DH+/DH485 Citect problems.

Hi guys,
I'm having problems with our Citect trends from our PLCs based on our DH+ network at our workplace.
I've recently been put in charge of administrating the local PLC network and Citect system.
I am mostly self taught, but do know a reasonable amount about both, but have been stumped with the problems we are having.

Firstly our networking layout is as follows:
-Ethernet network connecting all our PCs, phones, and Control/CompactLogix PLCs and panelviews (roughly 80 PLCs), which all seem to be working fine.

-Connect to the Ethernet is 2 DHRIO cards, each with 2 DH+ channels on them. 3 of the 4 channels have no problem, but one of them is having a lot of problems with a lot of data loss on the Citect trend pages.

-Connected to the DH+ network is 2 DH+ to DH485 adapters, which the DH485 network has about 8 nodes.

The problem DH+ network has 19 nodes, 15 Allen Bradley-5/04 PLCs, 2 Allen Bradley PV-700 HMIs, and the 2 DH+ to DH485 adapters.
The total network cable length would be about 400m, and longest run would be 50m.

Citect is monitoring about 250 tags from the PLCs on that DH+ network, about 1/3 of which have a 1 second sample time, and most of the others have a 5-10 second sample time.

While monitoring the tags on Citect runtime, there seems to be periods of data loss, 5-10 mins at a time, several times an hour. It seems random, and not happening more during day or night. This doesn't happen at all on any other networks DH+ or otherwise.

Also we are trending 8 tags from the DH485 network, which have not worked since the start.

We have gone over all the connections several times, made sure that it has end of line terminating resistors, and physically it looks fine. There isn't any double up station numbers, and the LINK_ID are all the same for the nodes on that network.

I am thinking that it may be because there is too much traffic on the network, and we may have to split up the network and put some nodes on the other less populated networks.

Just wondering if anyone has any ideas as what to check, or any suggestions?

Thanks,
Ken.

Hi & welcome to the forum,

A couple of points

- What is your DH-485 baud rate? I did have some degree of success on changing a DH-485 network baud rate from 9.6k to 19.2k when used with Citect

- With the DH-485 network - is the maximum number of nodes at each PLC set to 8 or so? If it is set to the higher (default) value of 31, it will operate slower.

- Are you using the recommended (#9842) belden cable?

Cheers
Ian

Thanks for the quick reply.

Baud rate for the DH-485 is 19.2K

I do believe the max nodes is set for 31, as default. I will try to change this and see if it will help.

We are using the recommended Belden cable, all seem to be connected correctly.

Also when using RSLinx, I am able to see the nodes on the DH-485 network through the DH+ to DH-485 bridge.

kenny_o said:

The problem DH+ network has 19 nodes, 15 Allen Bradley-5/04 PLCs, 2 Allen Bradley PV-700 HMIs, and the 2 DH+ to DH485 adapters.

Click to expand...

Hi Ken,

That is a LOT of nodes... I had issues with 9 but we had a lot of data that was on the network

Ken said:

-Connected to the DH+ network is 2 DH+ to DH485 adapters, which the DH485 network has about 8 nodes

Click to expand...

What kind of adapter are you using for the DH485? DH+ to DH485? and also your others

This has some info (chapter 8) http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1770-um022_-en-p.pdf

You said that part has never worked... did the Citect data ever work?

Hi Ken,

I had similar problems on a long DH+ network some time ago. Fixed it by splitting the network into several segments, don't remember the exact details. I used a ControlLogix rack with one (or more, can't remember) 1756-DHRIO card and an Ethernet card (1756-ENBT or ..) and a Power Supply of course. The DHRIO card has two DH+ channels so one card can split a network in two, which you can then access via the Ethernet card. It all works automatically and you don't need a processor in the rack, although you could put the cards in an existing rack with a processor in it.

Hope this helps.

Cheers,

Pete

Hi Peter

I think... stress 'think' that is what Ken has

Ken said:

-Connect to the Ethernet is 2 DHRIO cards, each with 2 DH+ channels on them. 3 of the 4 channels have no problem, but one of them is having a lot of problems with a lot of data loss on the Citect trend pages.

Click to expand...

