Measurements after Calibration Expiry

[Quality27]

Hi all,

Just got into a bit of a conundrum. If I am following a periodic calibration on a Pressure Transducer and a Floor Scale and conducted some tests after the calibration has expired, What are my options to validate the tests conducted?

1. Can I calibrate now and if I do not find any error, can that be used as a fair argument to validate the tests conducted using this equipment?

2. Can I ask my calibration vendors to issue a certificate on the equipment if it is found in calibration now?

Thanks,

Tony

 

[ScottK]

You can have the calibration performed now and if all is good you can rationalize that that is was still good at the time of testing.
Poor practice but not uncommon. You should also include this eventuality in your calibration procedure.

If the equipment is out... then you have a problem. Depending on how far out it is you may have to re-test, if you can, or at worst re-do the production run.

DO NOT ask your calibration service to back date the cert even if the equipment is in calibration. Be honest with yourselves and any potential future auditors - document the situation in an NCR or CAR, document what you did to correct it, what you'll do to prevent it from happening again, and move on.

 

[BradM]

Hello Tony!!

Hi all,

Just got into a bit of a conundrum. If I am following a periodic calibration on a Pressure Transducer and a Floor Scale and conducted some tests after the calibration has expired, What are my options to validate the tests conducted?

1. Can I calibrate now and if I do not find any error, can that be used as a fair argument to validate the tests conducted using this equipment?

2. Can I ask my calibration vendors to issue a certificate on the equipment if it is found in calibration now?

Thanks,

Tony

First question... do you have a system that identifies that a deviation or something be issued for the past due? Is anything typically documented for that? If yes, then begin the deviation for the past due, and identify in the deviation write-up whether the instrument pass/ failed and any impact it may have on the product/data measured on those devices.

I guess a lot depends on how your system addresses past due events.

Add edit... Scott posted while I was typing. If your system does not dictate a deviation or whatever, I would follow exactly what he said.

 

[AndyN]

Have you looked back at past calibrations to see what the "as found" conditions are/were? It seems that, as with most calibration "systems", when you base recall/recal on time, it doesn't actually work and then you get into such a dilemma.

Have you ever done a study on the equipment to determine if its response is a) linear and b) stable? Without these, shooting for time based recalibration might give a warm and fuzzy, but may not be too effective.

 

[dwperron]

First, using an instrument after its calibration due date is considered to be using nonconforming M&TE. That should be a violation of your quality system, with corrective actions required. Your quality system should have something for authorizing temporary extensions that would keep your equipment from becoming overdue.

You always evaluate your calibration results to determine that the results are acceptable. If they are acceptable then you can continue with no further actions required.

If unacceptable you will always have to follow your quality system in determining the impact any out of tolerance measurements may have had on your measurement process. You should have a documented procedure on how to handle these cases.

 

[Quality27]

You can have the calibration performed now and if all is good you can rationalize that that is was still good at the time of testing.
Poor practice but not uncommon. You should also include this eventuality in your calibration procedure.

If the equipment is out... then you have a problem. Depending on how far out it is you may have to re-test, if you can, or at worst re-do the production run.

DO NOT ask your calibration service to back date the cert even if the equipment is in calibration. Be honest with yourselves and any potential future auditors - document the situation in an NCR or CAR, document what you did to correct it, what you'll do to prevent it from happening again, and move on.

Thank you Scott. Poor practice was my first idea but usually when I submit a test report to customers, I ask my QC Techs to include calibration certs as a good practice.

I do not see any specs from the customer to include calibration certs but could there be instances where customer could ask for the calibration certs? Or for example, if I have to put together a AS9102 FAI, do I have to include these?

Typically we outsource calibrations, but in this scenario, can we certify internally after we get it certified from our vendor? Just feels too hokey to me.

 

[Ninja]

If the equipment is out... then you have a problem. Depending on how far out it is you may have to re-test, if you can, or at worst re-do the production run.

Just a note following on this...it's the same thing that you should be doing if you go to recalibrate/verify on time, but find that the gage is out.

Calibration timing is just a date on a calendar...equipment can go out at any time.
The point of periodic calibration check is to catch problems quickly, or prevent them...but the world isn't perfect...and sometimes a gage goes wonky early...

 

[AndyN]

Just a note following on this...it's the same thing that you should be doing if you go to recalibrate/verify on time, but find that the gage is out.

Calibration timing is just a date on a calendar...equipment can go out at any time.
The point of periodic calibration check is to catch problems quickly, or prevent them...but the world isn't perfect...and sometimes a gage goes wonky early...

In which case, you aren't managing the measurement resource effectively! The suitability of equipment for measuring and inspection shouldn't be carried out in isolation of the needs of the organization - which is exactly what happens in 90+% of cases. Recalling ITME annually isn't much good, if the plant's so busy they put on a second shift! However, I've yet to see any organization have that kind of intelligence in their controls. No-one changes their oil in the car annually (unless you're my ex-mother-in-law) or even 6 monthly. It's (should be) done on use and the conditions under which the oil operates.

 

[Ninja]

In which case, you aren't managing the measurement resource effectively!

Well...I agree with you about 93.68%...

Sometimes stuff just breaks without showing off. This is earth after all.

I have a CMM, it's been a workhorse for 8 years, never even thought of failing validations or calibrations...I moved it from 3mnths to 6mnths...3 years later I moved it to a year...then I randomly checked it 2mnths after a cal with no issue...checked it just 'cause I was training someone else how to do the verify and cal...

The stupid thing was WAY out, and we had to recheck 2mnths worth of work.

As Mr. Gump said, "It happens". A rarity...but it happens...

 

[AndyN]

A rarity...but it happens...

Which, in and of itself, no justification for treating EVERYTHING the same way.