A
aoe110
I meet the problem for dimension 70 +0.3 which needs to study Pp/ppk.
How to define this dimension as one sided or two sided spec?
Thanks.
How to define this dimension as one sided or two sided spec?
Thanks.
well, it depends...these things are almost never as simple as they may appear.
for this feature, is 70 a physical bound? in other words, is it possible for values to be less than 70? if not possible then the tolerance is one sided.
otherwise it's two sided.
BUT if the target is 70 and not 71.5, then Pp is calculated differently then one would do for a two sided symetrical toelrance.
If you can tell us more about the feature we can provide more specific advice...
for this feature, is 70 a physical bound? in other words, is it possible for values to be less than 70? if not possible then the tolerance is one sided.
otherwise it's two sided.
BUT if the target is 70 and not 71.5, then Pp is calculated differently then one would do for a two sided symetrical toelrance.
If you can tell us more about the feature we can provide more specific advice...
A
aoe110
This dimension is for the outer diameter of a steel pulley.
The target is 70.0, and the spec on drawing is 70 +0.3.
So how to work it?
Thanks.
The target is 70.0, and the spec on drawing is 70 +0.3.
So how to work it?
Thanks.
Maybe I make everything too simple. The specification is 70.0 - 70.3. Though you can get values less than 70, they all FAIL! This is where the confusion comes in. You can look at this problem as a two-sided specfication, but a perfectly targeted process with very low variability still gives you a minimum of 50% defective by the mathematics.
IF and ONLY IF you change the dimension to be 70.15 +/- 0.15 would you have a real two sided-specification. Though this gives a false target.
Since I am not a huge fan of the Ppk/Cpk paradigm, I just report out the PpU in this problem and explain that a target of 70 can lead to upwards of 50% defective for undersize.
As Statistical Steven stated, the specification zone is 70-70.3. Period. It may infer that the design preference is a target of 70, but that has no relationship whatsoever with the process and its tolerance or process target. It may be normal about the center-point, or it may be a uniform distribution at 75% of the tolerance, there may be a natural stop that causes the distribution to be Weibull or beta skewed towards (or even away from) 70...it is totally a function of the process, not the design target. Sorry for the designer's luck....but, we can thank them for inferring their desires anyway.
Rameshwar25
Quite Involved in Discussions
But my question is while calculating Ppk, we should take target of 70mm or 71.5 mm. In both cases the PPk value will be different because:
1. if we take a target of 70, closer the value of mean to 70, better the PPk value will be. eg. value of mean equal to 70.2 will show better PPk rather than 71.3.
2. if we take a target of 71.5, closer the value of mean to 71.5 better the PPk value will be. eg. value of mean equal to 70.2 will show poor PPk rather than 71.3.
Kindly help.
Rameshwar
1. if we take a target of 70, closer the value of mean to 70, better the PPk value will be. eg. value of mean equal to 70.2 will show better PPk rather than 71.3.
2. if we take a target of 71.5, closer the value of mean to 71.5 better the PPk value will be. eg. value of mean equal to 70.2 will show poor PPk rather than 71.3.
Kindly help.
Rameshwar
Rameshwar
What formula are you using? Ppk is min(usl-mean/3S;mean-lsl/3s). Ppk does not depend on the target.
Exactly, the point of Cpk is whether your distribution is centered within the tolerance. And it primarily applies to normal distributions of independent data. If your pulley is machined, it should not be independent, so that would also not make the Cpk correct or meaningful - even though it can be algebraical calculated.
M
Mr.Happy
Hi all,
I also have a question concerning one sided tolerance.
What if the specs ask a hardness off 63 HV minimal ? How should I put this in a SPC calculation?
Mr.Happy
I also have a question concerning one sided tolerance.
What if the specs ask a hardness off 63 HV minimal ? How should I put this in a SPC calculation?
Mr.Happy
Similar threads
