PLC Network Configuration

[renosis]

I work in fiber-optic networking and was tasked (a few months ago) with mapping the existing fiber network of a local waste-water plant in my area. The network was about 10 years old and it was a complete mess and nightmare (none of the strands were in standard color order). It took me a few weeks to get it all mapped.

The PLC network consists of 11 PLC's in 11 different buildings, running on a product called "Data Highway" by Allen Bradley on mostly PLC-5's. These 11 PLC's connect to a PLC in the main building with 5 x 2 channel network "cards". Each card has a channel A Transmit(Tx) and Receive(Rx) and a channel B Transmit(Tx) and Receive(Rx).

My question is, for some reason, each of the connections coming in is split between each channel on each card. For example, PLC-1 fiber comes in to the main building and is patched to CH A Tx and CH B Rx. I can't understand why they (whoever designed this originally) would split the connections like that?

I was thinking maybe they are daisy chained or are on some kind of funky ring topology, but this doesn't make sense judging by the way it is connected, because there would be "breaks" in the ring.

I was going to recommend to the plant operators to redo most of the fiber infrastructure to establish some kind of standards and I would like to get away from the split channel thing they have going on.

So I was wondering if anyone here knows of a reason why they might have it configured like this?

Thanks!

 

[catserveng]

DH+ is an older network, but still used in a lot of sites, here is an overall Allen Bradley reference document,

http://literature.rockwellautomation.com/idc/groups/literature/documents/td/1785-td006_-en-p.pdf

You can also look here for more specific information.

http://ab.rockwellautomation.com/Networks-and-Communications/Data-Highway-Plus-Network#/tab5

A lot of sins got committed over the years with these installed networks, the would continue to work, but slow way down. Take a look at teh info and see if that helps, otherwise let us know.

Mike L.

 

[renosis]

Thanks Mike L. for the response and the links.

I looked through them, but didn't find anything related to the fiber network configuration, but it was helpful regardless. I then realized that the fiber-optic modules are made by a different company called Phoenix Digital. I managed to find this manual from the Phoenix Digital Website:

http://www.phoenixdigitalcorp.com/pdf_directory/Rockwell.pdf

It says the following:

<quote>

INSTALLATION
Phoenix Digital’s fiber optic modules may be interconnected on the fiber optic network in an active bus
or fault tolearant ring configuration, using either multimode or single mode fiber optic cable. Fiber
optic Channel A and B inputs and outputs may be interconnected sequentially from fiber module to
fiber module to create a bi-directional, active bus. (Connection details are given in the product users
manuals.) This configuration may be made fault tolerant by cross-connecting the fiber optic modules
on either end of the active bus. (See Figures on pages 9 thru 12.) This effectively transforms the
network into a fault tolerant, redundant, self-healing, counter-rotating ring configuration for
ControlNet, Ethernet, Fast Ethernet (ALL Protocols), Data Highway Plus, Remote I/O, DH-485,
DF1, Modbus, or RS-232/ RS-485 networks, without requiring any further action by the user.

</quote>

So, I suppose I can now see why they may have thought it was a good idea to set the network up this way. But really, there is no advantage to it, it is actually a disadvantage. All 11 remote sites that have a PLC in them "home run" all the way back to the main building into this one PLC-5 1771 Module Rack Backplane thingy. So, if it is setup like a ring, I would imagine that would slow down this network significantly. And I don't see any benefit from redundancy, if one of these connections is broken, it is not like the ring will be able to double-back on itself, because the network isn't setup in a ring fashion.

But, I am also hesitant to completely change the configuration, because I know it at least "works" now. Still doesn't make sense to me though.

 

[jraef]

Sounds like an early implementation of a redundant ring topology. A lot of fiber optic networking ideas came from the Power Distribution industry in the early days.

This is a pretty good explanation of the design philosophy;

http://www2.selinc.com/techpprs/6087.pdf

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[itsmoked]

I see a lot of resistance to them, "Fixing something that ain't broke" And, it doesn't matter how slow it all is - if it's fast enough.

Keith Cress
kcress -

http://www.flaminsystems.com