Can someone clarify "Resolution", "Discrimination" and "ndc"?

[Clouds - 2007]

Hi,
Q1:
According to MSA(3.0), p.43, discrimination and resolution can be said the same. It's all about the ratio of minimul scale of instrument over the range to be measured. The range is also refered to the tolerance of product, sometimes process variation. (1:10 is a suggested rule).
So, can I make such statement, that resolution/discrimination is a relative index (it depends on how big your production Tolerance is, or how big your process variation is.) it's not a absolute feature adhered with instrument? Yet the minimul scale is an adhered feature of instrument.

Q2:
ndc indicates how well the instrument can differentiate from process variation. Say 1, 5 or 20 categories, etc. It's again an relative index, depends on the tolerance/process variation. My question is, why in ndc, it's recommended to have >5 categories, yet the resolution is suggested to be at least 1:10?

Pls. help to clarify. Thank you very much.
clouds

 

[harry]

Welcome to the forum, Clouds.

It's Sunday night over in the US and therefore pretty quiet over here. Please be patient and I am sure you would get your answers by Monday (US time).

We have a number of regulars in the 'MSA' thread and I am sure they will respond to your question when they drop by.

Meanwhile do browse around for these issues had been discussed in great lengths previously.

 

[Marc]

Some existing discussion threads:
Resolution
Discrimination
ndc

Anyone with specific answers to the posters' questions will be appreciated.

 

[Miner]

Hi,
Q1:
According to MSA(3.0), p.43, discrimination and resolution can be said the same. It's all about the ratio of minimul scale of instrument over the range to be measured. The range is also refered to the tolerance of product, sometimes process variation. (1:10 is a suggested rule).
So, can I make such statement, that resolution/discrimination is a relative index (it depends on how big your production Tolerance is, or how big your process variation is.) it's not a absolute feature adhered with instrument? Yet the minimul scale is an adhered feature of instrument.

Not quite. Resolution and discrimination are indeed the same, but they are not a ratio. They are the smallest incremental measurement that a gage is capable of reading. For example, a digital micrometer that reads in 0.0005" increments has a resolution/discrimination of 0.0005". The 1:10 ratio that you mention is an old rule-of-thumb (guideline) used before MSA methods were developed. This guideline, though not perfect, did provide some allowance for measurement error. MSA studies have show that the 1:10 guideline is imperfect though, so it is really only useful for identifying potential gages.

Q2:
ndc indicates how well the instrument can differentiate from process variation. Say 1, 5 or 20 categories, etc. It's again an relative index, depends on the tolerance/process variation. My question is, why in ndc, it's recommended to have >5 categories, yet the resolution is suggested to be at least 1:10?

Pls. help to clarify. Thank you very much.
clouds

The two are not really related. The 1:10 guideline was based around the tolerance and ndc is based around the process capability. The guideline does not incorporate any measurement variation, while ndc does include measurement variation. There are so many differences that any comparison of the two is irrelevant.

 

[Clouds - 2007]

Hi,
Thanks a lot for the knowledgable comments, and that's why I joined this forum.

rgds,
clouds,