X bar R chart - Bushing reaming ID - 7 points above or below the center line

S

Sebin Anithottam

#1
Hi,

We have a problem with the X bar R chart for a new bushing reaming process - bushing ID measurement. While plotting the chart we can follow all rules except for one - more than 7 consecutive points below or above the mean or 5 consecutive values with zero variation. We tried all we could to some how get some variation between measurements but we are not successful in doing so. The MSA study I conducted with the gage used shows that the ndc is 10 so I dont think there is a problem with my gage either.

Is there some thing I should do about this? Please advice.

Thanks
Sebin
 
Elsmar Forum Sponsor

Bev D

Heretical Statistician
Staff member
Super Moderator
#2
Reaming process? ousnds like you are simply measuring too quickly within your subgroups...a ration al subgrouping problem (but it's early and I'm only on my secodn cup of coffee and 3rd crisis)

a reaming process will typically have very tight variation piece to piece, but will display setup variation (changing the reaming tool) and tiem to time variation as the tool wears.

The better approach is to change your subgroup scheme; perhaps one reading every 2 hours adn at every setup / tool change? you would need to use a X MR chart tho...
 

Steve Prevette

Deming Disciple
Staff member
Super Moderator
#3
Yes, perhaps spreading the data values apart, or making larger subgroups may help. You may also want to look at the precision of the readings - are they being needlessly rounded off?

If the data are indeed telling you that you are shifting around, you may need to shift the baseline periodically to display this behavior. Then the question becomes - can you live with these shifts? I could also forsee a chart where you display the various baselines and see what their variation are in order to help with this question.
 

Miner

Forum Moderator
Staff member
Admin
#4
Bev D said:
Reaming process? ousnds like you are simply measuring too quickly within your subgroups...a ration al subgrouping problem (but it's early and I'm only on my secodn cup of coffee and 3rd crisis)

a reaming process will typically have very tight variation piece to piece, but will display setup variation (changing the reaming tool) and tiem to time variation as the tool wears.

The better approach is to change your subgroup scheme; perhaps one reading every 2 hours adn at every setup / tool change? you would need to use a X MR chart tho...
Bev is probably on the money on this one, but you need to verify one other item. You said that your R&R is good, but did you consider the gage resolution?

Inadequate resolution can throw off your gage R&R both ways (increase it or decrease it). If the resolution is acceptable, and the variation between consecutive parts is still zero, then Xbar charts are out. The best option would then be X/MR charts as Bev indicated.

You will need to determine the appropriate sampling frequency. This needs to be spaced far enough out to actually see product variation over time, but close enough to detect tool wear before you produce nonconforming product.

There are some more advanced statistical tools such as Autocorrelation that can help you determine the appropriate time interval. Test increasing time intervals until you no longer see significant autocorrelation. That point would be the shortest time interval to use. You may want to further increase the interval for economic reasons if the tool wear rate is much slower.
 
S

Sebin Anithottam

#5
Thanks for your replies.

I think the best option for me is to try the X/MR charts as sugested by Bev & Miner. The tool wear in this case is slow (3000 pieces - tool good for 5 days) since the bushing material is soft babbit material.

Almost everyone I talk to regarding this problem first asks me about the R&R, but in this case the resolution is fine too and the gage is fine.

Could some one give me a little more information regarding the autocorrection tool.
 

Jim Wynne

Staff member
Admin
#6
Sebin Anithottam said:
Thanks for your replies.

I think the best option for me is to try the X/MR charts as sugested by Bev & Miner. The tool wear in this case is slow (3000 pieces - tool good for 5 days) since the bushing material is soft babbit material.

Almost everyone I talk to regarding this problem first asks me about the R&R, but in this case the resolution is fine too and the gage is fine.

Could some one give me a little more information regarding the autocorrection tool.
It's not autocorrection; it's Autocorrelation
 

Statistical Steven

Statistician
Staff member
Super Moderator
#7
JSW05 said:
It's not autocorrection; it's Autocorrelation
Whenever you use a control chart, there is always some level of autocorrelation. This is because you are measuring and charting points in time order.

If you are measuring tool wear, you might want to look into methods that incorporate the slope of the points over time. You can look into regression analysis (though again, you have the autocorrelation issue with time).
 
