DH+ Network Mapping

[eight_bools]

This has been on my mind for a while now, wondering if anyone here has any best practices or insights. I have been working with a plant that has a pretty large DH+ network that causes them a lot of issues. It looks like the network was designed initially as a small packaging line, all communicating nicely. Over time it was expanded, I don’t have an exact node count in front of me but I think it’s close to 30.

To me it doesn’t make sense that it is as large as it is, and that it’s not separated into smaller sub networks of equipment that needs communications. It covers 3 production lines and some common support equipment. I also saw that there are some circular connections (both a & b channels connected on a DHRIO bridge)

Before I can help them put a plan together to move forward I would like to map the physical layout of the network to get an idea of what we are looking at. I have some software called NetDecoder to get an idea of which machines are communicating with each other. That should help me get a list of all the msg instructions that are being used. I am at a loss though of a good way to trace the actual cables.

The only thing I can think of is to unplug all the connections while they are down, and use a cable tracer on each of the connections. This makes me nervous, I would hate to miss a controller and fry the comm port. It seems less than ideal, it is a large facility and it seems like an exercise like this could be quite time consuming.

Anyone ever done something like this or have any suggestions?

 

[geniusintraining]

I had one customer that had maybe 20ish and I thought that was to many but of course it depends on the amount of traffic (data) you are passing on, the issue with DH+ is that it takes a puck and passes it from node to node and all the data is in that single puck, in your case I would try and break it into 3 and have one for each line, if they have the budget I would also add a bridge and get some of the traffic to Ethernet

Its very industrial/robust protocol so if your having issues make sure you have a terminal resistor at the end nodes and if your still having issues then odds are you have to many nodes and too much data

Reading material https://literature.rockwellautomation.com/idc/groups/literature/documents/um/1770-um022_-en-p.pdf

My customer had a fair amount of data but every couple hours they would load a new recipe and that had several large ramps of temp, vacuum and pressure so I think thats when it had issues.

 

[geniusintraining]

The only thing I can think of is to unplug all the connections while they are down, and use a cable tracer on each of the connections. This makes me nervous, I would hate to miss a controller and fry the comm port. It seems less than ideal, it is a large facility and it seems like an exercise like this could be quite time consuming.

So in RSLinx you should be able to get onto the network and see all the active nodes, then you just need to find the node and draw your own map, you will need to find out what is talking to what and that maybe easier said than done

 

[joseph_e2]

As geniusintraining stated, DH+ is (was?) a pretty solid industrial network and can tolerate a lot of mistakes in its wiring...but...best results will be found with best wiring.
It's supposed to be a daisy chain with no tees or branches and with a terminating resistor at each end only. I've seen it work well in weird arrangements with branches and stars, but that's definitely not ideal.

I hate to say it, but I think your best bet will be to trace the cables. I'd start with geniusintraining's suggestion to connect via RSLinx and browse the network with everything up and running. That will tell you what's out there and which channel of multi-channel devices is connected to what, but it won't tell you how the cabling is arranged. After that, I'd get out there and try to trace the cables as best you can. Visually first, then with a tracer if some disappear into wiring runs or conduits that are crazy. If you disconnect any cables from the terminals, be sure to note which conductor is on which terminal. It's not always done consistently and reversing them when you reconnect will make it not work and then you may have a frustrating time getting it back.

 

[Ken Moore]

Just a thought, is it one big network for some type of central SCADA or data collection?

 

[Jmeadows7676]

I was in the Same situation not to long ago and in the end just had to trace the wires. Net Decoders is probably the best Industrial communication software out there, but it just isn't capable of mapping the networks cabling. I'm Curious if the DH+ network started to deteriorate after adding the DHRIOs. This was my finding. The PLC 5 / SLC built in DH+ ports were far more fault tolerant than DHRIO not to mention way more powerful. Seems I remember getting ahold of a white paper that stated to only use 1 channel as DH+ communications and leave the other Off plus never have more than 16 devices attached to a single 1756-DHRIO card. If I can find the paper, I'll try to link it here. Good Luck.

 

[eight_bools]

Couldn't figure out how to do the multi quote.

@joseph_e2

They have had this issue before, where they were using non standard cable that had different colors and a vendor hooked it up wrong. Took them forever to figure out why it would not work. I believe they ended up running new cables before they figured it out.


@Ken Moore
It started as a small packaging line, with each machine talking to the conveyors and upstream equipment. Has grown well beyond that and the plant uses it mainly for connection with their laptop.

@Jmeadows7676
That's a good point about the channel 1, I hadn't heard that and if I can find the technote I can show it to them.

I am just guessing but they often have issues with brownouts, and losing power to the plant. One of the DHRIO bridges is set at node 1, and my theory is when they have a power outage one of their PLCs is losing its program and reverting to node 1. They then get a duplicate node and everything stops working. I know they struggle every time they start back up afterwards and the complaint is that they have issues on all 3 lines that make up this network.

 

[Jmeadows7676]

Couldn't figure out how to do the multi quote.

@Jmeadows7676
That's a good point about the channel 1, I hadn't heard that and if I can find the technote I can show it to them.


RA Technote 70199. Use the "to find an answer using a previous Answer ID, Click Here" button.
Alot is about Remote I/O but it also has a bunch of good information about DH+

Above you mentioned something about default DH+ address. I know with the PLC 5 DH+ only 1 channel "1A" will have an address on Powerup. Its set by dip switches. The others will come from the Program. As for the DHRIO cards. Address is set by Dip Switches.

 

[Jmeadows7676]

Also checkout the section about Bridging Messaging or Messing across the Backplane. Not following these rules has certainly caused me a lot of grief.