Help! The more I read the less I am sure about which network to use for long term SCADA viability, DeviceNet, Profibus, Modbus/TCP, EtherNet/IP from ODVA, etc., etc., etc. Each has its benefits, but so did Beta video tapes, remember! My experience says that the desire to integrate the office to the production floor, and the millions of lines of driver code supporting Modbus RTU that the most likely candidate would be Modbus/TCP. What do you nerds think?

Thanks!!

 

The "Busses" question depends on what you are using your system for but it should be capable of integrating a number of busses as each bus has its own benefits in different applications> if its for a process then you need as a minimum Foundation FieldBus and either ProfiBus Dp or ASi-Bus. For more help try visiting this
site. "www.easydeltav.com":http://www.easydeltav.com

 

Dear David,

A good start is to search this web site for "61850". IEC 61850 is a new standard for
substation automation - applicable for RTUs (providing a seamless communication means).

Almost all major vendors of substation systems are involved in the definition of the standard IEC 61850 (applying Ethernet, TCP/IP, ...). Most of them have first 61850 compliant products or have announced them.

A comparison of the major SCADA/RTU standards including IEC 60870-5-101, .., DNP3, and 61850 can be found under:

"http://www.nettedautomation.com/news/n_51.html":http://www.nettedautomation.com/news/n_51.html

IEC 61850 at the DistribuTECH in Miami Beach, Florida, February 27 - March 1, 2002

"http://www.nettedautomation.com/news/n_54.html":http://www.nettedautomation.com/news/n_54.html

Best Regards,
Karlheinz Schwarz, [email protected]

 

Depends on the application somewhat. If you're running remote I/O and devices around a machine and don't need full SCADA access to every possible
data point, then Profibus DP will work admirably. It's very easy to apply too, i.e. you don't need a post-graduate qualification in networkese to get
it up and running.

The only reason for choosing DeviceNet would be if you are tied to Allen Bradley. I'd be interested to know other perspectives, but I do feel DeviceNet is rather slipping away in terms of popularity and support (which is rather a shame as I just finished a slave implementation 8 months ago!). ControlNet and Ethernet/IP currently don't appear to me to be well supported, and are quite heavy in terms of resources required so you won't see a proliferation of low cost small devices and I/O units.

If you need DCS style access (i.e. to hundreds of data points on a network at any given time), I would personally look at running a Modbus/TCP backbone into RS485 Serial Modbus gateways. This combines the benefits of a 2 wire multi-drop network around a machine and cheap devices, with Ethernet's facilities for allowing multiple masters on the bus (i.e. allowing simultaneous connections from a PLC and panels).

All views personal, of course.

Tim Linnell (Eurotherm)

 

I think DeviceNet does OK at what it says it is -- a "Device Net". I think it was conceived partly as a way of networking sensors and actuators on a bus topology -- hence the included 24V power conductors. As a network for connecting racks of
remote I/O or distributed processors, I would prefer to use Profibus DP for the very reason you stated, and because of its much greater distance capability. Modbus Plus is worth considering, too.

[email protected]

 

You will be in this boat quite a while. So are the rest of us.

For what it's worth, we've choosen Profibus-DP and are adding Modbus/TCP. I do not hesitate to recommend either of them.

-James
Sage Automation, Inc.

 

I have used several of the main networks. I would choose a modbus network. there are many devices that support this and it is cheaper but just as consistant as many others. many plc's have modbus cards you can add , plus there are many converters availible. check out "www.niabrara.com":http://www.niabrara.com or square d, "www.powerlogic.com":http://www.powerlogic.com .

there are plenty of free opc server/client activeX freewares availible. Second choice would be to place a hardware gateway in the middle of the devices such as controlnet, that way you could bring many protocols back to one location, convert, and process.

Thanks!!

 

My opinion would be to use an ethernet network. The reasoning is that it does not require any expensive/proprietary hardware for the computers. Computer technology changes rapidly, but ethernet will be supported in the pc, however as can be seen when industry was caught off quard by the switch from ISA to PCI it may take time to get the industrial protocols supported under a new hardware/software platform.

The next question is which protocol? My recommendation would be Modbus TCP/IP as it is the most widely supported today. GE Fanuc, Modicon, Allen-Bradley (via prosoft) all have an offering for Modbus TCP/IP and this cannot be said for the other protocols. I would suggest that you look at all the device manufacturers you use and see what protocols they use of course.

 

To me SCADA is like BAUD it is still used but no longer carries the same meaning. You can still by a system that can supervise single loop controllers but why. Today systems Emerson Delta V, Honeywell PlantScape, GE Cimplicity, Foxboro IA, and even Wonderware are integrated Human-Machine Interfaces to a control system that can be controller remotely.

I am responsible for five offshore projects for my company. Three are Delta V/Hima using 4-20 mA, discrete inputs/outputs, OPC, Foundation Fieldbus, Profibus, and Modbus over RS-485. Two project are using Triconex interface to Foxboro IA workstations. The workstations are 50 mils away from the platform and the operators onshore are controlling the platform via satellite. The Triconex systems are using 4-20 mA and Modbus. So you can use several interface methods on a single control system. If you would like to as specific question please contact me at [email protected].

