SLC 5/03 Clock Losing Time

I have an application that requires the correct time controlling of some outputs, but during testing have discovered the SLC 5/03's clock looses a few seconds each day (unacceptable for this application). To compensate I added a line adding a few seconds just after Midnight hoping this will be close enough, but was wondering:

Is there a tweak to the clock speed on a SLC?

Is there an easy way for the SLC to read the real time on the PC based HMI at regular intervals? (Using DF-1 serial comm's) (1747-L532C 5/03 OS302)

Well brousing through the forums I stumbled across this:

http://www.plctalk.net/qanda/showthread.php?t=45409

In this tread Ken describes what you may be looking for.

Good luck
BCS

You're looking at ordinary oscillator drift; there's really no way to make the controller's clock any more accurate than it already is.

And hey.. I turn my back and get leapfrogged. How are things up at Dutch Harbor ?

The Clock Sync utility is the standard method for getting a good consistent time into A-B controllers.

I'd love to have a cheap serial-connected long-wave receiver from the atomic clock in Colorado, but I've been unable to find anything in the market that's appropriate and cost-effective.

And hey.. I turn my back and get leapfrogged. How are things up at Dutch Harbor ?

Click to expand...

Sorry about that did not even see that you were online.

Staying busy here, engine troubles, Well we will say they are fixed, we are at sea making money anyway, hopefully it will stay that way.

But back to the OP, so you say that the drift in time just is what it is, and will have to be manually adjusted from time to time?

BCS

Bering C Sparky said:

Well brousing through the forums I stumbled across this:

http://www.plctalk.net/qanda/showthread.php?t=45409

In this tread Ken describes what you may be looking for.

Good luck
BCS

Click to expand...

Sparky,

Thanks, that explains - but the Rockwell page linked to has been removed. I did a search & came up with "CIP Sync" but it's only for Ethernet comm's, nothing showed for DF-1 serial comm's.

Here's a screenshot of what I did to compensate (the adjustments are set to NULL so I can test the clock to see how much time I have to add daily & how much I have to add every xx days)

the drift in time just is what it is

Click to expand...

Yup.

Think of the CPU clock ticking along at 8 MHz or so. Some seconds it's right at 8 million cycles, some it's 8 million and ten, some it's 7 million nine hundred ninety. The clock design is from the mid 80's and isn't built for super precision, and it drifts a little depending on temperature.

So if your application needs a wall clock time that's accurate, you need to periodically synchronize it with a known-good or higher-precision clock.

Ordinary PCs have this sort of drift as well, but we don't usually notice because most of them are connected to a server or to the Internet and can get time updates from the big NIST servers daily.

well, now I know.

Thanks for the info.

BCS

It looks like the current version of the Logix 5000 Clock Update Tool is attached to RA Knowledgebase Article ID # 44321 (Access Level = Everyone).

There's an even newer Studio 5000 Clock Sync Service tool, but for SLC-500 controllers you should stick with the ordinary Clock Update Tool. It ought to work over any communication network.

(Using DF-1 serial comm's)

Click to expand...

I'd love to have a cheap serial-connected long-wave receiver from the atomic clock in Colorado, but I've been unable to find anything in the market that's appropriate and cost-effective.

Click to expand...

Well, I was not paying attention, this one went right over my head until I read it again later.

Dont I feel STUPID.

BCS

Ken Roach said:

but for SLC-500 controllers you should stick with the ordinary Clock Update Tool.

Click to expand...

Thanks,

I downloaded that & will try it - the accompanying PDF seems to show it will work with any RSLinx communication. Didn't look at it before because it's called "Studio5000 Clock Update" and Studio5000 stuff doesn't generally work with SLC's.

Well an option is GPS if you need high accuracy, GPS time is adjusted to the atomic clocks twice a day, you can import that data in to adjust the clock...I think you can atleast

I don't think it will update the clock directly but you can probably make a comparison of the numbers and adjust

What I did for a similar situation was buy a cheap, but accurate, digital appliance timer from Kmart & program it to turn on every 6:30 (AM & PM) for 1 minute.

When the time came on One Shot (twice a day) I moved 30 to minutes & 00 to seconds. I didn't have it come on at xx:00:00 because then I would have to figure out if the PLC was slow or not & have to change the hour up 1 if so.

If needed for more accuracy, the timer has 8 on times & can then be programmed to come on every 3 hours daily, or every hour if only for one shift.

I only check the time on the timer every few months & adjust it if needed, but it's been pretty reliable.

Plus with changing the minutes only it doesn't matter if it's EST or DST, and won't mess with that setting.

In North America the frequency (60 Hz) is maintained fairly closely and the utilities adjust as necessary for the proper number of cycles to within a few hundred parts per million (Wikipedia Reference) so a simple clock getting its reference from that would be fairly accurate in the long run, though there are proposals to allow that accuracy to slip.

Just to let everyone know:

I am using PeakHMI on this project that has it's own DF-1 driver & RSLinx can NOT be running it's DF-1 driver for the Clock Sync - so I ended up using Secpcb's idea of an appliance timer.

The timer I bought only had one ON time per day so I set it for 6:30AM for the PLC to be most accurate during the day. In the ladder I added a trap to see if it came on, and log the time after it came on to make sure it was 6:30:00 exactly.

When I checked it about 8:45AM the trap was set, it logged 6:30:00 correctly, but in the last about 2 hours the SLC5/03 had already lost 5 1/2 minutes. Throughout the day I monitored it & the SLC could gain or lose up to 4 minutes per hour. Now I'm going to get a timer like Secpcb said with 8 ON times & set it for every 3 hours.

I also had a project I was writing for a Micrologix that didn't have online editing - so I wrote & tested it on a SLC 5/05 - the timing of the cycles was exactly what was needed after final tweaking, so I copied it line for line to a ML project & downloaded it. To my surprise the cycle times were running fast - it seems a 8.00 second timer on a SLC and one on a ML are different times - the ML ran timers considerably faster, I had to calculate the difference & change all the timer presets to compensate.

Question: Does anyone know of a PLC with an accurate clock & timers for the next project where I need precise times?

I don't know if it would work for you but, several years ago somebody (Peter Nachtwey, I think) posted some ladder logic for a long term accurate clock.

The principle was to accumulate counts from a timer then at intervals increment the clock based on the timer and reset the timer. Counts above or below the interval were added/subtracted to account for drift.

I've searched for that post but couldn't find it.