to me like you might be creating loops in your network, without proper Spanning Tree setups.


Throw WireShark on your PC, and watch the traffic as you plug in the troublesome machines. I suspect you will start seeing lots of duplicate packets; if not ports shutting down.
 

Similar Topics

Hi there I am new with this thing and i don't know how to connect Allen Bradley Micrologix 1200 PLC to Raspberry Pi using USB to RS485 in NodeRED...
Replies
3
Views
2,233
Hello all, I am using a temposonic R series linear transducer for positioning of a lower ram on a hydraulic powdered metal press. The transducer...
Replies
7
Views
1,911
Hello Ladies and Gents, I have some PLC to PLC communications using Produce and Consume. The data is not time sensitive, and I really only need...
Replies
16
Views
4,614
I am not very experienced in Rockwell PLCs and have run into a small problem, so I could do with a little bit of help. We have a redundant...
Replies
3
Views
2,470
Dear All , I need help to resolve the issue of the RPI that has been not changing on my configured L35E processor ? it was running since but now...
Replies
2
Views
1,440
Back
Top Bottom