CONTROL CHART BASICS and the X-BAR AND R CHART +++++ EXAMPLE 

THANKS!!!!

Wow, thank you so much Iman!!! I am extremely blessed, and I want to be a blessing to others!

Hey Ash, I'm glad the video was helpful!!!!

You're absolutely welcome Rodolfo, and thanks so much for the great feedback and positive comment!

Does that +/-3x global stv still senstive enough to capture the abnormal signal? what is the pro and cons for those 2 methods?

Global standard deviation is far less sensitive to special causes because it includes ALL variations ALL AT ONCE with no regard to time in its calculation. This includes special cause variations which tends to heavily skew your limits. The entire purpose of a control chart and its calculations is to identify significant differences between short term (within subgroup) and long term (between subgroup) variation. For a stable process you would expect the short term outputs of a process to be the about same as the long term outputs.

Glad it was helpful!

wow, thanks so much!!!! I appreciate the awesome feedback!

You're absolutely welcome!!!!

Thank Yvette!!! You're absolutely welcome, I'm just glad I was able to help you!

You're welcome Lizzy!!!!

You're absolutely welcome!!!!

You're welcome!!!!

Thanks so much!

You're welcome Tammy!!!

You're welcome!!!

Thanks so much Devanshu!!

Thanks! I'm glad you liked it!

Wow, thanks Reynard!!!

Wow thanks so much, I appreciate that!!

You're welcome Alan!

Wow thanks Naveena, I'm glad you enjoyed it!! And thanks for the comment!

You're welcome!!! I'm glad you enjoyed it and found it easy to follow!

Hey There!! I'm glad you enjoyed that video! Okay about the requirements to sit for the exam, check out this video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-qPLGuXYPXUI.html And then to join the academy, check out this free course: cqeacademy.teachable.com/p/top-10-topics That free course will also come with details about the full course - The CQE Master Class

You're welcome!

Wow, thanks Jacques!!!

Thank you!

Thanks!

hahahaha, you're welcome!!!

You're welcome!!!

Glad you enjoyed it! -Andy

You're welcome Amsyar!

You're welcome!

You're welcome!!!

You're welcome!!!!!!!

You're welcome!

You're welcome!

Awesome, thanks for the feedback, I'm glad you like them!

@@CQEAcademy also make a video on MSA

@@akanksharai1366 will do! I'm sure i'll create a video on MSA here in the next 2 months or so. BTW - feel free just to call me Andy :)

Thank you!

Hey Sunny! That honestly depends on your process, and how stable and mature it is, and there is no hard and fast rule. If you go-live with a control chart and your process appears to be stable over time, then that re-calculation does not have to happen that often. If you go-live and your process is not stable, you can do both - address those sources of special cause variation, and re-calculate your control limits. But do make sure not to include any OOC (Out Of Control) data points when calculating those control limits.

You're welcome!!

Thanks Darcie!!!!

Thanks, I'm glad you enjoyed it!!!

Awesome, thanks Kaneki!!

Thanks Sameera!!

Hey West, Generally the answer is no, unless you're measuring some attribute that can also be zero. For example if you were monitoring the temperature of a cold process, you could realistically measure temperature values below zero. However, if you're simply measuring a small attribute that can never be a negative number, and your lower control limit is a negative number, the general practice is to move that up to zero. -Andy

You're welcome!!!

Hey Chong! For question #1, most tables stop at 25. . I tried to google it but couldn't find anything larger. Honestly, after 25 samples, you should stop using the range value and start using the standard deviation (X-bar and S chart). That transition to the S chart usually happens around 10 samples. For #2, no, you need to constants to calculate the UCL and LCL. You might not need the table I guess if you have the constants memorized. For #3, the question here is all about rational subgrouping, and you want to minimize the variation within a sub-group. I doubt you could argue that all of the measurements on a single roll are 1 rational subgroup. It seems more rationale that whenever you take multiple samples at each meter in length, those measurements would be considered a single rational subgroup. Honestly without knowing your process better it's hard to say what a rational subgroup is. #4, there are excel spreadsheets to perform all of the control chart calculations (UCL, LCL, etc), all you have to do is type in the measurement data

Thanks so much!!!

