Machine Shop Yield & Sampling Plan

[Quality Engineer66]

Hi All,

This is my first post here. I need some help in choosing what standard should I use for sampling. I am going to work in Machine shop where we produce lots of parts. We don't have any acceptance quality level set up. If I use Mil std 105e then for example if I have a lot size of 25 & I inspect 13 parts & found 3 were wrong then how will I know whether quality level is met.

Also what would be my first pass yield? Is it (13-3)/13 or (13-3)/25?

Looking for a great advice from the members.

 

[Stijloor]

Hi All,

This is my first post here. I need some help in choosing what standard should I use for sampling. I am going to work in Machine shop where we produce lots of parts. We don't have any acceptance quality level set up. If I use Mil std 105e then for example if I have a lot size of 25 & I inspect 13 parts & found 3 were wrong then how will I know whether quality level is met.

Also what would be my first pass yield? Is it (13-3)/13 or (13-3)/25?

Looking for a great advice from the members.

Can someone help?

Thank You!!

Stijloor.

 

[Stijloor]

Hi All,

This is my first post here. I need some help in choosing what standard should I use for sampling. I am going to work in Machine shop where we produce lots of parts. We don't have any acceptance quality level set up. If I use Mil std 105e then for example if I have a lot size of 25 & I inspect 13 parts & found 3 were wrong then how will I know whether quality level is met.

Also what would be my first pass yield? Is it (13-3)/13 or (13-3)/25?

Looking for a great advice from the members.

Bump.

First time poster; can someone give him a hand?

Thank you!!

 

[Jim Wynne]

Hi All,

This is my first post here. I need some help in choosing what standard should I use for sampling. I am going to work in Machine shop where we produce lots of parts. We don't have any acceptance quality level set up. If I use Mil std 105e then for example if I have a lot size of 25 & I inspect 13 parts & found 3 were wrong then how will I know whether quality level is met.

Welcome to the Cove.

It seems like you might be going about this backwards; you decide what your AQL should be (given a particular sampling plan) and that plan will tell you the sample size (based on lot size) and give you an accept/reject quantities.

Also what would be my first pass yield? Is it (13-3)/13 or (13-3)/25?

In most cases, if you were to find three defectives in a 13-piece sample from a 25-piece lot, you would inspect the other 12 pieces, so your yield would be from the lot, not from the sample.

 

[20110525R]

Hi All,

This is my first post here. I need some help in choosing what standard should I use for sampling. I am going to work in Machine shop where we produce lots of parts. We don't have any acceptance quality level set up. If I use Mil std 105e then for example if I have a lot size of 25 & I inspect 13 parts & found 3 were wrong then how will I know whether quality level is met.

Also what would be my first pass yield? Is it (13-3)/13 or (13-3)/25?

Looking for a great advice from the members.

Well, I am new to this forum myself, but I have 39 years in manufacturing with an Industrial Engineering and Manufacturing Technology background with 27 of those years in Quality Assurance. I am now Retired and am an ISO 9001 and AS9100 Consultant.
The scarey part here is, are you saying your company is manufacturing parts without a Quality Inspection System, or a Quality Management System in place? Is there a QC/QA group in the facilty or are you the only one there?
Most companies (DoD especially) now use Zero Based Sampling plans and Six Sigma, 5S, SPC or some other system and from what you described, you reject the entire lot with 3 out of tolerance parts of a sample size of 13.
I have to ask some questions to further understand your situation, but if you haven't read your company Policies and Procedures (QA Manual) or the company doesn't have them, then you have no QA/QC system, I for one, would not want to be one of your customers if this is the case. I suggest you talk to your QA/QC Manager, if the company has one, if not, then you have a serious Quality problem with your company. By the way, Mil-Std 105 is obsolete as DoD doesn't want an "Acceptable Quality Level", they want good parts, and I would hope all of your customers do also.
I would be happy to elaborate later, but would need to know more about what is in place with your company.

