View Poll Results: How should instrumentsl errors be specified?
Maximum error 2 100.00%
Standard deviations 0 0%
Some combination of the above 0 0%
Voters: 2. You may not vote on this poll

Register to reply

Limiting deviation

by 0xDEADBEEF
Tags: deviation, limiting
Share this thread:
0xDEADBEEF
#1
Aug26-14, 01:08 PM
P: 825
Long time no post...

I am very much disturbed after my research of measurement errors and corresponding norms, because I find so few results. There are mainly two ways to specify errors of measurement instruments. Standard deviation with its siblings and maximum error.

I am a great fan of the old German tradition to specify maximum errors only for measurement instruments. Specifying that some piece of equipment will have an error below x for all practical purposes when operated under the specified conditions is much easier to do quantitatively than statistical deviations and errors that depend on the measurement value and environmental fluctuations. Because in most cases the errors are far from being Gaussian. So you just add add all relevant maximum errors and don't care about independence.

But I couldn't find any norms about it. The relevant German ones don't have corresponding ISO norms. I couldn't even find a symbol for it. The German symbol is G for "Grenze" or "Fehlergrenze". I only get "uncertainties" and a number of k sigmas.

So after so much complaining my questions:
- Is there an international norm regarding the declaration of maximum errors.
- What symbol do you use for "maximum error" or "limiting deviation"
- If you have a better name for this and what do you call it? What is it called in your Physics books?
Phys.Org News Partner Engineering news on Phys.org
ESA investigates an alternative, environmental-friendly method of corrosion resistance
The oscillator that could makeover the mechanical watch
Engineers study bats to improve aviation travel
Mech_Engineer
#2
Aug30-14, 11:55 PM
Sci Advisor
PF Gold
Mech_Engineer's Avatar
P: 2,288
Generally speaking I would say "it depends." I've usually seen the "maximum accuracy error" specification as you've mentioned, usually defined in a +/- percentage value (example here: Honeywell Model 31 Force Transducer).

When qualifying the performance of a complex instrument or process, it might be prudent to use more advanced statistical methods e.g. Measurement Systems Analysis (MSA): https://www.moresteam.com/toolbox/me...m-analysis.cfm

Gold standard MSA books: https://www.aiag.org/source/Orders/i...m?search=MSA-4


Register to reply

Related Discussions
Comparing Absolute Deviation to Mean Absolute Deviation Set Theory, Logic, Probability, Statistics 4
Difference between sample standard deviation and population standard deviation? Precalculus Mathematics Homework 4
Limiting the Torque helps limiting the RPM of Shaft Mechanical Engineering 0
Standard deviation versus absolute mean deviation Set Theory, Logic, Probability, Statistics 3
Mean Absolute Deviation/Standard Deviation Ratio Set Theory, Logic, Probability, Statistics 6