Out of Tolerance Calibration Report

M

mutt68

We are trying to get a new customer and they asked if we had an out of tolerance report. If a gauge is found to be out of calibration and the parts that were shipped to them are suspect, they want a form that addresses the situation so the matter can be pursued.

I seem to be having difficulty designing one and wondered if anyone had a template.
 
G

George Weiss

This post discusses O/T reporting after a calibration. After other posts and re-reading the original post comments, I believe this post is not on topic, but on the chance that the original post description does in fact look for this information it is left.
.
The old MIL STD 45662A and forward wanted O/T reporting. In a strict Z540 calibration lab the item will be tested, and if found O/T the device will be placed on a hold shelf. The customer will recieve an O/T report. Upon recieving back from the customer an alignment go-ahead or return AS-IS the process continues. The O/T report can have many bits of information. All of the information can cause a question, so more information causes more questions. A full blown quote of repair or other defined report should be avoided.
An O/T report should report the O/T condition(s) for Z540.1 / 45662A
1. Description of test
2. Nominal value to be sourced/measured
3. Actual value sourced/measured
4. Tolorance limit
In the world of Z540.3 and ISO 17025 additionally you would add
1. Uncertainty of test
2. A description of degree of PASS* or FAIL* or Fail ("*" for in the uncertainty regions)
For a caliper, an example: All values in inches is reader simple versus a +/-tolorance spec.
This is a single O/T. The report line must be for each O/T-----(17025 only)
Description___Nominal____Actual____Spec low___Spec high__Uncertainty
Measure 1"___1.000,00"__1.005,00"__0.999,00"__1.001,00"___0.000,01"
(This is the AS-Found data only)
A 17025 test PASS* or FAIL* are classified as Unknown compliance, and 17025 test FAIL is an O/T with data.
Any measureable data is an O/T. If no data can be reported the test line becomes a Failure.
A test failure should generate a failure report without any O/T data. A minimal description of failure is listed.
.
The number of O/T can be great, to the point of limiting reporting to a
process system limit of 10 for example. (this is not defined in a standard, but a real world limitation)
Addition descriptions should be limited or absent at best for O/T reports.
Complete Failure reporting is another basket of issues to report.
It is an improper practice to claim a complete failure, when an O/T is present.
A failure is likely not to impact the process because the device could not be used.
An O/T can cause serious problems.
There is the condition/situation where both function/feature failures occur, and other parts of the DUT have O/T conditions.
This situation is a 1-of-100 in the O/T or Failure situation. It is many times not handled properly.
.
Another method of operation is a general O/T notice with limited description/reporting
The entire O/T can be applied to the calibration CERT in the remarks section.
The item was recieved/returned O/T or recieved O/T and adjusted, and returned In Tolorance.
In this case the O/T data can be reported for clarity. This is a highly recomended practice for the CERT.
O/T reporting seems to float about and get lost/misplaced.
The CERTs are protected by the customer and service lab, due to them being the prooof of service etc.
AN O/T report would be defined as having:
1. Customer info
2. Test Lab info
3. DUT info
4. Test date
5. O/T report lines, having one for each O/T
6. Asking for OK to repair/align or return AS-IS is a smart remark line.
(hold DUT until response is given, unless you have customer policy agreement)
Standards and method of testing is not required.
The format comes very close to the content of a Calibration Test CERT
Calibration evolution:
MIL-C-45662=>MIL-STD-45662=>MIL-STD=45662A=>ISO-10012&ANSI/NCSL-Z540.1=>ANSI/NCSL-Z540.3,ISO-17025
GUM from many places, but originating from @ JCGM, BIPM and others n France I believe.
.
.......I just realized the O/T report you are are asking about is not one issued after a calibration, but after finding a gage in an in-house process being O/T, and possibly causing O/T product being shipped. The O/T of a gage would cause an Impact analysis and report. If the O/T was believed to impact in-house product(s) or customer's product(s) as a result, then reporting of the event is required, and then also CAPA and other action(s). The following posts reflect those views
Hope this helps...............
 
Last edited by a moderator:
S

Sorin

We are trying to get a new customer and they asked if we had an out of tolerance report. If a gauge is found to be out of calibration and the parts that were shipped to them are suspect, they want a form that addresses the situation so the matter can be pursued.

I seem to be having difficulty designing one and wondered if anyone had a template.

You need a form that will address (in no definite order):

Immediate Action
Containment
Problem definition
Analysis
Solution
Assessment
Recall

Also:

http://www.gmpsop.com/sample/LAB-010_sample.pdf
 

Wes Bucey

Prophet of Profit
I was once asked for a copy of OUR form for that topic. My response was:
It's never happened. We have no such form. We have processes in place to prevent non-conforming material from EVER reaching a customer, including calibration/checking of instruments used to measure critical dimensions before and after every shift. I pledge, however, that if such an event ever occurs [that we discover an out-of-calibration instrument AFTER a shipment], I will call you personally and discuss how the situation should be handled.
 
