Evaluating software that measures Ppk, Cmk, Cpk, PPI, etc.

C

Cemm Mex

I am a newbie at this, and I am currently evaluating software that measures PPK, CMK, CPK, PPI, etc...

I the Old forums I've found most definitions but CMK does not appear. Anybody have a clue what it is?

By the way, does anybody have a nice, simple, cheap piece of software I can use for this purpose? So far I've only seen InfinityQS (www.infinityqs.com)
 

Proud Liberal

Quite Involved in Discussions
First off, InfinityQS is the best software on the market to doing actual SPC (statistical PROCESS control not statistical PART control). If you are serious about statistics, it is worth the effort required to impliment. But if simple is what you want, try QualityAmerica's SPC-PC IV (I used this one for more than 10 years prior to learning about InfinityQS.

I think that Cmk is another designation for Cpm. If so, Cpm differs from Ppk/Cpk in that rather than using the nominal (middle) of the tolerance zone as the "perfect" part, the dimension is targeted at some other value to accommodate processes that are skewed by nature. Examples would include:
  • Tool wear (you would target your process toward one end of the tolerance and allow the tool wear to move throughout a greater range)
  • Physical limitations - a die closing, parts locating in a punch press against a stop, etc
  • A chemical process that uses up a reagent or catalyst
  • Most GD&T designations (flatness, straightness, parallelism, angularity, perpendicularity, roundness, .....) that designate a limit are targeted toward zero.
What type of processes are you involved with?
 
C

Cemm Mex

Our Process

We are in the Maquila buisiness. Manufacture car parts, specifically automotive harnesses.
 
D

Darius

Sorry for the comment, but you are starting from the end, InfinityQS, I tink it's the most complete SPC software for now.

But for capability index almost any SPC or statistics software could do the job, it's almost a standard to obtain pp,ppk, cp,cpk and some of them cpm, cpmk and their conterparts with total variation estimate, others some more sofisticated indexes like non-parametical.

Remember that any index is just an estimate (because of the estimates of variation, the number of data in the sample, even the non normality that most of the capability index take for granted, the correctness of the specs, if is one sided or not, etc.), I my self find them usefull, but I always look for run chart (or spc chart) and histogram.

:bigwave:
 
Top Bottom