Machine Capability Studies on CNC Lathes

DWPOZE

Registered Visitor
I'm currently conducting capability studies on some CNC Lathes. I was going to have them run me a 100 consecutive pieces. The question I have should I allow them to offset the machine during the study.

Appreciate your help
 
G

galcantar

What do you mean to "allow them offset the machine" ?

Capability study main objective is to "know" common causes related to equipment; no adjustments, no operator changes, no measurement method- devices changes, material from same lot, etc.

Pieces needs to be consecutive and numered to check process stability
 
J

JaySturgeon

Agree with gal

What do you mean to "allow them offset the machine"?

Capability study main objective is to "know" common causes related to equipment; no adjustments, no operator changes, no measurement method- devices changes, material from same lot, etc.

Pieces needs to be consecutive and numered to check process stability


I am not a staistical guru but IMHO, you need 120 parts before you have a staistical sound sampling of the process. Those parts can not be influenced in any way to determine true capability.

Where did automotive get their 30 pc. therory is what I'd like to know.

:bonk:
 
A

Atul Khandekar

DWPOZE said:
I'm currently conducting capability studies on some CNC Lathes. I was going to have them run me a 100 consecutive pieces. The question I have should I allow them to offset the machine during the study.

If you mean adjusting or centering the process, then NO.

If the CNC has any built-in program/cycles for automatic tool nose radius compensation etc, then that should be considered a part of the process and should be allowed.
 
J

JaySturgeon

Question

Atul.............

Any comments regarding automotive selected sample size for cap. studies?

:cool:
 
K

ksanders

DWPOZE said:
I'm currently conducting capability studies on some CNC Lathes. I was going to have them run me a 100 consecutive pieces. The question I have should I allow them to offset the machine during the study.

Appreciate your help


I agree with the others. In our shop, the term 'offset' is another name for adjust. Adjustments within a study are not allowed.
 

DWPOZE

Registered Visitor
I understand that no adjustments (offsetting) should be made to the machine. But if set the machine up to the mean and run 100 consecutive pcs it will tell you how capable the machine is without operator making adjustments. But when your trying to hold a .001" total tolerance, adjustments will have to be made sooner or later during production
 
S

Sam

"I understand that no adjustments (offsetting) should be made to the machine. But if set the machine up to the mean and run 100 consecutive pcs it will tell you how capable the machine is without operator making adjustments. But when your trying to hold a .001" total tolerance, adjustments will have to be made sooner or later during production"

Two different processes.
- Initial process capability study; 100/125 pc sample size, run parts with no adjustments to set-up, measure oarts, plot on chart, evaluate, make adjustments and re-run study.

- on-going process control; evaluate charts periodically, make changes when process shows evidence of gion out of control.
 
J

JaySturgeon

On The Money

Two different processes.
- Initial process capability study; 100/125 pc sample size, run parts with no adjustments to set-up, measure oarts, plot on chart, evaluate, make adjustments and re-run study.

- on-going process control; evaluate charts periodically, make changes when process shows evidence of gion out of control.


Sam is right on the money. Are you being required for on-going SPC for the characteristic you've been discussing. On a CNC shooting for nominal is not a biggy. If you have tool compensation running you should be ok for the future SPC requirements, provided you have any.

We are a metal stamper. I continually have customers wanting Cpk numbers for holes and slots. In metal stamping punches (to put in the hole or slot) are always ground to the high side of the tolerance for wearability purposes, thus your Cpk numbers will always suck. I have done pretty well at convnincing most customeras to look at the Cp numbers and not the Cpk number.

Oh well......and they ask me why I drink!

:thedeal:
 
A

Atul Khandekar

JaySturgeon said:
Any comments regarding automotive selected sample size for cap. studies?
Like Sam said:

1.For initial process capability studies, it is necessary to run the process through all its natural cycles of variation, without adjustments. This is why a continuous sampling of 100-125 pieces is recommended.

2.You should not assess capability before ensuring that the process is stable and in control and 30 individual readings may not be sufficient to determine stability and control.

According to AIAG SPC manual:
Assessment of process capability begins after control issues in both XBar and R charts have been resolved (special causes identified, analyzed, corrected and prevented from recurring), and the ongoing control charts reflect a process that is in statistical control, preferably for 25 or more subgroups.

and also:
Generally, 25 or more subgroups containing about 100 or more individual readings give a good test for stability, and, if stable, good estimates of the process location and spread.

-Atul.
 
Top Bottom