Sir your video is superb regarding one doubt sir How to calculate the confidence level and risk, how do you say +ve side 2.5 and -ve side 2.5 in case am getting various + and - side value means how do i want to calculate Or else what ever the value observed by operator +/- side we should consider 2.5 %??
Hi Saravanaa, theoretically we have to consider that both side (- / +) 2.5% risk is there. But actually it may vary. So, it is only the assumption and used for calculation purpose
There is another way to judge whether bias is acceptable or not. That is through the p-value associated to t-statistic bias. According to AIAG MSA 4th edition, bias is acceptable if the p-value associated to t-statistic is less than alpha 0.05. Can you explain the same number example using p-value?. Because i can see p-value result always contradicts with the zero needs to be between lower and upper bound criteria.
I have a doubt at 18:13 . Confidence level is 1 minus Alpha divided by 2. But we are considering here as only Alpha divided by 2. I am confused on why 1-minus is not considered. Thankyou 😊
Hi Sekar, we are taking the t value based on alpha divided by 2 value only and not based on confidence level. This is the basic rule followed to find the t value.
The difference between mean value (15.007) and reference value (15.000) will give the bias (0.007).This bias is required for calculation of other parameters
When will the subgroup value be 2? As per BIAS, We can do only, 1 appraiser, 1 part, 1 parameter,1 place, if I am adding one appraiser to take the same parameter & same part in the same place could it change the subgroup value is 2? please clarify sir, Thanks in advance
Hi, I have been working in switchgear industry and lately our company has instructed to do MSA study. But we are mostly using instruments such as High voltage test instrument , low voltage test instrument , timing test instrument, snatch gap test instrument bla bla bla. Which MSA study fits for these kind of instruments, could you please assist?
Hi Mohammed, thank you for your comment. Hope you are producing non automotive parts. This MSA study is applicable for instruments which are used for checking the automotive parts. And especially for physical (mechanical type) measuring instrument and gauges only where variations may come due to instrument and inspectors. For electrical and electronic devices used to check the voltage, current , etc, i think calibration only sufficient. Pls check with any other source / experts for re-confirmation.
Hi Vignesh, if t statistics value is greater than significant t value, then the bias is more & will affect the measurement. If t statistics value less than significant t value, then there is very negligible bias only is there and measurement is OK. This point is not told in video. Thank you for asking this right question. Keep it up.
Sir Nanga use pandra formula vanthu bias-[d2×sigma b /d2*(tv,1-alpha/2] enaku d2 value 3.078 entha table la irunthu eduthanganu therila nanga use pandra sub group size 10 sir
Sir and also one more doubt t staticala degrees of freedom illa confidence level illa then ethuku athai derive pannanum. Athuku straight ah tables la poi t value edukulamla
I have verified your doubt and giving my comment. 1) You are not using correct formula as per MSA manual 2) You are using d2* - 3.078 - value taken from Appendix C d2* table for sub group size 9.but you have to use 3.18 as per table (use our table - link given in description) 3) Degrees of freedom to be taken from Appendix C d2* table for No of sub group & sub group size. Then actual t value to be calculated based on that degrees of freedom from T distribution table (see our video again). Because that table has values for natural numbers (1,2,3...) only. But we have to find t value for fractions (example 10.8). If you have further doubt. pls send separate comment (not replying to this comment). Then only i can easily view your comment and can reply
Sir, Very good explanations.Thanks for your time and efforts. I have one doubt on Bais Excel format In Excel format, you created a Histogram chart. In the histogram chart, you had used some formulas to find out the " Selecting no. of classes =" ( T24 Cell ) Could please clarify this formula and what is meant by 1.Selecting no.og classes ( T24 Cell) 2.FREQ.(W27) 3.CU. FREQ.(X27). Thankyou once again.
Hi Samynathan, 1)(T24) Selecting no of classes is to decide the width of Histogram (if range is small, small width of histogram will come, if range is big, bigger histogram will come by formula, because range divided by no of classes will decide the width). Pls study the formula again. 2) (W27) frequency is used to count the number of values (between histogram's each column) for histogram's bar graph 3) (X27) cumulative frequency is to confirm that all values taken for histogram (total 15) Hope now you understand.