So im thinking he also has on the 1 (thats giving him issues) is a 1756-DH485... butI do agree thats could be more data then the ENBT can handle, if you have a EN2T you could give it a try and see if it can handle the data

We build the same that you are talking about and have had good luck with them
http://www.plccable.com/allen-bradl...6-pa72-dh-gateway-data-highway-plus-ethernet/

We also have this one and its about half the price
http://www.plccable.com/ethernet-ip...way-plus-dh-gateway-converter-ethernet-ip-dh/

But I would like to know if it ever worked... if yes then we just need to figure out what changed and he would not need to buy anything

geniusintraining: Hi, yeah we thought it was a lot of nodes aswell, but before the Citect system was designed/installed, it worked fine getting online/downloading/uploading to PLCs through the DH+ network through RSLinx.

The DH+/DH-485 system has been in place working for more than 15 years, it has only been since we installed Citect that we noticed a problem with data loss on the trending pages, I think due to an increase in traffic from Citect requesting a lot of information from the PLCs.

The adapter we are using for the DH+ to DH-485 is 1785-KA5P/A, and I believe they have been working fine for 15+ years aswell, RSLinx wise, albeit a bit slow.

The DH-485 network has been working fine viewing it from RSLinx through the bridge on the DH+ network, it is just Citect that has never been able to view data from it. I believe that the tags are setup correctly, but I could be wrong.

We do have the setup that you linked in your latest post, with the ENBT and DHRIO cards in the ControlLogix rack.

Peter: Hi Peter, we were planning on splitting the network in half to see if it would help, and also removing the DH+ to DH-485 adapter and installing a 1756-DH485 card into the same rack as the other two 1756-DHRIO cards, which should also reduce the traffic on the DH+ network.

I just have to convince my boss to fork out the money for the 1756-DH485 card, which shouldn't be too much of a problem.

If we do have problems with the ENBT card, we should be able to upgrade it to a EN2T.

Today the factory is stopping production for 2 days, so I will try to split the DH+ network up and see if that helps with the data loss problem.

Thanks guys for the advice, and I'll keep you updated.

kenny_o said:

I'll keep you updated.

Click to expand...

Please do as it helps us all learn, if it were me.... you have a good chance adding a 1756-DH485 to a rack and it may solve your issues, you may have a better chance of talking them into it if you get a used one... check ebay

Im still a little confused on your system anyway, why did they install a 485 when they had DH+, the only DH485 networks I have seen were using 5/03's not 5/04's... just seams easier to keep it that way, I know you can configure the port on the 5/04 to DH485 and some do that for comms for when using HMI that are DH485

geniusintraining said:

Im still a little confused on your system anyway, why did they install a 485 when they had DH+, the only DH485 networks I have seen were using 5/03's not 5/04's... just seams easier to keep it that way, I know you can configure the port on the 5/04 to DH485 and some do that for comms for when using HMI that are DH485

Click to expand...

Heh, yeah the factory is about 100 years old, so it has had many many upgrades over the years.

We still have 2 PLC5s, and also a about 8 or so SLC 5/03s and 1 or 2 5/01s, and some of them are on the DH485 network.

The factory is closing for good in 2 years time, but i have managed to convince the boss to get a 1756-DH485, so that should sort out the DH-485 network, until the doors close.

I managed to split the DH+ network, so now it only has 12 nodes instead of 18, and once we get the 1756-DH485 I can remove the 2 DH+ to DH485 bridges, reducing it to 10 nodes.

We finish production on fridays and start back up on sundays, so i wont be able to test the Citect trends on the DH+ networked PLCs until at least monday.

Thanks again for your help guys, I should be able to let you know how the DH+ network goes on monday.
Have a good weekend

One of the best things you can do to improve traffic (and it may already be optimized) is to ensure that you are reading/writing as many big chunks of data as possible. It's much more efficient to read blocks of data than it is to read words scattered over several files.

I agree about the number of nodes being a problem so your plan of attack is a good one.

Hi guys,
I'm just replying to say that splitting the DH+ network has made a huge improvement to the reliability of the network. Almost no data loss the past 2 days on Citect trending pages.

DH-485 still not working, so we are planning on getting the 1756-DH485 card to resolve this issue.

We did find that between the factory and our workshop (where the DH+ to DH-485 bridge is located) the cable used is just a standard non-twisted 14 core cable, of which half the cores are used for part of our fire system network...

Shows you can't trust contractors to always do a good job.

As the 2 DH-485 bridges are the only nodes on the workshop side of the network, we will just terminate it on the next node in the network, and run a new cable to the 1756-DH485 card.

Thank you to all those that gave advice, it is much appreciated.

Kenny.