Control Chart out of control? Cpk - Process Out of Control

Q

qualitytrec

#11
Re: Control Chart out of control???

I am with the others who say there is not enough resolution in your device but there is something different in the measurments as well.
Do some measure hot of the wheel and others after it has cooled slightly, does one person have a tighter grip, are there more than one device in use, more than one operator, more than one machine, more than one appraiser???? At what point in the data is the whell dressed, where was the wheel bumped a little to hard by the work piece? All of this needs to be answered in my opinion if you are going to make the control chart work for you instead of being a "dead" tool.

Mark
 
Elsmar Forum Sponsor

Govind

Super Moderator
Staff member
Super Moderator
#12
Re: out of control???

We have very good inputs from the other contributors so far. If I may add…

Improving the measurement system to be able to understand real variability of the process will help. (and should do as a due diligence) However we are talking about a high Cp, Cpk (or Pp, Ppk) for that matter. Improving measurement system will rather show an improved Cp (Refer Larry Barrentine graph of actual/Potential Cp for various GR & R). This does not explain the pattern. My worry is more about the cyclical pattern. If you notice, the pattern starts at 16th subgroup. As suggested by other contributor, you need to collect more data to confirm if this patter continues.

My suspicion for this pattern is
a. Stability of the measurement system itself.
b. due to mechanical wear in the grinding equipment itself (also variables like pneumatic clamping)
c. The grinding wheel being dressed every 16th subgroup
d. We need to see time stamp for the data. Control chart looks like a subgroup size 3. Was there enough time given between subgroups (i.e rationally sub grouped?)

As another due diligence. The ground diameters have a feature called lobbing. They are best measured by a V micrometer or (if portable) it is best using a V block and single point plunger dial indicator. Rotate the ground diameter and read from the dial min/max. (Diameter is mathematically calculated). Using a normal micrometer will not provide good results.
Regards,
Govind.
 

Caster

An Early Cover
Trusted Information Resource
#13
data on a grind diameter and I attached the data and results in an excel file. Basically, the process range was only 0.0003” (3 tenths), .
My pitiful contribution is suggest you go back before the start and question the tolerance.

Do I understand the spec to be 1.7676 ± 0.0003?

Plus or minus three tenths?

Do you understand the design intent of this part?

Has an actual part designer been able to explain the need for this tight tolerance?

The good old fashioned tool makers rule is 1/10 the tolerance for the tool or gage (which you don't have as you are seeing from the poor MSA, probably failing ndc).

Would you call that three hundredths?

How did you intend to measure this?

How does your customer measure it? Do they have air gages, laser mikes, Marposs? Or same gage as you? If so, why call out this spec?

Should this have been picked up in the quotation stage?

In my world we have customer specs that exceed the ability of state of the art gages to measure, and that have no engineering/design reason to be so tight. They got that way from young designers trying to make their mark by tightening up all the specs, because they think tighter is always better. This causes no end of unhappiness and wasted effort.

If three tenths really matter to the part, get a better gage, if three tenths don't matter, work with the customer to set a meaningful spec.

I think it was Juran who said tight specs loosely enforced are the death of quality systems. It's hard to get operators following SPC rules if there are too many "exceptions".

Great thread so far...
 
D

Darius

#14
I saw the chart and altho is low resolution included in the package, the main problem I see is autocorrelation (if you get a better gauge it will not change a bit). :mg:

If on the meassures is there high value of autocorrelation, the control limits get too small for the data (because the UCL and LCL are calculed on the within sample variation and not the overall variation).

There are 5 ways to get into autocorrelated data:

1- take off autocorrelation (that one meassure has to do with the past one)
2- take more time between samples ( I don't like but somebody said it to me)
3- Make subgroups bigger (Somebody said it, but I don't like without rational sampling taken in account)
4- use an autocorrelation factor into your control limits calculus.
5- Use percentile to determine your control limits (non parametrical)

Good luck:magic:
 

Kales Veggie

People: The Vital Few
#15
I saw the chart and altho is low resolution included in the package, the main problem I see is autocorrelation (if you get a better gauge it will not change a bit). :mg:
Darius,

For our education:

Can you explain autocorrelation in SPC (or a good reference document) and how you determine that autocorrelation is a factor?

Thanks is advance.
 
