Hi,

We manufacture small plastic parts using injection-moulding machine. There are quite big batches of production (average 50 000 parts per lot). Please advise me how we should calculate Cmk factor:
1. When it should be calculated (once a year or may be after each new connection of injection-moulding machine with the tool (mould))?
2. How many samples must be taken (I assume that 1 sample always consists of 5 parts)?
3. How frequently samples must be taken?

Thanks in advance.

I do not a lot of experience in Machine Capability studies, but I took this from http://www.autotrain.org - Link was: /auto-home/main.html

Capability - Machine Capability

A Machine Capability Study is not, strictly speaking, part of Statistical Process Control, as it does not detect the presence or absence of special causes. Instead, it provides a "snapshot" of one particular component of a process, often a machine (hence the title). This could well be of use in the context of, for example, carrying out a trial run on a machine before agreeing to buy it (like test driving a car).

A Machine Capability Study is essentially a way of establishing the capability of the component/machine by assessing how it operates:
over a very short period of time
in an ideal environment

By an ideal environment, we mean one in which:
the machine is set as close as possible to its optimum values
it is being operated by an experienced member of staff
it is isolated from the normal working environment (and thus unlikely to be affected by special cause variation)
the study takes place over a very short period of time

Typically, about 50 outputs of the process would provide data for a Machine Capability Study. This may vary depending on the outputs. For example, an hour's output of nails might be more than enough, whereas producing 50 back axles might not be practical.

Regards,

Don

I'm not sure what a "Cmk" is, either. Does "m" stand for machine? If it does, then Don probably has the right conditions listed. However, if it is a standard process capability study, you must establish that special causes have been eliminated before you calculate the capability index. Otherwise, you get into questions of things like normality and its impact on percent of product out of spec. Perhaps could describe what you are trying to do in a little more detail.