if clocks rolled over already what happens to extra day?

greenspun.com : LUSENET : Electric Utilities and Y2K : One Thread

If Nova Scotia Power has rolled their clocks forward to Year 2000 already, as they have stated, and are leaving them there, what happens to the extra day (the result of Year 2000 being a leap year) at the end of this year? Does it have any effect at all?

Just wondering. Thanks.

Cath, February 11,1999

-- Anonymous, February 12, 1999

Answers

Cath,

There are, thankfully, seven days in a week and that will never change. No matter how many days in any particular year.

So if for example a building airconditioning system has a digital control, more than likely it will be on a seven day cycle. Just goes on endlessly, seven days at a time. Same sort of thing happens to elevator controls.

-- Anonymous, February 12, 1999


Does anyone know a source for this information that I may investigate?

-- Anonymous, February 12, 1999

Well now, good question, Cathleen!

Stupid answer: I don't know; if calendar clock synchronization was a consideration in system operation, obviously this would be a problem. I don't see this to be a significant issue in electric system operations, but I could be wrong! We'll see on March 1st, and possibly as late as a few weeks after for those facilities that are allegedly operating beyond 01/01/2000 currently. This is only one of the questions I have about advancing the "clocks" (bad term, but I'll go with it) in 1999. There's the potential for inducing problems in the systems in 1999 due to to things:

1) If the system can handle the transition to 1/1/2000, but for some reason can't handle the leap year function for 2000 (2/29/2000). There are quite a few documented cases of this characteristic - another instance of programming short sightedness.

2) Screwing up the calendar characteristics this year - ie. for those systems already advanced beyond 1/1/2000, the systems will be operating with a date of February 29, rather than March 1st.

Will this be an issue? Dunno. Stay tuned. We don't have long to wait to see.

Troy: again, Mark Frautschi's embedded systems whitepaper has a pretty good explanation of the whole issue of embedded controls and Y2k.

Hope all of this helps...

-- Anonymous, February 12, 1999


When the systems in these power plants are rolled forward for testing, are they certain that all embedded devices are moved forward? Or are the testers just concerned about the DCS and SCADA computer systems. It's my concern that no matter what date the computer systems read, there are embedded devices that don't get the current date & time downloaded by the control systems as part of normal operation and therefore aren't reset along with them.

These devices wouldn't fail during the tests because they're still operating within valid date parameters of 1999. They still present potential failure points at the rollover as the date become invalid.

People will believe that their plant is Y2K ready because it didn't fail when the computer's clock were advanced. Are there procedures in the roll-ahead tests which verify the embedded devices are advanced with the control computers?

Otherwise, it could be a rude New Year's greeting for some plant operators.

WW

-- Anonymous, February 13, 1999


Rick,

Perhaps I was to brief in my request for more information... What I was hoping for was more information concerning Nova Scotia Power's year 2000 simulation. I strongly believe that it is possible to find out for sure if all of the clocks were rolled forward. I am also interested in the procedure one uses in a plant in order to do this. All of the details I can find would be helpful.

I did read the report you mentioned and found it insightful (although at the time I read it the footnotes were off by one. I figured it out.) Thanks for your help.

-- Anonymous, February 16, 1999



Troy, on another thread in this forum there was a post by Carol of a letter from a rep in Nova Scotia. The e-mail for this person was:

jill.lawrence@nspower.ca

and Jill Laing-Lawrence also wrote: " If you have further questions don't hesitate to e-mail me or phone us at (902) 428-6578."

Perhaps you could contact this lady and she could direct you to one of their engineers involved in the date rollover? Might be worth a try, especially since you're in the industry, too. They may be glad to talk to you about the process.

-- Anonymous, February 16, 1999


Wild Weasel:

If the embedded chips are not time synchronized, there is no way of telling what date they think it is. They generally start with an epoch date at first power on. Depending on the design, it might restart at that epoch date every time it is powered on, or it might store and continue, or it might keep time.

The point of this is that this kind of embedded chip probably won't fail 1/1/2000. It may not ever fail, if it resets every time it is powered on. If it doesn't reset, then it could only fail when the internal clock hits 2000, which has little or no relationship to the external calendar.

As an example, think of such chips with an epoch date of 1980. If it resets every power on, it will have to be on continuously for 20 years to hit 2000 internally. If it doesn't reset and stores the date, it will have to have been powered on for 20 years. If it keeps time, it will fail 20 years after it received power. If the chip wasn't installed until 1991, it won't fail until after 2010.

Frautschi's (sp?) article covers this in much more depth.

-- Anonymous, February 17, 1999


John,

I'm just thinking that in my experience, most PLCs and other controllers will retain their date calculation information long enough to cover an plant outages where they may be installed. This would have the effect of power being continuously applied from their first power-up, which causes me concern that there will be some form of embedded device problems or failures on or about 01/01/2000.

In the industrial applications I'm involved with, we've done a good deal of checking embedded devices as part of our evaluation and testing efforts. I'm wondering if the power companies are checking every aspect of their facilities the way we're digging into ours.

Not to cast doubt on their efforts, but there have been press reports of follow-up audits finding enough "missed" embedded devices to possibly shut-down or limit operation of previously tested plants. A control panel with a half dozen embeddeds comes to mind, but I can't recall which audit report I saw it in. I'll have to retrace my steps on the web for the past few days to find that post.

WW

-- Anonymous, February 19, 1999


Moderation questions? read the FAQ