1-alpha/2 = 95% = confidence level alpha = risk = 5% alpha / 2 = both side (+side & - side) 2.5% risk from reference value (middle) so, t table will be checked by using both side risk ( two tail risk ) so both are different. Hope i have cleared your doubt
Let's improve our knowledge in Industry Subscribe here : ua-cam.com/channels/bsoHyXxtAelbMYicLaHQGQ.html
sir format link share sir. your whats up number
Really your explanation is super & excellent. Easy to understand the concept.
Thank you
Very nice explanation. Articulated very well.
Thank you Mr.Uthirapathi. Keep watching and sharing our videos to other friends also
fantastic bro, really very very clear explanation,
Thank you bhaskaran. Keep watching and sharing our videos
Thanks sir .... For supportive video
Thank you Manish Rathi. Keep watching and sharing our videos to your other friends
Thank u sir very nice explanation
Thank you Mr.Santhosh
Sir d2 star method to be used when m greater than 2 as in control chart. Here we have to use independent sample method. Please clarify
Super sir very good explanation
Thank you. Keep watching
Very good explanation
Thank you Muthu, Keep watching and sharing our videos
much more useful video sir
Thank you Ganesan
Very nice sir thank you sir 🙏🙏🙏🌹🌹
Thank you Pathmanathan. Keep watching and sharing our videos
As usual super
Thanks a lot
Sir your videos are useful and easy to learn .
I need one help , in bias study pls share any link for t table and c appendix table
Dear Narendran, now check the description in part-3. Download links for tables is provided.
Nice sir.
Thanks
Thank you vijayan
Sir your video is superb regarding one doubt sir
How to calculate the confidence level and risk, how do you say +ve side 2.5 and -ve side 2.5 in case am getting various + and - side value means how do i want to calculate
Or else what ever the value observed by operator +/- side we should consider 2.5 %??
Hi Saravanaa, theoretically we have to consider that both side (- / +) 2.5% risk is there.
But actually it may vary. So, it is only the assumption and used for calculation purpose
thank you so much sir👍
Thank you Mahendhiran. Keep watching and sharing our videos to other friends
Thank you sir 🙏
Thank you Aravinth. Keep watching and sharing our videos to your other friends
There is another way to judge whether bias is acceptable or not. That is through the p-value associated to t-statistic bias. According to AIAG MSA 4th edition, bias is acceptable if the p-value associated to t-statistic is less than alpha 0.05. Can you explain the same number example using p-value?. Because i can see p-value result always contradicts with the zero needs to be between lower and upper bound criteria.
Hi Muhammad, i try in future
Hi sir,
When bais study requirements for instructions.
What are the instrument applicable for bais.
It is applicable for all measuring instruments except gauges
t.q sir best explanation sir
Thank you Ganesan
I have a doubt at 18:13 . Confidence level is 1 minus Alpha divided by 2. But we are considering here as only Alpha divided by 2. I am confused on why 1-minus is not considered. Thankyou 😊
Hi Sekar, we are taking the t value based on alpha divided by 2 value only and not based on confidence level. This is the basic rule followed to find the t value.
Sir, is it common for the alpha values and then the risk factor 95%
Yes
What use of mean value (15.007)of observed measures
The difference between mean value (15.007) and reference value (15.000) will give the bias (0.007).This bias is required for calculation of other parameters
Dear sir, can you share zala format for bias study
Bias study format download link is given in description
r u taking any training class for MSA/SPC?
Yes. In my company I am giving training for employees.
Is tis enough for to crack qa qc interview sir
excuse me sir i have one doubt based on the bias value we can able find the error in instruments or inspector
Bias study is conducted for instrument only. It should be done by experienced inspector.
When will the subgroup value be 2?
