Under ISO 9001 when an operator is checking a part on a gage that has a variable data reading ( dial indicator) is it exceptible to write pass/fail on a log or do you have to write the actual reading or in some cases the ± reading ? We currently have them recording the reading because someone decided long ago we had to.

I think it depends on what you want or need to do with the information.
If you are going to do any analysis (Cpks, controls limits) and you go to the extent of actually measuring the dimension to read it then you may want to record the actual result. We have some dimensions that we measure but only record the variation from the nominal. These type records are also good for historical information if you need to see the long term effects, or provide verification. If your customer specifies actual readings rather than pass/fail then you need to do it.
However, if only want or need to see the results as good/no-good then develop a go/no-go gage to check the dimension, it is easier and quicker.
It is your decision based on need and requirements.

These are quality checks performed every 15 minutes by our operators and recorded on the back of their work instructions. We do not use this data for studies but it is used in part to fulfill 8.2.4 as well as some internal fulfillments.

I think it depends on what you want or need to do with the information.

However, if only want or need to see the results as good/no-good then develop a go/no-go gage to check the dimension, it is easier and quicker.
It is your decision based on need and requirements.


I agree, but would add, the answer applies whether you use a go/no-go gage or a variable gage.

Note: some auditors may not like the idea. I wrestled with it as well in the past. But, there is nothing in the standard that states the format you record.

I agree, but would add, the answer applies whether you use a go/no-go gage or a variable gage.

Note: some auditors may not like the idea. I wrestled with it as well in the past. But, there is nothing in the standard that states the format you record.

All go/no go checks are recorded as pass/fail, it's the variable gage readings I am unsure of. I can see where an auditor would may ask how I know if the person checking it knows how to read the dial gage but most all use zero blocks so they are checking ± tolerance. Dimensions and tolerances are on the work instructions and the operators have been trained.

All go/no go checks are recorded as pass/fail, it's the variable gage readings I am unsure of. I can see where an auditor would may ask how I know if the person checking it knows how to read the dial gage but most all use zero blocks so they are checking ± tolerance. Dimensions and tolerances are on the work instructions and the operators have been trained.


My concern was that many auditors assume that a variable gage reading must be recorded as a numerical value. There is however, no actual requirement for that.

Excellent suggestions on this one.

Apart from the pros who can definitively state how it applies to the requirements, I would state you have a large say in this matter.

The one advantage of recording the variable data is if the instrument used to make the measurements is found out of tolerance (O.O.T.) during the next calibration, it will assist in corrective action. By the magnitude of the error found during O.O.T., you can make an assessment with the variables data which measurements may have been significantly affected. If all you do is check yes/no, you may have a larger area of potential impact.

Don't know if you are into gauge R&R, but the variables data will help. Too, if you start getting different values between operators, that data may be nice.

Just a thought: You can further operationally define this. Say when inspecting, if the observed values are <25% or so of the maximum allowable deviation, a yes/no check could be utilized. If the observed values are >25% to 99% or so, you may want to record the variable value.

Too, all this may be irrelevant and discrete data would be just fine, if you are using proper standards, and performing cross-verifications, so as to minimize effect should be standard be found O.O.T.

My long-winded way to say that there are always risks/tradeoffs, and recording discrete data (pass/fail) as opposed to variables data is no exception.

Yes R&R is done on all gages and calibrations performed on a prescribed time table. Thank you all for the feedback.

I guess I have a more fundamental question, why is the dimension being measured to begin with? Is the data providing information to the operator for making decisions on the process? Or is it "report card" data - as in controls for other dimensions keep that dimension in spec, but the customer is concerned about the condition of that characteristic? And, of course, is the process a precision process, or non-precision? I agree with Brad, if the ongoing variation versus the tolerance is negligible (I would say less than 10%, about the level of the gage R&R), then pass/fail would be good. The other question is, has the dimension ever been out of spec? If not, you can likely spread the timing of the checks out, too - your risk being how many parts you would have to sort, scrap and/or rework if something did happen.