I recently read an article on Embedded controls about PLCs and the cyclic rollover problem; it seems this should be a well known problem amongst engineers, but I have never heard of it before: does anybody know about what this is??

Thanks

Carlos Fernandez

-- Anonymous, December 22, 1998

Answers

Carlos, The only place I have ever seen reference to a PLC "cyclic rollover problem" is in an article by Roleigh Martin entitled "One Year 2000 Embedded Systems Expert In Core Infrastructures Speaks Out Anonymously ". The article quotes the anonymous expert as saying about those involved in Y2K "They are also unaware of many problems in embedded systems which have been recently discovered. For instance, they were unaware of the cyclic rollover problem with certain PLCs... [One] must make sure the experts really are experts."

What is more likely the case is that no one knowledgable about PLCs has a clue what the term "cyclic rollover problem" is refering to - do a search for the phrase on the internet - you just won't find it. You may wish to ask Roleigh what the term means.

Now, if I had to guess at what Y2K PLC problem the expert was trying create a new name for, it might be the date rollover problem that occurs during power down for some PLCs. For some versions of the Allen-Bradley PLC-3 family, the company supplies the following statement at: www.ragts.com/webstuff/y2k.nsf/Pages/Brands-Allen-Bradley-Known+Issues ?OpenDocument

"The PLC-3 will rollover and have an accurate leap year in power-up mode only. The PLC-3 clock function, during power down, will hold the time and date at its last value prior to power down (assuming the battery is capable of holding data). Upon re-application of power, the time and date will need to be re-set. The PLC-3 did not rollover or recognize leap year in power down mode prior to 1/1/2000 and will notafter 1/1/2000. NOTE: The user program in the PLC-3 is completely backed by the battery."

Also, for some versions of the PLC-5 the vendor states: "The processors pass all the tests except leap year in powered down mode. The processor, if set to 2/28/2000 and powered down through the date change, will power up with 3/1/2000. If the power remains ON through the date change, the correct leap day is displayed. This is not unique to the Year 2000. Expanded tests shows the same results for 1988, 1992, and 1996. The important thing to remember is that this does not stop the processor from running unless the application code relies on the date for execution. There will be no fix for these processors."

These aren't exactly earth shattering Y2K problems and are easily managed. You'll find Allen-Bradleys home page is at www.ab.com/, and you will find links to more Y2K information there. You might also notice that many of AB's PLCs don't even have a real time clock.

Hope this helps, David Mustian mustian@bzn.com

-- Anonymous, January 11, 1999