ControlLogix network: Ethernet vs. Controlnet

Derek-

You have me there. I have to get going so I won't be able to look into it right now.

The C-Net side is pretty easy. RSNetworx for ControlNet won't allow you to set a NUT/RPI combination that it can't handle. So your NUT effectively defines your throughput.

With the Ethernet stuff I think I just kept on going down with the RPI for the produced tags until the data stopped getting through. I don't remember on that side, though. That's where it gets fuzzy.

Try some setups and let us know what you see.

Keith
 
Keith, You are correct when you state "deterministic" does not necessarily equate to "speed" but is does equate to "reliability" and "consistancy" when generally compared to Ethernet.

An example I have used before is a power station where 9 PLCs are networked back to a SCADA computer. There are about 3000 to 4000 words of data peer to peer on the network plus what the Citect SCADA is pulling out of the PLCs for monitoring and trending. The SCADA system reports as high as 15,000 digital reads per second and about 130 word reads per second.

The network is a proprietary PLC brand token ring and only runs at 2 megs over a twisted pair. The PLC network scan time rarely exceeds 30 milliseconds. That is very reliable when generators have to be taken off line sometimes in a great hurry.

On the other hand, I had a job where I had to use Ethernet. It was only the old BNC plug type which did not run real fast but was considerably higher speed than 2 megs - 8 megs if I remember correctly. The data peer to peer was minimal - about 30 words every 2 seconds. The SCADA data was also minimal - about 400 tags. The performance was atrocious in my view. I had no way to really monitor the response times accurately but sometimes a receive command in the PLC, that was only being requested every 2 seconds, took 4 o 5 goes to receive. I finished up hard wiring inputs from PLC to PLC because if the neutral earth switch opened (11 kV generation) the 11 kV goes floating well above ground and the generator breakers have to be opened in a hell of a hurry. By the way, it was a private network (address 10.....) and there were only 4 PLCs and the SCADA on the network. Really annoyed the IT people I can tell you but I told them that after 12 months and a day (defects liability) they could do what they wanted with the network.

On the other hand, the token ring network I normally use has a set of tables that can be put together and placed in every PLC. The data is automatically placed on the network and each bit/word is "seen" in every other PLC as the same bit or word number. If the network controller goes off line, the next one takes over. It is nice and redundant.

I am afraid your 10 second network scan time is absolutely useless in most of the jobs I do due to the fact that I have to get generators off line in a hurry sometimes.

It is really a case of horses for courses. If you can live with data crashes or slow scan times, Ethernet is fine, although I understand that it is much better with managed switches but have not had the opportunity to try that type of network out. With the number of data services companies out there that use this type of setup for high speed data, I can only assume that it is really quite good but there are still crashes.

Another reason for using something like ControlNet or another proprietary network is that IT do not understand it and will usually run the other way. That is one very good reason to NOT use Ethernet in my view.

I am in full agreeance with all the suggestions to use ControlNet for peer to peer and Ethernet for HMIs is fine. The SCADA system can run on Ethernet OK provided there is no really critical data that is required at very high speed. For example, on the power station job we were trending frequency on the main bus and each of six generator sets every 250 ms in the SCADA so that a diagnoses could be done of each set against any other set on line and the bus when a tree fell over a line for example. The philosophy is that the protection is set high enough to keep the generators going and "blow" the tree away without shutting down the power station. I might add that is works very well. Rarely a generator(s) is taken off line by a falling tree and rarely does a feeder trip from a fallen tree - it is simply blown away. But back to the high speed monitoring of frequency, it assists the station manager to assess the performance (recovery from a hit) with respect to another set. This data can then be used to determine whether there may be a sticking rack or injector in a set. Very useful. Due to the high speed trending requirements, Ethernet was not considered for the job.
 
TimeFluxCap said:
In 2007 Rockwell will have Ethernet I/O for its Guardmaster Safety PLCs. What does this tell us?

Ethernet can be determenistic?

Come on now, it can't be that hard.
 
Some manufacturers have a "global I/O" system that runs as a layer on TCP/IP. I here this is OK as it is not Ethernet and is deterministic.

The only thing that worries me is that BACnet runs that way and my experiences so far with BACnet have not been happy ones.
 
I think we all have to agree that technology is changing..Ethernet today is not what it was 2 years ago..Has control net changed any??

I must admit i am speaking from a morons point of view..i have never done anything with control net..however a few things do pop up..

I set up 1 control logix processor with 1 HMI and a remote rack over ethernet with NO formal training on ethernet or controlnet or realy any kind of training..It was a challenge for me but easy enough in the whole scope of things..after hearing about scheduling and stuff with Cnet i do agree that to the average Joe Blow this might be more than he can undersatnd..

Not that i understand ethernet!! However becouse it is so common place i am comfotable with it..perhaps it is the comfort thing that people look for..

In my application i have a resolver transmitting postion to the remote chasis.this in turn tells me whats happening..I have it set up that i dont recieve a postion from the resolver every x amount of ms then stop an fault the processor..(Postion is critical)..The processor shut down on me once..becouse i was doing other things on the machine so i changed the hub with a wirless.I think the wirless hub created a bit of lag..Its been the only shut down in about 1 year..(the machine runs 20 hours a day 5 days a week)

I guess what i am saying is go with what you are comfortable with..

D
 

Similar Topics

Hello guys, I have a network card ehternet 1756-en2tr configured in linear mode , not ring mode. And in RS Linx I can not see the devices that are...
Replies
5
Views
2,703
We're just looking at setting up a redundant ControlLogix chassis pair. The pair communicate to a single I/O rack via Ethernet/IP, and we will...
Replies
6
Views
3,352
Hi, we have a machine with a ControlLogix rack and 2 EN2T network cards in that rack. One is for the normal Ethernet/IP subnet (192.168.11.xxx)...
Replies
2
Views
2,873
We have a ControlLogix processor with a DHRIO module and are installing a PV Plus 1250 with a 2711P-RN6 DH+ module. There is currently only DH+...
Replies
0
Views
1,616
Dear all, I have just arrived to a new site and found very strange Control net network. Just to explain the situation. there is redundant CLX...
Replies
5
Views
5,208
Back
Top Bottom