MSA clause 7.1.5.1.1 - Statistical studies on counting weight scale

[SP-Quality]

How do others do statistical studies on a weight counting scale? We have our scales calibrated annually and confirm counts throughout the packaging process. Is a statistical study required? Any insight is appreciated.

 

[leftoverture]

Our company is not yet IATF, but we have never done any MSA activity on our counting scales. We have invested a lot of effort to train people to use them properly and to set sample sizes that yield the best accuracy (since our part weights vary so much). None of my customers has ever asked for any further MSA activity, but I have one customer who wishes we would do more frequent (daily) calibration checks. But based on my outside lab calibration results, the data doesn't suggest a need for further calibration effort.

 

[Johnny Quality]

Why would you do a statistical study on a weight scale used only to count parts?

If you are using the weighing scales for inspection/SPC to customer requirements then yes, you should do an MSA if you haven't already.

 

[AMIT BALLAL]

I agree with @Johnny Quality.
And 7.1.5.1.1 requires MSA study of measurement systems specified in the control plan. Refer control plan, if it specifies certain measurement system, it requires a MSA study.

 

[Bev D]

I certainly agree that a repeatability study is probably not warranted for most scales used to count parts by weight. However, a MSA is a measurement system analysis which is more than a ‘simple’ Gauge R&R study. The purpose of any control is to ensure the quality of whatever it is assessing. The MSA is intended to establish this ability. If the system can’t achieve it’s goal of assuring quality then you need a different system.

For counting scales the analysis is to see if we can reliably differentiate between 4 parts and 5 parts, or 100 parts and 101 parts,etc. essentially, the weight of the parts themselves has to be consistent enough that you actually can count them using the weight. If you can’t then you need a different system. Perhaps a ‘counting’ sensor can be used or packaging like an egg carton that specifically holds the correct number of individual parts.

 

[SP-Quality]

We reference weight counting scales and visual inspections on our Control Plan so our auditor is requiring that we do MSA studies for both. We already do it for all measurement equipment.

 

[leftoverture]

Remove weigh counting from your control plan. I see no need to have it on there. At my company, weigh counting is covered by the process work instructions and is not on the control plan and I have never had this questioned.

 

[Matt's Quality Handle]

Remove weigh counting from your control plan. I see no need to have it on there. At my company, weigh counting is covered by the process work instructions and is not on the control plan and I have never had this questioned.

I've been raked over the coals over parts counting by an Automotive OEM customer that rhymes with "Rhonda". So this may be relevant. It was a big deal for our customer, so it made it to the control plans.

 

[leftoverture]

I've been raked over the coals over parts counting by an Automotive OEM customer that rhymes with "Rhonda". So this may be relevant. It was a big deal for our customer, so it made it to the control plans.

That's interesting. I have a customer where we have experienced quite a few counting errors, mainly because they "kit" the parts on their end so they always find even the smallest counting error. We have found the biggest reasons for counting errors are: 1) operators who don't properly set the scales; 2) using too small of a sample size for the part (light parts need larger sample); and 3) using too large a scale for the box count/part size. A traditional "gage R&R" study would not resolve those issues, though I think if we'd have documented the studies we've done on counting it would certainly qualify as an MSA, just not in the traditional sense. How did you handle the MSA with "rhymes with Rhonda"?

 

[Bev D]

Having worked for Rhonda and being a Customer of 'weigh counted' parts where the count is critical and supplying 'kits' that must have the correct count of parts in them (medical imperative) I can state that there are times when the correct count is absolutely necessary and not just because of regulatory 'accountability'.
Rhonda and us: a line stoppage if the count is too low is very expensive.
Medical: if the disease is somewhat rare and the test can expire the Vet may only buy one kit. the disease can be not only fatal but highly contagious. ensuring that a critical part of the test is in the kit box is essential.
Regulatory: some materials are highly regulated and every part or drop must be accounted for as they cannot "fall into the wrong hands".
these are just the times off the top of my head when there should (really must) be a validated control mechanism. Not all counting belongs in this category but some do. As quality professionals we need to recognize this and more importantly understand it and ensure that it's done correctly.

You are correct that a 'traditional' gauge R&R will not help. but a measurement systems analysis will help. see my post above for that and other ideas about ensuring counts. no one can count so we need to move on from that and error proof the process when counts are necessary.