J

#### Joe Bloggs

I have been asked to make a comparison of the two systems and offer an opinion of whether an existing PLC-5/60B system using the BCM hot-backup should be upgraded to the ControlNet option. Some people on site are pushing for the latest technology while others are resisting change.

We have experienced random problems were the BCM's do not change over despite all the right diagnostic information and are concerned controlnet will not offer better perfomance.

Can list members offer their experience on these two.

Joe

M

#### Mark Fairbaugh

I am not sure about the control net but we use the BCM's here on PLC5/40 and we have a very high BCM failure rate. The channels just lock up and take down the whole DH+.

Mark Fairbaugh
PT Newmont Nusa Tenggara
Sr. Process Control Engineer
e-mail: [email protected]
Phone: (62)-370-636-318 Ext. 47948

M

#### Michel A. Levesque, ing.

We've used PLC-5's with ControlNet hot backups before as well as BCM's. Unless the customer specifically asks, we will always use the ControlNet hot-backup system from now on. The ControlNet version is much easier to implement and to troubleshoot. The reliability is far better than the BCM's. The ControlNet version is
much closer in function to a real hot backup than the BCM's. Setting up a ControlNet is not hard but you MUST have the proper versions of PLC-5 and 1771-ACN's (or ACNR's). The ControlNet hot backup only works with 5/40's and 5/80's as I write this. You mentioned resistance to change, this is probably due to the fact that setting up BCM's and getting them reliable is a hard process. Your guys must not want a repeat of that scenario. I would just mention that setting up a ControlNet hot-backup is easier, faster and more reliable. One important point, we only use ControlNet linked I/O and not RIO. If your plant is on RIO then you still need the BCM's for RIO switching. The big winner for ControlNet is how easy it is to do crossloading and the amount of diagnostics available. I highly recommend the ControlNet hot backup. But, don't just take my word for it, call AB and get them to demonstrate a ControlNet hot-backup, then judge for yourself.
Email me directly if you need more details.

Michel A. Levesque eng., mcp
Directeur Bureau Montreal
AIA Inc.
[email protected]

B

Hello Joe,

I would suggest that potentially the ControlNet backup solution could be a superior solution provided it is implemented correctly. The biggest single problem with the A-B "Hot Standby" system with BCM modules and standard RIO is that it provides processor redundancy only! The question you have to ask yourself is; "Is the PLC-5 processor the failure point "weak link" in the
control system??" It has been my experience that the answer to this question is always NO. With the PLC-5 BCM module "Hot Standby" system, the only redundancy you get is the processor.

The ControlNet "Hot Standby" system could potentially be better but the cost involved in implementing correctly will be very high.
1) Processors need to be changed to ControlNet processors.
2) The existing RIO network must replaced with ControlNet.
4) For maximum effect the redundant cable option must be implemented.

Once again the question to ask is; "Where is a failure most likely to occur??" The answer is; "Not at the processor!"

My $.02 worth. Brad Hite Intertech Incorporated mailto:[email protected] http://www.myplc.com Teaching Practical Skills for a Technological World J #### Jack Gallagher Brad, I think that your reasoning is correct in saying that the processor is not as likely to fail, but a processor failure is a fatal failure. Most other failures are non-fatal, where you may lose an I/O card, or a wiring connection. These failures are localized, where the loss of a processor is a complete node failure. Anyone else have an opinion or suggestion on redundancy? B #### Brad Hite Hello Jack, I agree that an I/O failure or card failure is localized and therefore may not be as serious as a processor failure. However, it has been my experience especially with the A-B PLC-5 product that about the only time a processor failure occurs is do to a programming error (i.e. a negative value in a timer preset). If the error exists in the Primary PLC the error will also exist in the Standby PLC. Both processors will go down and the money spent on "Redundancy" will not have accomplished anything. <Flame shields on> My personal opinion of the PLC-5/BCM "Redundancy" architecture is that it is a nice marketing/sales tool but not overly useful in application.<Flame shields off> My$.02 worth again.

Vice-President
Intertech Inc.
mailto:[email protected]
ICQ# is 38746036
http://www.myplc.com
Teaching Practical Skills For a Technological World

S

#### Sagi Sebastian

I agree that an I/O failure or card failure is localized and therefore may not be as serious as a processor failure. However, it has been my experience especially with the A-B PLC-5 product that about the only time a processor failure occurs is do to a programming error (i.e. a negative value in a timer preset). If the error exists in the Primary PLC the error will also
exist in the Standby PLC.

The processor will fail mostly due to power supply failure individual)/loss of program in memory/electronic card failure etc. in all these cases since the program for primary and secondary are loaded separately in different RAM / EPROM
I don't understand how the present value of primary will get into secondary. There is no chance at all. If the primary processor fails then the secondary processor will run the show with the old values if the values before the failure occurs

sagi sebastian
Asst. Manager
Reliance petroleum
mail to : [email protected]

V

#### Vitor Finkel

I do. The I/O rack communication module fails almost as much as the CPU, and that is also fatal failure. As far as I can tell there is nothing you can do to improve that. There is no
hot-stand-by for those communication modules, as for the CPU, so you pay lots of money to improve overall reliability due to CPU failures, but still
have to live with communications failures, that are as fatal as CPU's.

That is probably why I was never convinced to spend money on Hot-stand-by configurations. One more thing I've heard more than once: When really needed, hot-stand-by CPU's carryovers are known to fail, even when they've been successfully tested before.

Vitor Finkel [email protected]
P.O. Box 16061 tel (+55) 21 285-5641
22222.970 Rio de Janeiro Brazil fax (+55) 21 205-3339

C

#### Chuck Miller

Several factory supported hot standby systems are available that do include power supply, processor and communications redundancy for either simplex
and/or redundant I/O configurations. These systems typically counter the downtime caused by "Common Mode Failures" (fatal failures). These systems have been thoroughly testing with millions of hours of reliable operation.

Shutdown could be the result of a detected I/O failure in a system where the logic is designed to provide that function. The real problem results from not having the diagnostic capability to detect those anomalies (i.e. the I/O cannot respond to a system demand). To maintain reliable control, the I/O must be intelligent and assume at least part of the decision as to the health of the internal communications subsystem as well as provide basic diagnostics on its own health.

AB is certainly a strong competitor in the PLC marketplace but I don't see them as the industry leader in redundancy. It might be good advice to look for products that are designed and tested for this application. Otherwise, you will always be relying on the ability of the configuration engineer who wrote the switchover program ........ if you can find him.

PS - If you are going to have two PLCs anyway, why not have them both on-line and voting. There is really minimal increase in cost or complexity?

Chuck Miller
GE Fanuc Critical Control Business Development Manager
15511 Rio Plaza Dr. Houston, TX 77083
(281) 495-0333 Fax: (281) 495-0370
www.gefanuc.com