Control charts of finished product characteristics..NonSense?

[guyramesh]

My dear quality gurus;

I have always believed in plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product.

However, my supplier is providing 'control charts'? based on characteristic of end product batches packed on different dates,complete with USL, UCL, LSL, LCL.
Does charting end product characteristics over many batches considered SPC control?

Please refer to attached chart example. Need your support to face my mgmt and straighten this thing once and for all.

rgds,
Ramesh.

 

[DsqrdDGD909]

My dear quality gurus;

I have always believed in plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product.

However, my supplier is providing 'control charts'? based on characteristic of end product batches packed on different dates,complete with USL, UCL, LSL, LCL.
Does charting end product characteristics over many batches considered SPC control?

Please refer to attached chart example. Need your support to face my mgmt and straighten this thing once and for all.

rgds,
Ramesh.

I'm sorry, but I don't understand. Is the chart showing individual lots or averages of lots?

 

[Geoff Withnell]

My dear quality gurus;

I have always believed in plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product.

However, my supplier is providing 'control charts'? based on characteristic of end product batches packed on different dates,complete with USL, UCL, LSL, LCL.
Does charting end product characteristics over many batches considered SPC control?

Please refer to attached chart example. Need your support to face my mgmt and straighten this thing once and for all.

rgds,
Ramesh.

What your supplier is giving you is evidence that his process is in control and the characteristics charted are within the requirements you have set. I would not myself send outside my organization the actual "process control" charts. How I control my process is proprietary. What you as a customer need is evidence that the process producing the product is in control (evident from the output characteristics being in control) and the product meets the specs (evident from the control limits on the output characteristics).

Geoff Withnell

 

[bobdoering]

My dear quality gurus;

I have always believed in plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product.

However, my supplier is providing 'control charts'? based on characteristic of end product batches packed on different dates,complete with USL, UCL, LSL, LCL.
Does charting end product characteristics over many batches considered SPC control?

Please refer to attached chart example. Need your support to face my mgmt and straighten this thing once and for all.

rgds,
Ramesh.

Well, it is a type of charting methodology called "report card charting". It indicates that the aggregate of controls prior to the finished product is effective. Is it the best use of charting? No...but is is one use. If the aggregate is so far in control, then it may be suitable for verifying outgoing product quality level and significant changes, when they occur. The best use is plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product!

 

[Darius]

IMHO, while is the same product (and no change whatsoever on it) and no special filtering on the data, could show the control your supplier have on the process, so SPC, your supplier could give you a cpk or ppk value, but is giving you the actual data in order to show that that property is a don't-worry characteristic for you. You may say that you can't actually do nothing on a death corpse, but as a statesman said more or less "the country that doesn't know it's history is condemed to repite it"

, the stats hide information if the data doesn't behave in a "normal gaussian way", so no trends or "non gaussian distribution" that can be hiden on an indicator. I am agree that SPC is more or less distribution sensitive (without the rules), but the indicator is not, but we (I include my self) are more confortable with a number than with a chart in order to compare the actual behaviur against the previous ones.

btw. the chart looks with very few data (bad thing, so the chart is telling you nothing) and a little close to the limit (not centered on specs), remember that "Quality is low variation on target".

 

[Jim Wynne]

What your supplier is giving you is evidence that his process is in control and the characteristics charted are within the requirements you have set. I would not myself send outside my organization the actual "process control" charts. How I control my process is proprietary. What you as a customer need is evidence that the process producing the product is in control (evident from the output characteristics being in control) and the product meets the specs (evident from the control limits on the output characteristics).

Geoff Withnell

I've seen situations where suppliers didn't want to send a customer a PFMEA and sometimes a control plan--documents that do include proprietary information, but I've never had one refuse to supply control charts, especially where PPAP is concerned. There's nothing that should be considered proprietary or secret in a control chart, unless the supplier is trying hide something.

 

[Jim Wynne]

Well, it is a type of charting methodology called "report card charting". It indicates that the aggregate of controls prior to the finished product is effective. Is it the best use of charting? No...but is is one use. If the aggregate is so far in control, then it may be suitable for verifying outgoing product quality level and significant changes, when they occur. The best use is plotting and monitoring control charts of input parameters which directly affect characteristic of process output or end product!

Yes! But it's also important to note that the thing that's missing from ex post facto charts is the chronological element.

 

[bobdoering]

Yes! But it's also important to note that the thing that's missing from ex post facto charts is the chronological element.

It is grossly chonological...but, I agree, it will provide very little information for adjusting the process - or even saying where in the process there needs an adjustment- if an out of control condition occurs. It can only initiate a corrective action investigation. Clearly, report card charts are the weakest....but better than nothing.

 

[Bev D]

I actually require these charts from our suppliers. I think Context is important here. Bob/Jim: you are correct that for the supplier control charts on inputs are much more valuable especially when the intent is to control the process (as you say Bob for adjustment). However, for the Customer, as the Customer, I really need to know that the supplier is meeting my drawing requirements and is stable and capable. This way I can get rid of incoming inspection of parts and have some confidence in their ability to meet my delivery requirements. I also have to remember that sometimes the factor that goes wacky and makes my parts not meet spec isn't oen that is on the control plan and so a control chart on the output (blueprint) characteristic makes sense and has value.

Now I do require the supplier to understand his input characteristics and how they effect the output, set appropriate tolerances and have the appropriate controls in place for those inputs. I will aslo audit those charts on occassion but I do incoming inspection on the output control charts not the parts.