Hey Samuel!!! Great suggestion, that's definitely a good future video! okay, so it is possible for special cause variation to impact your R chart, but not your X-bar chart. That's why we have both charts, because it's important to monitor both the center of your process (X-bar) and the variation within your process. And either chart can alert you to special cause variation!

Thanks so much!

Thanks! I'm glad you like it!

Thanks Rushikesh!!!!!!!!

You're welcome Gretel, and thanks for the comment!

Same questions too 🙋

hahahaha thanks!!!! I'm glad you love it!! -Andy

Awesome, thank you Rupesh!!! I'm glad I can help you grow!

@@CQEAcademy Ya, it helped me in my diploma exam... 😅 Last minute studies!

@@rupeshchoudhari96 fantastic, what are you getting your degree in?

@@CQEAcademy I'm currently in last year Diploma (Mechanical)

@@rupeshchoudhari96 Best of luck!!!! I got my degree in mechanical. Let me know if there are any other topics that you'd like to see. -Andy

Thanks!

Wow thanks so much Jarod!!!!

Great catch, and you're correct, thank you though!!

You're welcome Muhammad!!!

Hey Manikandan!! I love this question - they actually can be derived from the d2 values, and sample sizes.

Thanks Vikas!!

You're welcome Peino! I do cover this topic in alot more detail in my course, The CQE Master Class

Thanks

Hey Agra, I assume that when you take those 2 measurements, that you're measuring the same characteristic - not two different things (dimensions, etc). I haven't run into that situation before - why would you measure the same sample twice?

@@CQEAcademy Yes, it is the same characteristics but 2 replications. I am not really sure why there are 2 replications because I just retrieved it from old data, probably because there is a doubt in the gage precision

@@agradhanurwedhasakti4865 That's what came to mind when I first read the two replicate measurements, that someone was worried about gage precision. Honestly, what I'd do is run a Gauge R&R, and if the R&R is acceptable - go down to only a single measurement! That might be a huge form of waste (over-processes) by measuring the same sample twice. And if the gauge R&R is bad. . . it might be time to find a new gauge - too much precision should not be overcome by repeated measurements.

Hey Great question!!! Okay, so a rational subgroup of 5 would mean 5 unique samples measured 1 time. It's not a single sample measured 5 times. That type of variation would reflect the variation in the measurement system, not the variation in the process. -Andy

@@CQEAcademy Oh, thank you for the quick reply! I wondering, how this kind of UCL and LCL calculation can happen in the reality. Imagine there is a production line, (a work station e.g. turner machine). How can I set up the control line for the process? Should the control lines be there, before he begins the job? As planned, every half hour a sampling will be measured (e.g 3pcs/sampling). After 4 hours there will be 8 subgroup- measurement results. After I have 8 result, i calculate the control line. 1. How the process will be monitored before the control lines' calculation, in the first sevaral hours? 2. What is the meaning of the control lines' calculation, if the next and next and next subgroups' result will again and again change the UCL and LCL values? Sorry for the detailed question, i just would like to understand this very well.

@@surnamegm Hey There, Okay, so generally, you want to collect around 25 sub-groups to calculate the control limits. Until you have control limits in place, just consider the chart to be a "run-chart" with no control limits. Also, before calculating the control limits using those first 25 sub-groups, you want to feel reasonably certain that the process was only experiencing normal (inherent) variation. You should not re-calculate the control limits for each new data point. However it's not unreasonable to re-calculate the control limits if a change is made to the process. This is something that you'll get better as assessing over time (the need to re-calculate the control limits). It's hard to create rules for this scenario because every process is different. Hopefully that helps! -Andy

@@CQEAcademy Thank you very much!

@@surnamegm You're welcome!

Great question! I would recommend using a consistent sample size over time because it will make it easier to manage the control chart over time. Changing sample sizes is not impossible, but you have to make sure that as you change the sample size you don't violate the requirement of having a rational subgroup, and you will have to re-calculate your control limits with your different sample sizes.

Hey Nicole! Can you clarify a bit, what do you mean by the performance target?

Thanks!!!!!!