D

Darius

#8
Maybe the main question in't solved, all the rules don't need to be applied as Don Wheeler said in the attachement article of quality magazine.
http://elsmar.com/Forums/attachment.php?attachmentid=2431

It's true that if you make more sparse the points the less autocorrelation exists, a good idea as stated by Bev is to wait more time between samples, but you may concider the Don Wheeler article if the cost of sampling is too low and you don't want the oportunity of an event get unnoticed.

I added autocorrelation to the control limits, not to know time between samples (there is real no difference in costs if I take automatically a measurement every 15 minutes or every 2 hours), if somebody can explain more or give a link could be nice. :D
 
Last edited by a moderator:

Jim Wynne

Staff member
Admin
#9
Darius said:
Maybe the main question in't solved, all the rules don't need to be applied as Don Wheeler said in the attachement article of quality magazine.
http://elsmar.com/Forums/attachment.php?attachmentid=2431
Wheeler says that there are 3 rules that should be applied more or less uniformly, and one of them addresses the violation in question. Also, the Wheeler piece assumes that the integrity of the data isn't being questioned; application of the various rules is useful in detecting fudged numbers. It's been my experience that people who are likely to fudge the data are also the ones least likely to do it skillfully.

A bit
, perhaps, but I find this bit from Wheeler very questionable:
While a computer program can implement dozens of detection rules, it cannot begin to make sense of them. The computer will never know the context for the data, and for this reason it may flag trivial signals and miss some important signals. This is why I personally turn off all detection rules (except for a point outside the limits) on my statistical-process-control software.

The fact that software can't make qualitative judgments is no reason to turn off the ability to flag detection rules. It's obvious (or should be) that significance depends on context, but the user is free to make those decisions for himself, and ignore warnings that aren't applicable.

 
Last edited:

Miner

Forum Moderator
Staff member
Admin
#10
Sebin Anithottam said:
Thanks for your replies.

I think the best option for me is to try the X/MR charts as sugested by Bev & Miner. The tool wear in this case is slow (3000 pieces - tool good for 5 days) since the bushing material is soft babbit material.

Almost everyone I talk to regarding this problem first asks me about the R&R, but in this case the resolution is fine too and the gage is fine.

Could some one give me a little more information regarding the autocorrection tool.
Here are a couple links that should help.

This link gives a good overview of the autocorrelation method:
http://www.itl.nist.gov/div898/handbook/eda/section3/eda35c.htm

Good statistical software such as Minitab will perform an autocorrelation analysis, but if you do not have access to one this link gives guidance on how to calculate it in a format that you should be able to duplicate in any spreadsheet.
http://www.minitab.com/support/answers/answer.aspx?ID=1163

In summary, you calculate an Autocorrelation index for a specific time interval (called a lag). Continue calculating the Autocorrelation index for increasing time intervals (lags). Plot the Autocorrelation index (y-axis) versus lag (x-axis). You should see the plotted curve drop from a high value until it levels out (becomes asymptotic to the x-axis).

Once the curve levels out, your measurements are no longer autocorrelated, and you will now see actual process variation. This will be the shortest time interval at which you would want to sample. You may want to increase it further for economic reasons if wear is slow.
 
