SPC And Concentricity

[William Johnson]

What is everyones opinion on doing a 30pc capability study on concentricity. Is it simular to True Position for CP,CPK'S.?I have a customer that wants 1.67cpk on concentricity they also want true position. I will only give them the cpk on each axis, because I need to know that the process is in control first. I have never given much thought to concentricity and cp,cpk. Our engineer says that I can report cp,cpk on concentricity, He also agrees that I can not do true positon but thinks concentricity will be ok to do.?????? What do you all think.

[Matthew D. Hoedeman]

Well, I think that you will run into the same issue with that as you would with TP since you don't know which direction each one varies from one to the other. But, others will tell you that you can do it. If there is no way to 'clock' the part (ie ID to OD on a tube) then it would work, but if it is clocked in any way (ie c-bore to through hole in a detail with datum edges) I don't think it would work any better than TP.

[#2#]

I agree with Matt. I don't mean to hi-jack but this is sort of related. Can we extend the question to GD&T measurements in general? TP, concentricity, roundness, flatness, perpendicularity etc. All of these are a method of comparing a part to a tolerance. The reported value for any of them does not indicate which direction a part deviates from nominal.

[bw_bob]

http://pcdmisforum.com/showthread.php?t=3632....Look at this thread


Oh justncredible were are you.......

[John Kingston]

http://pcdmisforum.com/showthread.php?t=3632....Look at this thread


Oh justncredible where are you.......

Shouldn't you be saying, "Oh he who shall not be named where are you......."

[cmmguy]

Ok, Here is how a product that I recommend(QCCalc) performs a variant of CPK called CPKu. It only calculates the upper side for single sided zero-based tolerances such as True Position, Runout, Concentricity.

White Paper on CPKu

Maybe this will help... If you can only do capability on bilateral toleranced features, that sure doesn't leave much.

[William Johnson]

Where do you get QCCalc at? I would like to investigate it. I can not go to your link. "NOT ALLOWED" DAMMM IT

[cmmguy]

Here is the file posted.

[img]http://www.*******.com/misc/*******/CPKSingle Sided tolerances.gif[/img]

[kbotta]

IF the part is not a feature of revolution (you can stop "anit-rotation"):
Remember - Concentricity is the median point, or mid point of "diametrically opposed points". Because you do have a cylindrical tolerance zone here, and you are inspecting each point individuly - I think concentricity actually lends itself better to SPC then position. Lets say your zone is 1mm. Your mid points are going to be nominaly "zero" +/- 0.5.
Don't you guys think this lends itself better?
Kev??
This would be per ASME 14.5M-1994, and you would not use the concentricity function in Pcdmis of course...

[cmmguy]

... Your mid points are going to be nominaly "zero" +/- 0.5....

I missed where you get the minus part of the tolerance from? Are you think rotational location from centerline?

[kbotta]

Yes. See attached...
kev

[cmmguy]

Yes. See attached...
kev

What would be Plus or Minus on one part would be in a different direction on the next part. I guess I am not following your example very well.

[bw_bob]

What would be Plus or Minus on one part would be in a different direction on the next part. I guess I am not following your example very well.

Diametric tolerance zone of .10........+/-.05.......has to fall in there somewere.

[kbotta]

potentialy. but That could be true of all studies. I am simply trying to communicate a way to inspect conc. with "spc" techniques. of course, you would have to look at individule points. that is what you would be looking at. An individule point. Times - how many? For the the entire surface. Just a thought...
kev

[kbotta]

Yes. In my 1st post I spoke about an anit-rotation feature. This would allow you view each point - the same each time.
kb