Real Time Clock setting changes

Hello everyone –

I have a quick question about an A-B ML1200 real time clock.

I have an application that has been in service for over two years that (among other things) produces a test sound on a siren at 8P on the 4 Tues of each month.

The owner called me 2 weeks ago and said that someone noticed that the siren was about 30 min early.

I checked RTC under the function file, and sure enough, the time was about 30 minutes ahead.

My experience with PC's is that their clocks can fall behind when the PC sees a lot of activity (maybe a busy CPU??) or other times I have notice that the time actually gains (I seem to notice this when the PC was on, but not used much for a long period of time).

Does anyone else have any experience or comments on PLC real time clock time changes? Is there something I should be checking?

Thanks in advance -

OFG

I've never seen a computer clock that was perfectly accurate. I know little about micro-electronics, but someone once explained it as variations in clock crystals in the computer chips combined with voltage variations to produce these errors. I have two networks of PLC5s and a network of SLCs and have a screen that shows the "clock error" on all of them at once. The SLC clocks seem to all keep time better than the PLC5s which lose (or gain, I forget) a few seconds a day!.

As far as I know, you just have to come up with a procedure to correct it, either manually or through a communication channel automatically.

Okie -

Thanks for the response.

I had kind of assumed that their clocks weren't that accurate, but this is my first ML1200 and I did not want to automatically lump it in with the old stuff.

I will make arrangements to check it more often for a manual reset.

OFG

I don´t know what is the accurance it has and what do you need, but if you need presition you should make an external sincronization with a presition clock or with a PC (or whith a display) that keep the time being chequed.

If it's connected to an HMI, you should be able to do as Hector suggested and provide a means for a trained opreator to synchronize it. If not, his other suggestion of using a more accurate external clock to turn on an input at midnight every day, for example, sounds interesting too, but having a means to manually set it, and making it readily available will always be useful, if not required, since the application is time dependent.