D

Darius

#16
I forgot to tell wich method I like more in this case (non parametrical , the Cnp and Cnpk calculated by me = 1.333, good but not too high)

Can you explain autocorrelation in SPC (or a good reference document) and how you determine that autocorrelation is a factor?
There are a few articles about that topic, but I will try a short explanation, If is not clear enought, I could explain more (I don't like long speeches)

Sorry for fast drive, I scanvaged my oun posts.
Here is a quick review:

In "advanced topics on SPC" by Donald Wheeler I saw that procedure (to fix control limits).

The autocorrelated control limits (also explained on the same book) inflate the variation estimate by an autocorrelation factor, if I remember it: 1/(1-r2)^0.5

An autocorrelated variable looks like a bunch of trends, changing up and down (the trends of course).

About autocorrelated process, I like to explain a little more:

There are two ways is wich autocorrelation affects periodically collected data, excessive autocorrelation will have visible impact upon:

1. the running record (bu is something that most of the times happen in a continuos process without being a special cause condition)
2. the control limits calculated according to the usual formulas, the "contaminated limits" will be narrower than they should be in order to properly characterize the process.

Calculation:

Autocorrelation_Coefficient=1/(1-r^2) ^0.5

being r the sample correlation coefficient, calculated using

v(1),v(2)...v(n)

x(1)=v(1), y(1)=v(2)
x(2)=v(2), y(2)=v(3), and so on


Sigma = Range_Average/d2*Autocorrelation_Coefficient

Darius.
:cool:
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#17
Your process is not out of control. It looks just fine. You are using the wrong statistics for grinding. Grinding should not be a normal process, it is the uniform distribution. In fact, grinding is the gold standard for the uniform distribution. Using the uniform distribution, your control limits should be 1.7572/1.7678 There is a very simple answer to your dilemma. You need to use the X hi/lo-R chart to control the process. Without rehashing the entire discussion on this thread, please read through the following thread:

http://elsmar.com/Forums/showthread.php?t=20935&highlight=SPC+precision+machining

If you have any further questions after reviewing that thread, I will be more than happy to discuss it.

I will say this, though - you gage R&R might be an issue. For SPC you should have a NDC of 10. Looks like you might have an NDC of 4. That is a problem. You would be better suited to have more resolution in you measurement. However, that is a lesser issue than using the wrong statistics. You will chase your tail forever doing that! :cool:
 
C

CMfgT

#18
Thanks to all, I was offline for a day and wow, you guys/gals were busy. Thanks for all of the post and I will read through them.

Thanks again
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#19
Re: out of control???

As another due diligence. The ground diameters have a feature called lobbing. They are best measured by a V micrometer or (if portable) it is best using a V block and single point plunger dial indicator. Rotate the ground diameter and read from the dial min/max. (Diameter is mathematically calculated). Using a normal micrometer will not provide good results.
Excellent point! This is why using X hi/lo-R is the best way to control grinding. It captures the lobing issue and eliminates that error from the control system. It also clearly shows how much of your tolerance is used up on roundness. It is possible that what may look like a cycle is measurement error from that issue. At this point, there is no way of knowing. Measuring one diameter of a circular feature is nonsense. There are an infinite number of diameters in a circle - how can it be described by one? Calculating the average of of an insignificant number of representative data points in a circle (1 out of infinity?) is even less meaningful. That is one (of several) key defect of X bar-r charts in precision machining. With the resolution, and the lack of hi/lo measurements, no real conclusions can be made on cycles, etc. X hi/lo-R resolves that. See previous link. :cool:
 
C

CMfgT

#20
I got the excel file for hi/low, but i have minitab, anyone know if minitab has hi/low control charts??
 
