Metrology Best Practices

[nonobaddog]

In my years as a metrologist, it has been my practice to report dimensions to 1 place further than the specs given on the print that I'm working with. I know that I have seen this practice in writing as a "best" practice. My boss, who does not understand metrology, wants me to start reporting to the number of decimal places shown on the print. Can anybody provide a link to, or a direct quote from a source specifying this "best" practice?

 

[drawoh]

nonobaddog,

I am not aware of a procedure that defines all of this. If it were me inspecting, I would report the number of decimals I had confidence in.

If you are doing metric drawings to ASME Y14.5M-1994, the trailing zeros are supposed to be deleted from the drawing dimensions, making your problem moot.

JHG

 

[nonobaddog]

I know I've seen it in writing somewhere... and it's not so much a procedure or requirement as it was a documented "best practices."

Thanks.

 

[KENAT]

It makes sense I suppose. I can't think of a reference either though. How about putting the 'official' result to the same sig fig/decimals as appropriate and a () for the actual measured value?

KENAT, probably the least qualified checker you'll ever meet...

 

[MetrologyHam]

It is often the practice of Calibration/Metrology professionals to report data to the last digit, however, it is not always the best method. Writing extranious digits leads to transcription errors. It is customary to report the Standards indication to one decade greater than the item being calibrated.
However, that is not a hard rule. What you really need to concentrate on is the reporting of SIGNIFICANT digits per ASTM E 29 – 02. This means that you would "Round-up" data that is over 5 in the following decade. SUch as 2.3505 would be rounded-up to 2.351.

 

[tomwalz]

From the manufacturing side (our version) we would classify this under SPC (or SPQC) for statistical process quality control. If a tool is wearing we want to know before it goes out of spec.

Not sure if you use it but extra digits beyond what is specified are key to the whole process.

E.G. 2.3505 then 2.3504 then 2.3503 might indicate wear in a drill bit.

EG. # 2 – a set of solid 2.3505’s would be good. If you see 2.3505, 2.3504.2.3506 it might indicate collet wear.

Tom


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[qualitygeek]

I have found that in general practice that what ever you are checking you should use anything that will produce a result that has an accuracy of 1/10th of what your are trying to check, when ever possible.

 

[macgyver247]

quality geek

I too used to use the rule of "10".
But have recently been told that a "rule of 4" ( four times more accurate ) can be applied to CMM results. Although this could lead to the measuring tool uncertainty actually using up 25% of your tolerance.

 

[bugnut]

Just happened across this thread and recalled a similar thread from a while back. I too used to use the read to 1 digit more significant and round. Was corrected to read all digits. Please see this thread for the detailed prior discussion and an applicable ANSI standard. thread286-122849
the quote is from the thread "According to ANSI Y14.5M-1994 (and 1982), Paragraph 2.4 on page 25

"All limits are absolute. Dimensional limits, regardless of the number of decimal places, are used as if they were continued with zeros.

Examples:
12.2 means 12.20...0
12.0 means 12.00...0
12.01 means 12.010...0

To determine conformance within limits, the measured value is compared directly with the specified value and any deviation outside the specified limiting value signifies nonconformance with the limits."
"