Thread starter Similar threads Forum Replies Date
L The ARL for an X bar Control Chart (n = 5) Statistical Analysis Tools, Techniques and SPC 5
G I-Chart and X bar Chart Control limits Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 3
L How to interpret the average R bar Value shown in the R Chart Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
O Main difference between X-bar sigma chart / X-bar range chart Statistical Analysis Tools, Techniques and SPC 4
M What is the difference between X Bar - R Charts and Average Chart? Statistical Analysis Tools, Techniques and SPC 7
S What decision rules should be applied for x bar-R chart? Statistical Analysis Tools, Techniques and SPC 10
I I-MR Chart vs. X-bar & R Chart - Differences Statistical Analysis Tools, Techniques and SPC 8
R Calculation Mean for Control Chart -X Double Bar Statistical Analysis Tools, Techniques and SPC 13
R X Bar R chart - Interpretation in MSA study Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
M X-Bar Chart - CQE Problem to Solve Statistical Analysis Tools, Techniques and SPC 3
K Individual x Chart - Why use the equation is barX+2.66 bar(MR) Statistical Analysis Tools, Techniques and SPC 6
S Validity of Craig Hane's Global Local Index for X-bar R Chart Assessment Capability, Accuracy and Stability - Processes, Machines, etc. 3
Q X-bar R chart question from CSSBB - QCI CD Professional Certifications and Degrees 2
A Monitoring Out-of-control situations on X-bar and R chart Statistical Analysis Tools, Techniques and SPC 15
S Creating X bar R Chart using XYZ coordinates Statistical Analysis Tools, Techniques and SPC 4
D X Bar Chart - How can I deal with missing values in subgroups Statistical Analysis Tools, Techniques and SPC 12
D x-BAR, p or U chart? Plastics compounding line Statistical Analysis Tools, Techniques and SPC 1
K X bar & R Chart Result - Analysis of the Graph Statistical Analysis Tools, Techniques and SPC 28
W Warping / Flatness SPC - X bar-R chart - Control is Max 0.15 Statistical Analysis Tools, Techniques and SPC 5
K Formulas for X bar & R Chart with one sample size Statistical Analysis Tools, Techniques and SPC 7
R X bar s chart template / form for Excel .xls Excel .xls Spreadsheet Templates and Tools 2
K Seeking: X bar R chart software - Or how to do one in Excel Statistical Analysis Tools, Techniques and SPC 6
K Control limit - How is table for x-bar & R control chart derived? Statistical Analysis Tools, Techniques and SPC 7
F Minitab 19 Anova or X bar # of distinct catagories Using Minitab Software 1
1 How to measure straightness on a HEX Long Bar Manufacturing and Related Processes 7
I Can subgroup size change daily? (X-bar & R) Capability, Accuracy and Stability - Processes, Machines, etc. 5
M CE Marking of Robotic Devices which use 5 Bar Pneumatics CE Marking (Conformité Européene) / CB Scheme 3
Jen Kirley How to evaluate hazards in using bar code scanners in chemical manufacturing Occupational Health & Safety Management Standards 1
A Bar Lights vs LED-606 in vision inspection Inspection, Prints (Drawings), Testing, Sampling and Related Topics 1
J SST .625 dia Cold Rolled 303 bar and Magnetism Manufacturing and Related Processes 6
J Determining if there is Tin Plating on an Aluminum Bar Inspection, Prints (Drawings), Testing, Sampling and Related Topics 6
M Interpreting X bar and R Control Charts Statistical Analysis Tools, Techniques and SPC 2
B New website for AIAO-BAR - Unaccredited Accreditation Body ASQ, ANAB, UKAS, IAF, IRCA, Exemplar Global and Related Organizations 33
smryan Does the NIST certificate that came with my new ball-bar have any value? General Measurement Device and Calibration Topics 8
D One-Two-Three Sigma Control Limits - Calculating Control Limits for X-bar Charts Statistical Analysis Tools, Techniques and SPC 3
Sam Lazzara Bar Sealer vs. Impulse Sealer for Medical Pouch Heat Sealing Qualification and Validation (including 21 CFR Part 11) 4
D ANOVA vs. X bar & R - MSA Study in Mintab - Which to use and why? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
B If I know x-bar, spec limits, and Cpk, can I calculate the control limits? Capability, Accuracy and Stability - Processes, Machines, etc. 16
C 1040 vs. 1040 Modified Bar Stock - Material Question Manufacturing and Related Processes 7
M How to Fix the Standard Bar of a Screw Thread Micrometer General Measurement Device and Calibration Topics 9
J ANOVA vs. X-bar/R Calculations Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
kedarg6500 Writing the symbol for R bar bar (double bar) in Excel Excel .xls Spreadsheet Templates and Tools 3
D For Reference Only or Calibration/Verification - Tape Measures for cutting bar stock ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 26
S Medical Gas Pressure Regulators - Anybody had testing conducted at 450 bar? ISO 10524 Other ISO and International Standards and European Regulations 5
D Relying on Supplier Records for Material Traceability - Bar stock and forgings ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 11
hogheavenfarm Subjective vs Objective Finish Issues - Stainless Steel Bar Inspection, Prints (Drawings), Testing, Sampling and Related Topics 20
M ANOVA / X Bar R Results are different - Why is the difference is so great? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 9
G S30100 steel bar spec - related specs, ASTM or AMS? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 2
M CE Product Registration Bar Codes for each Product - Anyone? EU Medical Device Regulations 2
K Bar coding of gages on usage instead of time General Measurement Device and Calibration Topics 4

Similar threads

Top Bottom