Number of Distinct Data Categories - NDC - AIAG's MSA Manual 3rd edition

[sramasam]

The question that i have is about the NDC... [Number of distinct Data Categories] which is there in 3rd edition MSA.

Queries are,

1 What's the exact Definition of NDC?

2 How is the NDC different from Least count of an Instrument?

3 Will the least count of an instrument influence the NDC [Lesser the LC Higher the NDC]

Secondly, 3rd edition MSA has a guideline for the selection of an instrument. It states that the least count of the instrument shall be one tenth of the sigma of that particular parameter of the part....{optional is 1/10th of the Tol}...
How do we get to know about process variation [sigma] even before the R&R % is known.

Pl. clarify the doubts.

Thanks!

 

[ben sortin]

The number of distinct data categories within the measurement study. If you conduct a gage R&R with ten parts that are all the same then you have a NDC of one. If you conduct a gage R&R with ten parts that are all different then you have a NDC of ten. It is a good practice to select ten parts across the whole tolerance or process.

Least count refers to the resolution of the gage. The smallest increment that can be read without interpolation.

The least count does not effect the number of distinct categories. The improvement of resolution will reduce measurement error if the gage is read correctly.

Choose a gage with a least count of 1/10 the tolerance if you are only going to measure the parts for conformance to specifications.

Choose a gage with a least count of 1/10 of your process variation if you are going to improve the process.

Obtain your estimate of process variation from a surrogate process if at all possible.

You are only as good as your tools (gages in this case).

 

[Atul Khandekar]

Hi sramasam,

Welcome to the Forums.

NDC is the number of different categories into which your measurement system is able to classify the data.

You generally start with a ratio of 1:10, so that your instrument can at least detect 10 categories within the tolerance band. Due to the r&r error, the 'discrimination power' of your measurement system may be reduced. Note that within the spread of R&R error, the system cannot distinguish between different readings. For the measurement system to be acceptable, nDC should be more than 4.

This issue of nDC has been previously discussed. For example:
ndc in MSA 3rd Edition

Hope this helps.
-Atul.

 

[Atul Khandekar]

3 Will the least count of an instrument influence the NDC [Lesser the LC Higher the NDC]

LC is the minimum change in the reading that an instrument can report (without interpolation, as ben said.). In case of graduated scales, you can actually try to interpolate and read upto half graduation. Generally, the data should be recorded to as many decimal places as possible. Any rounding-off (possible with higher values of LC) may adversly affect the r&r and nDC.

 

[sramasam]

nDC

Atul & Ben,
Thanks for the response.
Got a question after going through the example given by YKT....

In that example the min weighing unit is 5 grms... Available packs weigh 1.010, 1.007 & 1.011 grams...

So it will fall under one category...and the weighing scale would show the same value for all 3 packs. Clearly understood....

Questions that i have are,

1 Whatif the minumum weight in the weighing scale is 0.001 gram and the least count also is 0.001 gram?

2. What if the weight of the packs are 5.010, 5.007 & 5.011 grams and the least count is 0.001 gram?

3 What if pack weights are the same as given in Pt# 2and the least count is 0.010 gram...

Got stuck up here....

Appreciate your clarifications..