 

Hello David,
Without getting into all the permutations of legacy constraints, IF you have the latitude to install anything you want, AND your factory is more like a process plant than an assembly line, AND you can live with the device and vendor selection THEN you may consider open source Foundation Fieldbus (FF).

While technologically sound, FF doesn't yet seem to be the solution for every application.

As you probably know, one of the big benefits of the advanced protocols in the process industry is that smart field devices can be networked together across the plant without hard-wiring to a local controller, eliminating significant hardware and labor installation costs. There are other benefits as well, eloquently enumerated at:

"www.fieldbus.org/About/FoundationTech/":http://www.fieldbus.org/About/FoundationTech/

The downsides include somewhat limited vendor support (lots of big-name vendors, few actual FF registered products), process industry bias, and spotty local expertise.

There is an interesting archived webcast sponsored by Control Engineering magazine that overviews FF, free registration required (registration link at upper right):

"www.controleng.com/webcast/archives/fieldbus.htm":http://www.controleng.com/webcast/archives/fieldbus.htm

Surely you will receive as alomost many different opinions on this topic as replies. I'm very interested to hear if anyone has experience using Fieldbus in a manufacturing floor environment.

Regards,
Brandon Nichols, PE
[email protected]
http://www.bwnichols.com

 

Considering the fact that concept of open systems is gaining popularity, all will try to join hands in some or the other way.

Ethernet and Modbus/TCP are commonly understood by a large part of our society.

I think these two are here to stay for a longer time....at least till the time Marvin* thinks of something new.

*=Hitch hickers guide

 

Is the choice of a protocol really an issue? What you have already received as replies should show you the true answer: there is no "right" protocol at the moment and even if there was one, it won't be relevant next year!

My opinion is that you should introduce in your reasoning the "front-end" box. This box should help you to keep your previous hardware & software investments if any AND prepare & protect your next investments.

OK, this is an additional initial cost but put that in front of the cost of add-ons you'll have to develop in the lifetime of your system.

All of us we have had such problems to tackle with.

Be honnest what's the cost of such "front-end" bos w.r.t. the engineering work to go round the limitation of existing HW/protocol/SW?

For instance, modern and old fashion protocols will fit more or less to your technical requirements but to date they won't give you an answer for security against eye-dropping. If you intend to reduce cost of your system by using browser oriented SCADA, you may need to take care of that: where are the standard solutions/protocols?

Your choice shouldn't be only technical but also economical: look at the life time cost of your system!!!

There exists already such boxes, depending on your requirements and system architecture, prices will vary.

If you want further comments, please contact me by email, I.T.Lity Ltd does not sell or represent these companies nor sell consultancy.

Kind regards,
Patrice ([email protected])

 

Modbus ??
Certainly TCP/IP, Industrial ethernet, and Profibus... I believe Profibus was recently adopted as the standard in China, it already is in Europe. Nuf said.

 

Depends on so many things... Are you using SCADA systems to talk with PLC/DCS or RTUs or field devices. Are you setting setpoints on recipes? are you using it for fast (ms) data acquisition? level of knowledge of your crew? Is IT (dirty word, dirty word) involved in the setup? Will you need access from your computer? Will you require access from remote sites like your home? etc etc.

I personally see SCADA systems as a bridge between the plant floor and the enterprise. Somewhere in the MES world. In the MES world, I am a firm supporter of Ethernet technology. If you go to your plant manager and start talking about FF, Devicenet Profibus, Modbus, etc, he/she will probably look at you like you are talking chinese. But you mention Ethernet, and they say "yeah, I know that...I have seen the drop in my work area.". Ethernet is becoming more and more robust.

Communications can be managed via Switches and routers. And it connects nicely with plant floor and enterprise networks. And you IT (spit again), have an idea of what to do to run a drop to your area. And they feel happy because they think that they are winning the communications war...Hahaha

Marco, PE

 

I think the best one is Ethernet/IP, nowadays everything is focusing on the ethernet, Lots of equipment.

 

As a proponent, the answer for me is easy. Modbus TCP is the way forward at this point in time. The modbus instruction set is truly open, not proprietary. Check out "http://www.modbus.org":http://www.modbus.org Here you can check out the standards, realize the virtues and download some useful information. Let's face it, almost every automation vendor, many control instument providers and other equipment vendors have Modbus connectivity. Using TCP/IP as the transport is the next logical step. Move on or move over I say.

 

My past experience with this issue has led me to look at it from the standpoint of what will best suit my needs in the long run using the following criteria.
1. How large is the installed base or how many places use it? (If you know someone, get feedback)
2. How long has it been out there in the marketplace?
3. If 2 = a long time(5+ years), is it soon to be obsolete?
4. Is it used only in specific industries or does it crossover into other industries? (like eithernet does)
5. Is it proprietary or open technology?
6. If proprietary, does the manufacturer support it and what is the cost of the support?
7. Does it require special tools/training/personnel and large costs or hard to find materials to support?
8. What is my required product lifecycle and will it meet this need?
9. In todays world situation will I always be able to get and support it?(Foriegn vs. domestic Mfr.)
10. What is the long term cost of supporting the installation? (Continually required updates/ revisions?
11. Will it meet my future needs? (remember, imaginings of the future often fall short of reality in the controls business)