Calibration Program Creation vs Manufacturer supplied

Qual_Eng_17

Registered
Hello All,

I am a little out of my knowledge area on this one but ill get to the point. We have a new piece of Inspection Equipment (Cognex) which is going to be measuring product. In past practice rather than using the System Calibration Program, which is supplied by Keyence, we create a program to measure a know artifact that is held to a tolerance significantly tighter (1:10) than what the system is intended for.

With this new equipment it came with a Calibration Artifact and a Pre-Build Program, but the engineers involved in this would rather create their own program to measure a similar artifact to that of above as it is normal practice here.

Is this acceptable in Med Device to do this type and justify by measuring something with a significantly tighter tolerance and repeating or should we be using the manufacturer pre-built programming?
 

Jen Kirley

Quality and Auditing Expert
Leader
Admin
I don't use Cognex, and I don't know what kind of program you are referring to for calibration.

I am nervous about what you are proposing though. I found a Cognex page about calibrating its equipment (does it address the the type of equipment you use?), and I know that 13485 wants us to be very careful with calibration and our software to be validated. I don't know if you are 13485 registered, but even if you aren't, following its requirements seems like a good defensive measure to me.

In case those sites aren't what you need, I found an assortment of others you can review.
 

Ron Rompen

Trusted Information Resource
reading the initial post, it appears to me that you are using your in-house artifact to VALIDATE the accuracy of the Cognex equipment, rather than to perform a calibration.
i like to do the same thing - this also allows me to perform cross-machine comparisons.

IF, however, you ARE using this artifact to perform calibrations (using the built-in software functions) then I would be very careful in doing so - machine programming is easy to change, and just as easy to screw up - miss editing one line of code (or even one variable within a line of code) and your results can be screwed up just enough to cause you a lot of grief.
 

Qual_Eng_17

Registered
Yes we are, looking at ISO 13485 it only talks about being tracible to a international standard. Which to a counter part is a NIST Certified Gage Artifact. However this is not the manufacturers recommended calibration method, they provided a NIST Artifact that uses a machine specific program for calibration.

Please tell me if I am completely "Off my rocker" here, i do have some un-easiness here about not using the manufacturer supplied program and standard but if we are meeting the requirement for it then i could be a little more at ease here.
 

Jen Kirley

Quality and Auditing Expert
Leader
Admin
In addition to the calibration clause, we need to include clause 4.1.6, which calls out for validating software used in the QMS.
 

Ron Rompen

Trusted Information Resource
Qual_Eng: the requirement for traceability can be satisfied by having the artifact (either the supplier artifact or your own in-house artifact) measured on equipment (other than the Cognex of course) which has sufficient accuracy and precision, and which has been calibrated traceable to NIST.
You are not 'off your rocker' to be concerned.
before making the changeover to the 'new' artifact method, do a correlation study between the two artifacts.....do it several times, and compare the results. if you notice any significant discrepancy (correlation <90%) then you need to reassess the plan, investigate WHY the results are different, and determine what actions are to be taken.
 
Top Bottom