Measurement Uncertainty Forum

Should Uc Budget for Pt include UUT LTS?

James D Jenkins
Re: Should Uc Budget for Pt include UUT LTS?
by James Jenkins - Wednesday, July 1, 2009, 9:26 AM

winkUncertainty of Measurement never includes the tolerance of what is being measured. Only the contributors that make the data imperfect, such as:

  1. Traceability of Reference (Cal Uncertainty)
  2. Allowed Deviation and Drift of Ref Device (MTE Spec)
  3. Resolution(s) of data (Two when comparing two indicating devices)
  4. Process Repeatability & Operator Reproducibility
  5. Environmental influences on MTE and UUT (Typically, NA if done in adequate environment. Although you should check the specs on any electronics involved before dismissing this contributor.)

Torque wrench PT's are often failed due to failure to evaluate repeatability correctly or not properly exercising the wrench prior to taking data.

For an added measure of correctness (no pun intended), exercise the wrench 3 to 5 times, then repeat the measurement of the snap wrench at least 10 times (collecting data) and then compare the standard deviation with that in your "representative" budget. If the standard deviation is larger than that in your budget, update your budget and report. 

If your process normally take 5 readings (as recommended for torque wrench calibration) and computes the mean value to be reported as the measured value, ensure that you divide the standard deviation by the square root of the number of samples used to compute the reported mean. (If using UncertaintyToobox, put a 5 in the "Reported value = mean of X samples" entry on the Type A worksheet.)

Regarding samples; eventually we will have some torque wrench samples in the User Group, however, these are typically not complex estimates, rather using the "Basic Five" contributors, as given above, covers the significant uncertainty contributors.

If you would like your analysis reviewed for correctness, you can submit it in the User Group and I will review it, correct it if needed, and post it as an example for other users. Please note; if you do NOT want your name or lab name used in the example, remove all entries containing the names from the analysis prior to submitting, otherwise I will assume you wish to keep that info in the sample (free advertising). 

P.S. Next I will check to make sure I have an upload section set up for this in the user group, if not, it will be there by tomorrow.wink