As per BIAS, We can do only, 1 appraiser, 1 part, 1 parameter,1 place,
if I am adding one appraiser to take the same parameter & same part in the same place could it change the subgroup value is 2?
please clarify sir,
Thanks in advance
Yes, in Bias study the sub group size is based on number of appraiser only
@@newmechanicalmind8818 Thankyou sir
Hi,
I have been working in switchgear industry and lately our company has instructed to do MSA study. But we are mostly using instruments such as High voltage test instrument , low voltage test instrument , timing test instrument, snatch gap test instrument bla bla bla. Which MSA study fits for these kind of instruments, could you please assist?
Hi Mohammed, thank you for your comment.
Hope you are producing non automotive parts.
This MSA study is applicable for instruments which are used for checking the automotive parts.
And especially for physical (mechanical type) measuring instrument and gauges only where variations may come due to instrument and inspectors.
For electrical and electronic devices used to check the voltage, current , etc, i think calibration only sufficient.
Pls check with any other source / experts for re-confirmation.
Sir simple "t statistical =bias/ sigma b " what it's role in bias study
Hi Vignesh, if t statistics value is greater than significant t value, then the bias is more & will affect the measurement. If t statistics value less than significant t value, then there is very negligible bias only is there and measurement is OK. This point is not told in video. Thank you for asking this right question. Keep it up.
U r doing great job sir thanks for ur explanation.
Minitab msa pls upload with explanation
Dear Anbu, i will plan in future
What if we have tolerance spec is + side only ex : 9 /+0.05
Take 9 as reference value. But it is not advisable to take one sided tolerance for conducting bias study
Sir Nanga use pandra formula vanthu bias-[d2×sigma b /d2*(tv,1-alpha/2] enaku d2 value 3.078 entha table la irunthu eduthanganu therila nanga use pandra sub group size 10 sir
Sir and also one more doubt t staticala degrees of freedom illa confidence level illa then ethuku athai derive pannanum. Athuku straight ah tables la poi t value edukulamla
I have verified your doubt and giving my comment.
1) You are not using correct formula as per MSA manual
2) You are using d2* - 3.078 - value taken from Appendix C d2* table for sub group size 9.but you have to use 3.18 as per table (use our table - link given in description)
3) Degrees of freedom to be taken from Appendix C d2* table for No of sub group & sub group size. Then actual t value to be calculated based on that degrees of freedom from T distribution table (see our video again). Because that table has values for natural numbers (1,2,3...) only. But we have to find t value for fractions (example 10.8).
If you have further doubt. pls send separate comment (not replying to this comment). Then only i can easily view your comment and can reply
Sir can I get SOP for MSA
Hi Mr.Sushil, MSA manual only will explain about the studies in MSA. It is freely available in internet.
Sir, Very good explanations.Thanks for your time and efforts.
I have one doubt on Bais Excel format
In Excel format, you created a Histogram chart.
In the histogram chart, you had used some formulas to find out the
" Selecting no. of classes =" ( T24 Cell )
Could please clarify this formula and what is meant by
1.Selecting no.og classes ( T24 Cell)
2.FREQ.(W27)
3.CU. FREQ.(X27).
Thankyou once again.
Hi Samynathan,
1)(T24) Selecting no of classes is to decide the width of Histogram (if range is small, small width of histogram will come, if range is big, bigger histogram will come by formula, because range divided by no of classes will decide the width). Pls study the formula again.
2) (W27) frequency is used to count the number of values (between histogram's each column) for histogram's bar graph
3) (X27) cumulative frequency is to confirm that all values taken for histogram (total 15)
Hope now you understand.
@@newmechanicalmind8818 Thank you for your clarification
Formula la "1- alpha/2 " nu yen iruku. Nama "Alpha/2" vachithana sir t table check pandrom
1-alpha/2 = 95% = confidence level
alpha = risk = 5%
alpha / 2 = both side (+side & - side) 2.5% risk from reference value (middle)
so, t table will be checked by using both side risk ( two tail risk )
so both are different.
Hope i have cleared your doubt
Please share this ppt sir...
Hi Mani keep watching and try to make your own for better understanding