How To Conduct Machine Capability Of An Electrical Discharge Machining (edm)?

[maximillen010]

can anyone assist me on how am i going to do machine capability of an EDM machine? what is the procedure? i'm new to this area. need assists on this matter.

 

[Govind]

can anyone assist me on how am i going to do machine capability of an EDM machine? what is the procedure? i'm new to this area. need assists on this matter.

Unfortunately, there are not many literatures available in the machine capability study. Let me try to recollect from my memory. Machine Capability study is to measure the variations attributed to machine only and not machine and process combined. The main difference between the 2 studies (Machine Capability & Process Capability) is “data collection”. In the machine capability study, the consecutive parts are measured. No changes to setting should be done. In the Process Capability study, a sample of parts are collected in a periodic frequency. The calculation formula remains same for Machine capability study. The Machine Capability indices are identified as Cm and Cmk.

I remember a thumb rule in this data collection.
1 outlier is allowed and removed in a size of 30 consecutive parts measured.
2 outliers are allowed and removed in a size of 50 consecutive parts measured.
3 outliers are allowed and removed in a size of 100 consecutive parts measured.

If the outliers exceed this number, the data collected until that point is discarded and a fresh count is started from 1,2,3..
Cm, Cmk is expected at > 1.66

If I find an authentic reference, I will add to this post at a later date.

Regards,
Govind.

 

[Md.Saddam Hossain]

i need to know details about cmk. For which parameter, what will be the best cmk for buying a new machine. Can any one give me best answer.

 

[bobdoering]

If you wish to do a machine capability, you should take a standard part - perhaps a 2" by 2" by 1" depth plunge and run it 125 consecutive times into a common base metal. Measure the widths and the depths and chart each one individually in time ordered sequence. You should see each dimension get smaller and smaller as the probe wears. This will give you a wear rate function. What does that mean for capability? First, your output should not be random, or it is out of control. Second, wear rate should be a function, you can likely develop from the linear regression of the individual charts. That will prove the process is non-normal, and capability calculations are not applicable. BUT, the process is predictable based on the function.

You can find Cm and Cmk in Bothe's book "Measuring Process Capability", but they won't apply...or at least shouldn't.