Real Time Clock Accuracy

I am on a quest to find an accurate clock in a PLC- by accurate I mean a minute or two a year. I currently use Unitronics PLC's which are accurate to +/- 1 minute per month at room temperature. Every other small PLC vendor has the same spec if they even have a clock at all. I don't have a way to communicate with the installed units so we can't reset the clock that way.

I'd like to throw this topic out and see what experience others have had.

This is an interesting topic, often overlooked. PC and PLC clocks are much more subject to temperature and other drift factors than many people suspect.

Even my Company's big, powerful controllers, which have a microsecond-accurate clock for motion control, accumulate enough error on the realtime clock to need periodic resetting for time-of-day timestamping over the course of a year.

Some models of PLC's, when attached to a network, can use the Network Time Protocol (NTP) to get the current time from a time server. Most of these servers are synchronized with the NIST atomic clock in Colorado.

One very accurate method I've seen used in far-remote locations is a module that received GPS satellite signals, which embed an accurate timestamp.

I've messed around a little bit with the output of a serial GPS receiver and the serial port of an SLC-500, but haven't figured out how to parse the signal well enough to handle the clock. Many small PLC's might not even give you the programmatic ability to set their clocks through internal ladder code.

You're right about the syncronization with a time server - Unitronic's Vision series has a function block specifically to do this if you're connected to Ethernet and can get to one. Unfortunately, most of my customers don't have this luxury and they don't want to pop for a GPS and then figure out how to interface with it. I can't count the hours I've spent making "simple" serial communications work.

I'm curious as to the actual method the PLC manufacturers use to make a real time clock- do they use an internal oscillator on a counter register or do they use a dedicated clock chip?

Joe,

Ya gotta know... ALL Computer Clocks are based on an Internal Oscillator.

Those oscillators have the inherent errors that they have... there's nothing to be done about that... at least, not in terms of the oscillator.

However, the oscillators control the current count in (content of) the Clock-Chip (Clock-Memory).

In most cases, the content of that Clock-Chip (Clock-Memory) is accessible... that is, writable!

If you really, really, really, need the accuracy that you indicated, then you would do very well to acquire a Time-Signal from the National Time Standard in Denver.

Every so often, you could read the signal from Denver and make the necessary correction to your on-board clock-chip (clock-memory).

You could do this on a weekly, daily, hourly, or minute-by-minute basis.

You just have to acquire the signal from Denver. There are recievers (H/W and S/W) available just for this purpose.

How can a $10 watch keep better time than every PLC and PC I've come across?????

1756hp-gps

1756HP-GPS 1756 Hiprom GPS Module
​

This module can be used in the Clx platform to sync the RTC from the GPS system no worry about time drift!!!!!

An old joke

"If you have a watch, you'll always know what time it is.
If you have two watches, you'll never be sure."

Terry Woods said:

Joe,

Ya gotta know... ALL Computer Clocks are based on an Internal Oscillator.

Click to expand...

Yeah, I know. The question is how much care is put into the oscillator design. This is what I'm talking about-

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/4984

This is a $2.35 clock chip. What magical smoke is used to make this work that can't be incorporated into a PLC design?

rPraveenkum said:

1756HP-GPS 1756 Hiprom GPS Module
​

This module can be used in the Clx platform to sync the RTC from the GPS system no worry about time drift!!!!!

Click to expand...

This module looks pretty cool, but I think it's overkill to keep your clock right. Plus you're requiring a Clx platform, a club I haven't paid the dues for. How much is this, anyone?

I appreciate the input so far. This grasshopper has learned much.

JoeTauser said:

Yeah, I know. The question is how much care is put into the oscillator design. This is what I'm talking about-

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/4984

This is a $2.35 clock chip. What magical smoke is used to make this work that can't be incorporated into a PLC design?

Click to expand...

That oscilator is temperature compensated. Those that aren't have errors of 20-30 ppm insead of 2.5 ppm. 2.5 ppm is abount 79 seconds per year. This is supposed to be worst case.

Most people don't care and the manufacturers build for most people.