Manufacturer Accuracy Specs
I recall a post many years ago about the accuracy specs from instrument manufacturers being reported to one standard deviation. In looking for the standards they follow, some do not even report that they are specifically following any published standards for developing their accuracy report and some state they are following the ISO 17123 protocols.
1) Does anyone have the ISO protocols so that you could confirm the language showing the reporting as being at one standard deviation?
2) If the manufacturer does not specifically call out the ISO procedures, does this mean they are not using them or is it to be assumed that they are?
I too find it odd that there's not more official documentation on this floating around.
What I have heard, unofficially, is that the DIN and ISO standards evaluate the same thing but with different methodologies.
That is, they are evaluating the standard error (one standard deviation) of an observed direction - meaning the average of one F1 and one F2 pointing at the same target.
Jon Payne; It would be interesting to know why you have these 2 questions(?).
Your #2 question is never to assume anything. If the manufacture does not state what he is using then ask him.
Your #1 question; ISO is a set of field procedures to use to check what the manufacture is saying about how accurate his instrument is at some confidence level. ISO shows you how to calculate "s = experimental standard deviation" and then does that pass a statistical test at a "confidence level" of 95%.
Does that help??
@john-nolton Thanks John.
Q1 - That helps. I understand that the ISO provides a procedure for testing the instruments and determining a value for s then seeing if it passes 95% confidence interval.
Some of the datasheets/manuals have stated RMS, some have just referred to ISO sections. Are there any statements in the specifications that indicate when using ISO as the reference for reporting an instrument accuracy, it will be reported as the calculated s value (as opposed to the 1.96s) or is that something that would need to be verified with each manufacturer if they haven't stated anything more such as RMS?
Q2 - Yes.. Never assume, but after finding reference to ISO standards hidden in various areas in fine fine print in different locations for different manufacturers, I was getting a little perturbed at hunting for what seems like should be front and center in advertising measuring equipment.
Reason for the Qs - I would like to have a discussion at out state society chapter meetings about the language in our statutes for Relative Positional Accuracy as a means for those in attendance to gain an hour CPD credit for attending the meetings. That might entice a few more attendees and would give current attendees additional value beyond the fellowship and a good meal. To claim CPD credit the information needs to be a formal presentation, not just a discussion amongst colleagues. I would envision this being a discussion over several meetings going over a variety of sub-areas of the main topic (i.e. where the language for our statute came from, error sources, degrees of freedom/redundancy, normal distributions and confidence intervals, etc...). I figured a discussion about instrument specs and methodology would be valuable, but don't want to look even more foolish than usual by not having done some homework first.
@jon-payne Thank you for your nice response. I sure hope all goes well with your class.
You ask a question "Are there any statements in the specifications that when using ISO as the reference for reporting an instrument accuracy"
Answer: probably just the opposite. In ISO 17123-4 says the following in the introduction;
"This part of ISO 17123 specifies field procedures for adoption when determining and evaluating the uncertainty
of measurement results obtained by geodetic instruments and their ancillary equipment, when used in building
and surveying measuring tasks. Primarily, these tests are intended to be field verifications of suitability of a
particular instrument for the immediate task. They are not proposed as test for acceptance or performance
evaluations that are more comprehensive in nature."
The manufacture can say something about how good his instrument accuracy is, but the USER should always check to
see if it is so.
Web site for ISO = www.iso.ch
Many times I've heard the instrument accuracy is advertised at 1 SD, so I was curious if that was just what manufacturers decided to do, if it was observed by practitioners based on years of experience with a variety of instruments, or if it was actually a reporting requirement from a set of standards such as ISO (i.e contained in ISO/IEC Guide 98-3:2008 Uncertainty of measurement — Part 3: Guide to the expression of uncertainty in measurement (GUM:1995) Section 7 on reporting uncertainty).
As you say, always best to check for yourself anyway.
I think given the link to the message thread Bill provided and your input here I'll have enough to work from in a general discussion.
Reason for the Qs - I would like to have a discussion at out state society chapter meetings about the language in our statutes for Relative Positional Accuracy as a means for those in attendance to gain an hour CPD credit for attending the meetings. That might entice a few more attendees and would give current attendees additional value beyond the fellowship and a good meal.
I would love to attend that series of your chapter meetings but I don't think I can pull off a 10 hour roundtrip and a workday. Keep up the good fight. Love to see state society endeavors.