Organizing Uncertainty Budgets for different Measurement Equipment

[ScottBP]

I know there's plenty of uncertainty calculator programs and spreadsheets out there, but when it comes to things with many, many individual uncertainty analyses (like gage block sets or multifunction calibrators), it usually means making umpteen dozen (or even hundreds) of copies of the spreadsheet (one for each reading), then saving them along with other associated calculations (e.g. control charts for Type A repeatability analysis) somewhere.

I know that some software (such as Fluke MET/CAL) can calculate uncertainties automatically as an instrument is being calibrated, but what I'm talking about is linking together the uncertainty budgets, control charts, etc. for the parameters that we list on our 17025 scope of accreditation, in some sort of logical fashion that can be easily understood.

I've been doing it on spreadsheets and saving them to a networked drive, but as the company I work for grows and has more equipment added to our capabilities list, this task has become overwhelming. To top it off, now we have auditors from regulated industries (e.g. nuclear) telling us all those spreadsheets have to be validated, approved, revisions counted and archived, etc., but none of them tell us how to go about doing it. So my question to everybody is how do you do it? How do you keep up with everything?

 

[Marc]

A quick "Bump". My Thanks in advance to anyone who can help with this one.

 

[Hershal]

Some good tough questions there.

First, the validation. Some of the regulators may require all kinds of validation that an AB may not. For example, using Excel and its built-in capabilities is often considered validated and the only thing you need to do is validate that specific application. So, an uncertainty budget done in Excel that is sufficiently generic would only need to be validated once.

As for the number, I suspect a better move for control is to match your uncertainty studies to the accredited scope assuming you are a test lab. Gage blocks for example, may then need only 4 or 5 uncertainties, rather than 81.

Of course, if you are a cal lab then you may be stuck, as you have to provide uncertainties to your customers for calibrations performed.

 

[ScottBP]

One auditor told us to simply work out the calculations on paper with a calculator and see if the results agree... Really? :mg: Instead I managed to distill it down to a basic budget with the minimum contributors, and then I'll do what I found out one of our consultants did- I'll take a screenshot of the most critical calculations showing in the formula bar. If the auditor doesn't realize the amount of scruitiny Excel has gotten over the years, then maybe we need a new auditor.

As far as keeping track of changes, I've heard of people using something like Microsoft Sharepoint or some other collaboration software that has built in revision control, but how would I set something like that up? The devil is in the details, and believe me, I'm on a first name basis with the devil right about now...