Is a GR&R necessary for Weighing Scale?

[sinned]

If I consider the operator of a weighing scale has no effect on the measurement result, then the need for GR&R study will essentially reduces into one of repeatability study.

If I consider the operator putting the part onto the scale vary between different people, then the GR&R is needed to see if there is any operator effect.

Which one above make more sense ?

b.rgds - sinned

 

[Ron Rompen]

Re: GR&R necessary for Weighing Scale?

Just to be devils advocate here, how have you determined that the operator has no effect on the measurement result? I have used that arguement (long ago) with a CMM holding parts on a fixture, and it was explained to me that the operator CAN have an effect on the results - is the part/sample properly loaded and centered, did ALL the sample get placed on the scale, did the operator remember to properly zero the scale, enter the tare weight, etc etc etc?

 

[howste]

Re: GR&R necessary for Weighing Scale?

I believe that the operator can have an effect on the measurement results. If you do the R&R it will tell you if the operator has an effect or not.

 

[Bev D]

Ron and Howste are correct. the only way to answer the question concerning the operators' effect is to do the test. It's a very easy test to do as well so there is really no need to debate. debates only result in opinions, the test results in facts.

 

[sinned]

Thanks for all of your contribution.

For sure, I know that doing a GR&R does no harm and that the result can answer the question whether the measurement system is adequate or not.

Nevertheless, it is exactly what bother me that many engineers jump to "just do a GR&R" with no thought of what measurement system they are really working with. That this become a no-brainer answer without defining and understanding the variables of the measuring system at all.

For example, i saw a measurement which has only one authorized operator in the whole plant. But for the sake of making up the three persons and three set of data, they just got someone who will not involve in the measurement (probably for the rest of their life) to take part in the GR&R.

We have weighing scale that are linked to computer and not requiring any manual "zeroing" when doing measurement (this was an attempt to avoid missing zeroing); and the effect of placing sample on different parts of the scale surface been studied to be insignificant. And sadly this all efforts will not be recognized.

 

[Ron Rompen]

You appear to have done the majority of the work already in validating your measurement system - but (looking at it from a customer or auditor point of view) you have no data or other evidence to support your contention.

In some cases (particularly in automotive) you eventually have to bow to the weight of the standard, and do what they ask even if it doesnt always make sense (at the time).

Thanks for all of your contribution.

For sure, I know that doing a GR&R does no harm and that the result can answer the question whether the measurement system is adequate or not.

Nevertheless, it is exactly what bother me that many engineers jump to "just do a GR&R" with no thought of what measurement system they are really working with. That this become a no-brainer answer without defining and understanding the variables of the measuring system at all.

For example, i saw a measurement which has only one authorized operator in the whole plant. But for the sake of making up the three persons and three set of data, they just got someone who will not involve in the measurement (probably for the rest of their life) to take part in the GR&R.

We have weighing scale that are linked to computer and not requiring any manual "zeroing" when doing measurement (this was an attempt to avoid missing zeroing); and the effect of placing sample on different parts of the scale surface been studied to be insignificant. And sadly this all efforts will not be recognized.

 

[ncwalker]

Nevertheless, it is exactly what bother me that many engineers jump to "just do a GR&R" with no thought of what measurement system they are really working with. That this become a no-brainer answer without defining and understanding the variables of the measuring system at all.

The engineers are not wrong and you are not wrong. They are right in that you need to do a GR&R. You are right in that you don't "just" do one. You need to give it some thought so that it is conducted properly.

Let's talk about no-brainer answers.... I hear ya. When you are in the business of doing something and know it real well, you develop a sense that things are no-brainers. I have a lot of expertise in an industry that is considered a "black art" and there are lots of guys with experience that say they do things because "that's the way it's done, it's a no brainer." And yet, often I conduct an experiment and prove them wrong. So be careful with these type of statements. At first glance, a scale may SEEM like a no brainer. And I'm not looking at your scale, it's resolution, etc. I know my bathroom scale is pretty good at weighing me and my family. I'm less sure it's good for weighing my cat because my cats weight is small compared to the operating range of the scale. So I am more inclined to say "It's a no brainer" on a person, but I'm not willing to say it's effective on cats without testing it. Now, if this was the umpteenth scale you were calibrating and you have 20 studies in hand where it passes EASILY against the fluctuation and tolerance of your new situation, THAT's a no brainer. But if it's not, it MAY not be (I say with only knowing it IS a scale. Point is - be careful).

For example, i saw a measurement which has only one authorized operator in the whole plant. But for the sake of making up the three persons and three set of data, they just got someone who will not involve in the measurement (probably for the rest of their life) to take part in the GR&R.

This statement tells me there's a lack of knowledge of what GR&R is and why it is needed in your facility. Also, probably a lack of how to conduct the studies.

And as the guy above me (who's name I can't remember because I'm in the edit window and don't know how to find it, sorry) said - sometimes it is easier to just comply with the silly automotive requirement (they are there) than fight it. Not a bad path, as long as it isn't compromising actual quality.

 

[Ninja]

... the only way to answer the question concerning the operators' effect is to do the test. <snip> debates only result in opinions, the test results in facts.

...Nevertheless, it is exactly what bother me that many engineers jump to "just do a GR&R" with no thought of what measurement system they are really working with. That this become a no-brainer answer without defining and understanding the variables of the measuring system at all....

Understanding is a good thing. Thought is a good thing. Using a brain is a good thing. But without data, it is thought, theory and debate.

"Just do a GRR" can come from a lack of thinking, true.
It can also come from a desire for data to make a well founded and defensible conclusion.

You think that there is no influence by operator.
That is an opinion...albeit a well informed and likely correct opinion.
Get the data...then it won't be an opinion anymore...that's what I'm reading up above...

To quote others: In God we Trust...all others must provide data.