 

[Jim Wynne]

Well, I am new to this forum myself, but I have 39 years in manufacturing with an Industrial Engineering and Manufacturing Technology background with 27 of those years in Quality Assurance. I am now Retired and am an ISO 9001 and AS9100 Consultant.
The scarey part here is, are you saying your company is manufacturing parts without a Quality Inspection System, or a Quality Management System in place? Is there a QC/QA group in the facilty or are you the only one there?
Most companies (DoD especially) now use Zero Based Sampling plans and Six Sigma, 5S, SPC or some other system and from what you described, you reject the entire lot with 3 out of tolerance parts of a sample size of 13.
I have to ask some questions to further understand your situation, but if you haven't read your company Policies and Procedures (QA Manual) or the company doesn't have them, then you have no QA/QC system, I for one, would not want to be one of your customers if this is the case. I suggest you talk to your QA/QC Manager, if the company has one, if not, then you have a serious Quality problem with your company. By the way, Mil-Std 105 is obsolete as DoD doesn't want an "Acceptable Quality Level", they want good parts, and I would hope all of your customers do also.
I would be happy to elaborate later, but would need to know more about what is in place with your company.

As far as the obsolescence 105E is concerned, it might have been superseded by the government, but it's still in use otherwise in many companies. The nice thing about math is that it's never obsolete.
I'm pretty sure that before 105E was supplanted the government (and all other customers) wanted good parts. ANSI/ASQ Z1.4 (the current version of the MIL standard) has accept/reject criteria just like its predecessor. The point that needs to be made here, imo, is that lot sampling plans on the producer end are of dubious value and should be obviated by process control. Nonetheless, there may be times when suppliers are faced with questionable lots and will (a) need to do sampling and (b) make a decision based on a rational assessment of risk. That's what sampling plans are for.

 

[20110525R]

As far as the obsolescence 105E is concerned, it might have been superseded by the government, but it's still in use otherwise in many companies. The nice thing about math is that it's never obsolete.
I'm pretty sure that before 105E was supplanted the government (and all other customers) wanted good parts. ANSI/ASQ Z1.4 (the current version of the MIL standard) has accept/reject criteria just like its predecessor. The point that needs to be made here, imo, is that lot sampling plans on the producer end are of dubious value and should be obviated by process control. Nonetheless, there may be times when suppliers are faced with questionable lots and will (a) need to do sampling and (b) make a decision based on a rational assessment of risk. That's what sampling plans are for.

Jim,
My point wasn't whether or not the Mil-Std was obsolete or the math, but if Quality Engineer66 has no accept/reject criteria, then who has been responsible for the metrics and processes before he came to work at the facility to validate the Processes and Outcomes? Rational assessment of Risk is directly related to criticallity of the part (FMEA in most cases) and if he's just going to work at a machine shop has he talked to QA/QC folks yet? Does the company have an Inspection System in place? If up till now, sampling has not been used, then how has this company been evaluating the product. My guess is, there is something in place, procedures/system, to guide him in assessing the Processes and Outcomes and he needs to review that system (i.e. talk to QA/QC folks) before making an accept/reject criteria decision.
Mike

 

[Quality Engineer66]

Thanks all for the help. I am afraid to say that we don't have any standards to follow. We do have the check list but no sampling plan. What sampling technique should I use in the machine shop? Also how to calculate the yield based on the sampling plan.

 

[20110525R]

Thanks all for the help. I am afraid to say that we don't have any standards to follow. We do have the check list but no sampling plan. What sampling technique should I use in the machine shop? Also how to calculate the yield based on the sampling plan.

Guess I am a little confused by your question now on calculating yield. Are you using SPC? If not, why? Have you had training in Statistical Process Control (SPC)? You are a Quality Engineer correct? Sorry, not to be down on you, but your questions now that I re-look at this thread are pretty much basic SPC.
Maybe time for some "Light Reading" on W. Edwards Deming, Joseph M. Duran, Walter A. Shewhart and Kaoru Ishikawa. Okay, I jest, it's not light reading, pretty boring stuff (wait till you sit through the Deming Tapes), but definitely worth the study for your field. If you have had this training, then I apologize for the inference here, but what you are asking here is pretty much normal stuff for a QAS/QAR or QAE to do.

 

[Quality Engineer66]

I am fresh graduate from college & working as a quality engineer. I am given the role to set the standards for the machine which doesn't have any system as its a new addition to the company. Can you please help me how to choose a sampling plan & how to know about the yield?