Rockwell Automation-Any documentation concerning accuracy of timers,I/O for audit

Good Morning ,

I am working on a small project for a food safety audit . We ran some tests in our ovens concerning speed and temperature to prove when bacteria is killed. Our food safety department is always trying to be prepared for an auditor questions. The question was asked to me , " How do we calibrate the timers and response time of the I/O " ? and I said we don't. The electronics are super fast , and it is completly within what your specifications . Of course they ask , " how do we know that ? ".

My stupid question is , does Rockwell Automation have any certificates that state that , example , their timers are within a certain micro-seconds , or their I/O response time is within a certain rise and fall specification ?

Have any of you been asked that , before ?

Thanks so much in advance ,

http://www.plctalk.net/qanda/showthread.php?t=90776

The above post may help you out.
Regards,

It is an interesting question that really has two aspects in the context of the audit. First is the absolute accuracy of the timer, which I presume is related to hardware/firmware in the processor. Rockwell may publish a value for limits on this as "drift" or something like that.

Possibly more relevant is how the program running in the processor examines the timer and responds to it. I would expect this to have somewhat more variance, though could be limited by use of periodic tasks and watch-dog faults.

Ultimately, if pressed, you could have external hardware such as "safety rated" timing relay that would trip if not pulsed by a PLC output at a defined rate. These relays have tight specifications and published failure rates that satisfy safety applications. Be careful about offering this option because someone might actually tell you to do it.

It's an interesting question but i/o plus timers are only one small part in an uncertain chain for regular plc. Do you have additional safety plcs? There are some design rules in food processing but I would guess that your company is in charge of writing your own practice or standard. Formulate it in such a way that gives you the least work.

What I need to do is count plates going thru an oven. We need to run at least 203 plates in 60 seconds at 345 degrees F , to kill all possble bacteria. It is going to be a very small program . I like the idea of keeping track of scan time. Do you know of a chart that shows scantime vs timers , etc.

Thanks for your input.

/soapbox On

If actual time is of concern in a PLC program, then using ANY form on ladder-logic timer is incorrect.

Use the Real Time Clock in the PLC. Even on the lowest end PLC's, the RTC has much lower drift than a timer instruction.

The Logix Platform Real Time Clocks are easily read in microseconds if you need.

/soapbox Off

Yea , I see what you mean .

I assume you have an HMI? I would just show the timer .ACC on the screen in a nice format HH:MM:SS. Give it a 'calibration test' button so it can be triggered outside running logic. Tell them to run the PLC timer against their known good stopwatch and document the results.

They can write a formal SOP document if they want, action being to contact the vendor should results fall outside of acceptable tolerance.

Move on with life...

EDIT: Don't talk about the technicalities of HMI comms and PLC scan times. They just want a document for CYA purposes should this ever get asked.

Paully's5.0 said:

I assume you have an HMI? I would just show the timer .ACC on the screen in a nice format HH:MM:SS. Give it a 'calibration test' button so it can be triggered outside running logic. Tell them to run the PLC timer against their known good stopwatch and document the results.

Click to expand...

Considering the context of the application, this is a reasonable approach. And if no HMI, you could flash a pilot light at some regular interval that could be timed.

I will only add two things to consider, depending on how picky and technically savvy your auditors are. For a "good stopwatch" it is worth the cost for something like the following, which I found by googling "NIST traceable stopwatch."

Calibrated Stopwatch

The other thing would be to conduct a Measurement System Analysis or Repeatability and Reproducibility study on your timing check. If you have stat people or process engineers at your facility, they may be able to help.

Again, it all depends on requirements of the audit process. Even though we all "know" the PLC is plenty accurate for a situation like this, sometimes people with clipboards need more assurance.

Mispeld has some good points, I personally wouldn't touch anything to do with calibration test instruments. I'd give the tester the means they are asking for and see what procedure/test equipment they come up with. It's a pretty good indicator on how serious the 'verification' actually is.

Reminds me of the time I wasted days with a customer telling me mass flow meters were not as accurate as their load cells, they were convinced the flow meter needed to be calibrated. Took a lot of wasted raw materials and me walking around their "isolated" platform, demonstrating 50lb changes on their batch tank for them to realize that they should bring a third party in. Easy to predict the ending of that one, although we were close to one of our process engineers and the customers validation engineer throwing fists. We even had theoretical tag-team partners and were about to take some bets....

Is the auditor from the FDA by any chance?