Uncertainty and Fluke equipment

[NC-Cals]

If you read Fluke's manuals about their specifications, they say that their operational specifications take into account drift, resolution, environment and calibration uncertainty. The only additional line item I put in the budget was UUT resolution. I believe I am about to get into a discussion with our auditor because I put the exact wording from Flukes manual in my budgets. I did not include calibration uncertainty because Fluke has already done so. Since Fluke calibrates my equipment, I feel the budget does not require the certificate's calibration uncertainty. They have accepted that in the past, but they say they won't accept it anymore. Has anyone else run into this argument with their accredation body?

[Hawaii596]

One issue I have with Fluke's "uncertainty" is that it isn't consistent between models, and I have to be careful. Some are at 99%, some are at 95%, some have both "absolute" and "relative" uncertainty, etc.  I remember a number of years ago getting gigged by an eagle eyed auditor because (and I think this was on the 5520A maybe??), the specs were at 99% and I treated them as k=2 (oops).

The other side of this is that since we have to calculate "Measurement Uncertainty" (which is more than the "uncertainty" of the standard), there are other details that need to be RSSed together any way.

So when dealing with a Fluke spec in my uncertainty budget, I now pay close attention:  What % confidence is their spec?  Is it absolute or relative uncertainty? If I use their Relative Uncertainty spec, I need to add cal uncertainty from the cal cert for that standard. Note that if they don't state whether it is relative or absolute uncertainty, I include MU from the Cal Cert. What is the applicable temperature range of their spec? What is the cal interval of their spec? I nitpick my way through a lot of their ever changing details when building a budget.

And I make sure my budget has the big six (as applicable):  Spec of the Standard, Resolution of the Standard, Resolution of the T.I., MU from Standard Cal Cert, Repeatability, Temp Coefficient (only if applicable).

[NC-Cals]

Wouldn't treating the K3 spec as K2 spec be more conservative? You wouldn't be over-stating the uncertainty. And since Fluke claims their accuracy spec already contains the traceability uncertainty of the calibration, wouldn't it be redundant to include it in your budget. Again, you'd be expanding your uncertainty larger than it needs to be.

[Hawaii596]

Still an audit hit when the assessor digs into the budgets. I've had them review spec details and compare my calculations to what's in the manual.  If they are significantly different, they want an answer.  Some of the assessors have certain things memorized (like when a 5520A is k=3 for example). I think as long as I show that I am following the big 6 contributors, using the correct math algorithms, never have the MU on a cal come out smaller than CMC, I am pretty much ok. 

I agree that if I know that answer and tell them I am leaving it based on a k=3 spec and estimating more conservatively, they will likely leave me alone.

[NC-Cals]

I found an interesting paper on oscilloscope calibration uncertainty. There is a lot more to take into account apparently. I don't get paid enough for this level of analysis.
https://www.vsl.nl/sites/default/files/rtf/TCEM-1022_Euramet_Oscilloscope_Guideline_draft1.pdf