G

George Weiss

O/T, (Out of Tolorance), is a common occurance in calibration service industry.
The customer, who has their equipment calibrated would recieve an O/T notice, as I listed previously.
It is generally the case that equipment used in a manufacturing environment is calibrated/serviced by an outside organization.
The general calibration cycle is 1 year.
A device could potentially be O/T for a majority of a year, and effect all product during that time.
Other controls in the process can sometimes reduce the impact of such an O/T.
If the O/T is minor, (100%-125% of specification), then impact can generally be considered minor.
A significant O/T would be when tolorance is exceeded 200%+ (meaning a +/-1% spec. device reads >2% high or low)
A customer's impact can not be remotely evluated, but an example would be:
A system with a gage having >4:1 TUR / TAR will now have a >2:1 TUR / TAR,
and this increases the chance of false acceptance/rejection of effected product.
The exact analysis and remedial actions varies and is not pretty.
This is a wound which get just get uglier, and uglier, and uglier.
The customer who recieves your O/T form after you recieved it, now has to judge if he has to submit a O/T report or recall.
I have found this to be an area where all partes wish it would disapear.
 
Last edited by a moderator:

Wes Bucey

Prophet of Profit
O/T, (Out of Tolorance), is a common occurance in calibration service industry.
The customer, who has their equipment calibrated would recieve an O/T notice, as I listed previously.
It is generally the case that equipment used in a manufacturing environment is calibrated/serviced by an outside organization.
The general calibration cycle is 1 year.
A device could potentially be O/T for a majority of a year, and effect all product during that time.
Other controls in the process can sometimes reduce the impact of such an O/T.
If the O/T is minor, (100%-125% of specification), then impact can generally be considered minor.
A significant O/T would be when tolorance is exceeded 200%+ (meaning a +/-1% spec. device reads >2% high or low)
A customer's impact can not be remotely evluated, but an example would be:
A system with a gage having >4:1 TUR / TAR will now have a >2:1 TUR / TAR,
and this increases the chance of false acceptance/rejection of effected product.
The exact analysis and remedial actions varies and is not pretty.
This is a wound which get just uglier and uglier.
I have found this to be an area where all partes wish it would disapear, and to many times it does. (quietly)
(a personal observation of occurance)
Gee, George, in over 40 years of running operations, I never had it happen when I was a supplier and can only recall one instance where it occurred with one of our suppliers and we caught the N/C before we installed it in any of our products and we, in fact, were the ones who alerted our supplier there might be a problem with his instrumentation.

If a customer does a proper job of investigating suppliers prior to placing an order, it has been my experience that part of that investigation is looking for processes that PREVENT errors and nonconformances, not processes that DETECT them after they've occurred.
 
J

Jeff Frost

Don't create a special form for out of tolerance gages. The method I have been using for years is to have the person responsible for calibration document the out of tolerance on a nonconformity report which is sent, along with a copy of the calibration report, to MRB for review and investigation. MRB investigates possible effect on WIP and delivered product as required by Clause 7.6 and then documents their actions on the nonconformity report and takes any needed action in-house. A copy of the nonconformity report with disposition is then returned to the person responsible for calibration so that it can be added to the history file.

If the out of tolerance condition did have an effect on delivered product the customer is notified by written letter (or email), containing all relevant information by the Quality Manager.
 
G

George Weiss

.....I have had to reject many an instrument for being Out-of-tolorance. The equipment would come from every walk of life
and organization. Claiming 100% coverage, and lack of O/T, is a gamble with future reality for most organizations. In some
cases the daily check of a caliper with a gage block set can cover an inspection tool, (as in your case Wes), but a large
number of facilities have significant numbers of complex equipment standards, which do have qualities critical to a process
test specification. Because of the extent of some equipment testing, and the numbers of specifications of some equipment,
it is possible to find O/T with no real impact on product or process. This why an O/T analysis has to be performed.
The degree of O/T and the specific parameter are key. A company with an environmental chamber which has Temp.
R/H and timing specifications, may have a Ramp Time O/T which is not a factor in their process, and so has no impact.
So, in the end I am glad you, Wes, have not had to deal with the problem of in-use equipment being O/T, but it is a very
real and common problem.
 

BradM

Leader
Admin
We utilize an out of tolerance report, which encompasses most of the items George listed earlier.

In short, it alerts the customer (internal customers) that their item failed calibration.

We also add a list of history for the instrument, going back three times (if we have the history for it).

We attach a copy of the calibration performed. We then work with the customer regarding any remedial action necessary, and have the customer sign the failure notice. This way... the calibration department is not responsible for the remedial action. Which would be difficult anyway, as we have little knowledge of all the processes which they are used. :2cents:
 

BradM

Leader
Admin
Gee, George, in over 40 years of running operations, I never had it happen when I was a supplier and can only recall one instance where it occurred with one of our suppliers and we caught the N/C before we installed it in any of our products and we, in fact, were the ones who alerted our supplier there might be a problem with his instrumentation.

If a customer does a proper job of investigating suppliers prior to placing an order, it has been my experience that part of that investigation is looking for processes that PREVENT errors and nonconformances, not processes that DETECT them after they've occurred.

It's fairly common with what I've seen, to have calibration failures. When you are calibrating a bunch of process equipment, it's totally out of your control how they handle the equipment, what they purchased (many times), etc. Thus, there are going to be failures.

I do agree with your logic when it comes to standards within a lab. If you perform cross-verifications, practice redundancy, and keep close tabs on the cross-verification results, you can significantly limit any impact from an out of tolerance for your standards.

It still happens though. I have received a few from calibration vendors outlining exactly where it would out of tolerance, and how we might be affected.
 
Top Bottom