Thread starter Similar threads Forum Replies Date
A Fabric roll inspection - What type of Control Chart to use? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
Z Putting back excluded rows/data points in a control chart Using Minitab Software 0
B Control chart and sample time Statistical Analysis Tools, Techniques and SPC 1
J Control chart for huge sample size Statistical Analysis Tools, Techniques and SPC 9
M What is "setup control chart"? Statistical Analysis Tools, Techniques and SPC 2
A Redesigning our process flow chart, PFMEA and Control Plan Process Maps, Process Mapping and Turtle Diagrams 4
D Control chart applicable? Percentage of compliance with a standard Statistical Analysis Tools, Techniques and SPC 2
D Link between FMEA, flow chart and control plan FMEA and Control Plans 10
B How to do a control chart - Machining product that requires flatness control Capability, Accuracy and Stability - Processes, Machines, etc. 1
S Understanding control chart and measurement capability Statistical Analysis Tools, Techniques and SPC 2
F Appropriate Control Chart Statistical Analysis Tools, Techniques and SPC 5
S used excel formula calculated the ARL's with Rule1&Rule2 for Shewhart control chart Statistical Analysis Tools, Techniques and SPC 0
H How to define Root Cause when some points are out of control chart Statistical Analysis Tools, Techniques and SPC 6
L The ARL for an X bar Control Chart (n = 5) Statistical Analysis Tools, Techniques and SPC 5
L P-Chart Control Limit Calculation help Statistical Analysis Tools, Techniques and SPC 1
L A Complete Control Chart Decision Tree Statistical Analysis Tools, Techniques and SPC 5
S Variation Study on Mined Material - What do you call it (not a control chart) Statistical Analysis Tools, Techniques and SPC 7
G I-Chart and X bar Chart Control limits Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 3
S Regression Control Chart - Process Capability Statistical Analysis Tools, Techniques and SPC 18
J Control Chart Pattern Detection Statistical Analysis Tools, Techniques and SPC 3
R Control Chart Constants and Confidence Interval Statistical Analysis Tools, Techniques and SPC 7
P Specification and Tolerance in Stability Control Chart Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 9
I What is an "X-Rs" Control Chart? Statistical Analysis Tools, Techniques and SPC 4
I What is an Ideal Control Chart for Hospital Surgery Infections? FMEA and Control Plans 2
M Shewhart Control Chart for SPC In Excel - Explain the yellow and red marked column Statistical Analysis Tools, Techniques and SPC 2
K Advice on Calculating Control Chart Control Limits Statistical Analysis Tools, Techniques and SPC 13
J Statistical Significance and SPC Control Chart Reports Statistical Analysis Tools, Techniques and SPC 9
P SPC Data Collection and Control Chart Type question Statistical Analysis Tools, Techniques and SPC 4
R XBar & R Chart Spreadsheet to apply a Control Process Statistical Analysis Tools, Techniques and SPC 1
Chennaiite Control Chart for Non-Manufacturing Process Indicators - Is it worth? Statistical Analysis Tools, Techniques and SPC 1
O Is it an Assignable Cause only if the Control Chart says so? Statistical Analysis Tools, Techniques and SPC 16
C Control Chart Roadmap Template Document Control Systems, Procedures, Forms and Templates 0
T Shewhart Control Chart - ISO/TS 13530:2009(E), Page 22. General Measurement Device and Calibration Topics 3
M MSA (Measurement System Analysis) - Do we have control of all states in the chart? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
H Control Chart applications in Cement Industries Statistical Analysis Tools, Techniques and SPC 2
T Implementing SPC - Control Chart for Individual Measurements Statistical Analysis Tools, Techniques and SPC 9
R "Control" as used in 4.1 c) - Process Flow Charts (Process Sequence Chart) ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 12
I Control Chart for Mortality Rate on Broiler Farms Statistical Analysis Tools, Techniques and SPC 4
G Control Chart Selection for Crimping Operations for RJ 45 Plugs Statistical Analysis Tools, Techniques and SPC 5
K When to create a new Control Chart? Quick question Statistical Analysis Tools, Techniques and SPC 4
T Control Chart - Salt Spray Test Hours - Powder Coated Parts Statistical Analysis Tools, Techniques and SPC 3
S Calculation of Upper and Lower Control Limits for Control Chart? Statistical Analysis Tools, Techniques and SPC 7
R Control Chart Metrics/KPI - Measuring Effectiveness of Control Charting System Statistical Analysis Tools, Techniques and SPC 4
L Measurement Uncertainty using Control Chart Measurement Uncertainty (MU) 2
R Calculation Mean for Control Chart -X Double Bar Statistical Analysis Tools, Techniques and SPC 13
J What Control Chart should be used for Microbial Limit (NMT 100 cfu/m3) Statistical Analysis Tools, Techniques and SPC 12
B Color Individual Points on a Control Chart? Using Minitab Software 7
A Organization Chart Control and Quality Manual Integration ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 15
S Pre-Control Chart for Single Side Tolerance Statistical Analysis Tools, Techniques and SPC 1
T Control Chart to use when Sample Size is Constant Statistical Analysis Tools, Techniques and SPC 4

Similar threads

Top Bottom