If an instrument is OOC from unknown days before and analyst could not encounter the error as it is much internal and critical to get,what to do with the samples that run on the instrument from some before days?
How one can get the time from when it was out of calibration?

The foolproof method to correct the situation, is :
1)Containment action:
a)Stop production untill you get a correct instrument
b) inspect all the parts available in your custody.
c)Send an express message to your customer to hold all parts available with him
d)Rush to your customer premises and segregate those components.

2)Corrective action:
a)Calibrate the instrument in question and start production.

3)Prevention:
a)Calibrate more frequently
b)Check the instrument before each time it is used as we do in case of hardness testers and micrometers.
c)Introduce a 'Poka Yoke' if possible.

I HOPE THIS IS ENOUGH!.
V.J.Brahmaiah:agree:

You may never know exactly your intrument went out of calibration. What you must do is move on from this point. I agree with bramiah's procedure. It should take care of your situation. One of the biggest mistakes we know is not letting the customer know what is going on.

If an instrument is OOC from unknown days before and analyst could not encounter the error as it is much internal and critical to get,what to do with the samples that run on the instrument from some before days?
How one can get the time from when it was out of calibration?


In our organization, we developed a calendar for internal an external calibrations. It lists all measurement equipment used by production and its calibration frequency (weekly, monthly, quarterly, etc.) Also, each time an equipment is calibrated, we use a label with the calibration date and due date. It has worked for us so far. :agree1:

Just make sure the equipment listed on the calibration calendar is up to date and try not to run out of labels!:notme:

............ How one can get the time from when it was out of calibration?

What are the 'Risk' involved?

If the risk (or consequential loss or damage) is high, you may want to validate your instrument every day or even every shift. If it is low, you may choose to do it once a week/month or so. In this way, you may be able to contain problems - otherwise, you can't tell by your present practice.

The foolproof method to correct the situation, is :
1)Containment action:
a)Stop production untill you get a correct instrument
b) inspect all the parts available in your custody.
c)Send an express message to your customer to hold all parts available with him
d)Rush to your customer premises and segregate those components.

2)Corrective action:
a)Calibrate the instrument in question and start production.

3)Prevention:
a)Calibrate more frequently
b)Check the instrument before each time it is used as we do in case of hardness testers and micrometers.
c)Introduce a 'Poka Yoke' if possible.

I HOPE THIS IS ENOUGH!.
V.J.Brahmaiah:agree:


So after this, what about the reports released before? how to correct them?

You may never know exactly your intrument went out of calibration. What you must do is move on from this point. I agree with bramiah's procedure. It should take care of your situation. One of the biggest mistakes we know is not letting the customer know what is going on.


I agree to you but we have to inform to customers at least when we have failed any batch or we have passed the batch on threshold values

So after this, what about the reports released before? how to correct them?

You may release a revision to all related reports stating 'The inspection report No.XXX dt.DD MM YY is invalid due to use of defective measuring instrument.
-VJB

The foolproof method to correct the situation, is :
1)Containment action:
a)Stop production untill you get a correct instrument
b) inspect all the parts available in your custody.
c)Send an express message to your customer to hold all parts available with him
d)Rush to your customer premises and segregate those components.

2)Corrective action:
a)Calibrate the instrument in question and start production.

3)Prevention:
a)Calibrate more frequently
b)Check the instrument before each time it is used as we do in case of hardness testers and micrometers.
c)Introduce a 'Poka Yoke' if possible.

I HOPE THIS IS ENOUGH!.
V.J.Brahmaiah:agree:


While this is "foolproof" it may be more than necessary. It is best to do a little analysis first. it is quite possible to have an instrument out of calibration and still be sure that all material shipped is conforming. As an example, suppose we have a characteristic of nominal 1.00 +/-.01. It is checked with an instrument with a gage tolerance of .001. When the instrument is checked at calibration, it is found to be at .0015 above nominal, out of calibration. A review of inspection records shows that the results recorded for units checked with the instrument are all in the range 0.995-1.004. If we allow for the inaccuracy of the gage, the range would have been 0.9965-1.0055, well within spec limits. The gage should of course be corrected, and steps taken on the calibration frequency, but as no non-conforming product has shipped, it is NOT necessary to inform the customer.

Geoff Withnell

Check out this thread on the same subject -

http://elsmar.com/Forums/showthread.php?t=33595

While this is "foolproof" it may be more than necessary. It is best to do a little analysis first. it is quite possible to have an instrument out of calibration and still be sure that all material shipped is conforming. As an example, suppose we have a characteristic of nominal 1.00 +/-.01. It is checked with an instrument with a gage tolerance of .001. When the instrument is checked at calibration, it is found to be at .0015 above nominal, out of calibration. A review of inspection records shows that the results recorded for units checked with the instrument are all in the range 0.995-1.004. If we allow for the inaccuracy of the gage, the range would have been 0.9965-1.0055, well within spec limits. The gage should of course be corrected, and steps taken on the calibration frequency, but as no non-conforming product has shipped, it is NOT necessary to inform the customer.

Geoff Withnell

Thanks Geoff!:agree1::applause:

It's of prime importance to know how much out of specification the equipment is, and then to compare it to the product specification! All of the procedure posted could well be a total waste of time......

While this is "foolproof" it may be more than necessary. It is best to do a little analysis first. it is quite possible to have an instrument out of calibration and still be sure that all material shipped is conforming. As an example, suppose we have a characteristic of nominal 1.00 +/-.01. It is checked with an instrument with a gage tolerance of .001. When the instrument is checked at calibration, it is found to be at .0015 above nominal, out of calibration. A review of inspection records shows that the results recorded for units checked with the instrument are all in the range 0.995-1.004. If we allow for the inaccuracy of the gage, the range would have been 0.9965-1.0055, well within spec limits. The gage should of course be corrected, and steps taken on the calibration frequency, but as no non-conforming product has shipped, it is NOT necessary to inform the customer.

Geoff Withnell

To expect that the whole lot produced is with in half the allowed tolerance (as in the example} is unrealistic..If it is true then it means that the process is achieving '0' PPM or better.
v.j.brahmaiah

To expect that the whole lot produced is with in half the allowed tolerance (as in the example} is unrealistic..If it is true then it means that the process is achieving '0' PPM or better.
v.j.brahmaiah

I believe you're missing the point of Geoff's post. Please don't get lost in the technicalities of the precise measurements he used, he's trying to lay out the case where virtually none of the procedure posted is necessary - if one has data and analyzes it...........

To expect that the whole lot produced is with in half the allowed tolerance (as in the example} is unrealistic..If it is true then it means that the process is achieving '0' PPM or better.
v.j.brahmaiah

Not really germane to the original poster's question, but in a well run manufactuing process, 0 ppm is what we expect. If your process is truly capable and in control, you are not producing defects! BTW, how would you get BETTER than 0 ppm?

Geoff Withnell

Not really germane to the original poster's question, but in a well run manufactuing process, 0 ppm is what we expect. If your process is truly capable and in control, you are not producing defects! BTW, how would you get BETTER than 0 ppm?

Geoff Withnell
You are right! Better than '0' PPM is impossible.There was a slip in my expression.
V.J.Brahmaiah:agree: