Self define the UCL and LCL instead of by Calculation

[dvd101]

Hello all,

I am using X and Moving R chart due to only 1 data per date.

Can I just simply default the UCL, LCL = +/-1% of Mean X and MR Limit will default by 1% Center line of Mean of X instead of calculate the UCL and LCL???

Any concerns if auditor look into such SPC Chart? How about if I address this issue in MRM, make sure them have considered and agreed on such SPC monitoring practice, is it okay to the auditor?

The SPC always out of control (UCL , LCL, and MR Limit) if we use the calculated UCL,LCL and MR Limit.

Regards,
David

[justncredible]

My guess is your useing the wrong chart, have you tried a weighted moving range, a IWMR chart?

As for what the auditor questions, I do not think a auditor will double check the math of a chart. Even if it is out of control your customer has agreed to use it.

[Tim Folkerts]

I suppose you could use something like +/-1%, but that removes most of the purpose of SPC.

SPC is designed to help you understand and control your process. The rules are designed to alert you to the possibility of unusual, non-random occurrences. By widening the control limits, you decrease the probability of noticing those occurrences.

And just because a point is OOC (out of control) doesn't mean you have to change or adjust something. There will be occasional OOC indications even for a process that is perfectly in control. The OOC indication is a reminder to check the process. Hopefully nothing is wrong and the process is fine. But perhaps there is a problem that needs fixing. If the process is often OOC, that is an indication that you might want to address the underlying problems that are causing the process to perform erratically.

But remember, it is ultimately a question of moeny and value. If the process is often OOC according to the rules, but still well within the spec limits all the time, then there is less cause for correction. On the other hand, if you can identify and eliminate the causes of the OOC, you may be able to improve the process further - potentially gaining new customers or lowering costs.



Tim F

[bobdoering]

In Reply to Parent Post by justncredible

My guess is your useing the wrong chart.

I agree. What is the process, what is the distribution?

[dvd101]

It is the Testing Process, we need to establish the SPC for the Tester which use to sort out the good and bad units.

In general, I think the tester may not need to perform the SPC since "Tester" will not mfg / fabricate anything just sorting. Am I right?

[bobdoering]

In Reply to Parent Post by dvd101

It is the Testing Process, we need to establish the SPC for the Tester which use to sort out the good and bad units.

In general, I think the tester may not need to perform the SPC since "Tester" will not mfg / fabricate anything just sorting. Am I right?

There is something to be said about that, although there may need to do attribute charting if there is a customer requirement for recording the trends. It would be "report card" charting for all practical purposes. Not the best use of SPC, but a use nonetheless.

Let's not forget that X-bar/R is not the only SPC chart!

[Tim Folkerts]

When assessing test units or meters, Gage R&R is often a more appropriate and valuable approach than SPC.

[dvd101]

The GR&R was performed, however, the stability of the Tester also need to perform. Therefore, we use SPC to monitor the Tester's stability.