What's new
What's new

What is a typical defect per million rate in machining?

honda1988

Plastic
Joined
Oct 5, 2017
We run very high volumes. Some jobs are 2 million per year through vertical machining centers. What is a typical defect per million in a machining environment?

Thanks,
 
We run very high volumes. Some jobs are 2 million per year through vertical machining centers. What is a typical defect per million in a machining environment?

Thanks,

Depends on "who" you are and what you are making.

“I guess the question I'm asked the most often is: "When you were sitting in that capsule listening to the count-down, how did you feel?" Well, the answer to that one is easy. I felt exactly how you would feel if you were getting ready to launch and knew you were sitting on top of two million parts -- all built by the lowest bidder on a government contract.”
― John Glenn
 
like Gordon says, it depends.

I know a guy who was qc manager at a bumper factory. 3ppm scrap was average. injection molded part.
could you imagine machining a complete heat treated stainless valve body and only losing 3 per million? that would be insane
 
The scarier thought is now the reporting requirements are now becoming PPB (parts ber billion) as people were reaching low digits on the old PPM on machined parts.
How far we have come.

This kind of crazy ratio also depends on what you are counting.
Does a setup or changeover part count? Sometimes you know you will loose one or more and only count flaws once running.
Parts caught inside by your error proofing or out of spec parts shipped?
And volume, a 2 million part a day producer normally has a more stable process so lower numbers.

What matters is that you know your number and try to improve on it. Not benchmarking against others.
It's a continual race against yourself. Other shops/processes/products are not you.
2,10, 500, 2000 are all good numbers.
Bob
 
We run very high volumes. Some jobs are 2 million per year through vertical machining centers. What is a typical defect per million in a machining environment?

Thanks,

.
wouldnt you look at cost of
inspection
slower safer machining process
tooling
part
.
if rejected part a $1. problem or $1,000,000. problem ? for example SpaceX rocket a structural strut fails and the whole rocket is lost. the strut is far cheaper than the whole rocket but its failure caused everything to be lost.
.
definition of critical component
would anybody care if it fails ?
what problems does failure cause.
.
nothing worse than $1. part causing a car or automobile to not start and need to be towed to be repaired
 
my radio in my truck the plastic knob of volume and on/off is broke. maybe $1. part and the truck radio dont work. makes me think the $20,000 truck is low quality junk all cause what they saved $1. making the volume control
.
just saying cost against does anybody care if part fails ?
 
Also, are you making something of the tolerance of a washer, or the tolerance of an optical assembly?

Also, what is your base rate? If you're making a part with 8000 drilled holes, are you counting defects per million parts or holes? For that part 1000 defects per million parts would be amazingly good. 1000 defects per million holes would ensure that you never, ever make a good one.

Without context the rate is meaningless. Early-stage semiconductor manufacturing often runs a defect rate as high as 50% and that's commercially viable in that field.
 
company makes photographic paper. walmart complains every so often big defect on picture. paper not 100% inspected. found out roof leaks rain water getting on paper. put up tarps to drain water away. installed electronic scanner to look at paper moving over 1000 feet per minute quickly detect rain water defect. scanner was like $100,000. but well worth it rather that piss off customers
.
didnt matter defect rate was low it was pissing off the customers
 
A guy I knew who was a machinist but did not own his equipment. He rented time in a shop or had some type of arrangement with the owner. In a run of parts he would check the first, middle, and last part. That was it, he said it takes too much time to check every part. He did a few jobs for me before I started on my own. He was never between jobs, always busy.
 
A guy I knew who was a machinist but did not own his equipment. He rented time in a shop or had some type of arrangement with the owner. In a run of parts he would check the first, middle, and last part. That was it, he said it takes too much time to check every part. He did a few jobs for me before I started on my own. He was never between jobs, always busy.

When I worked for a medical implant manufacturer, bone screws, we made approx. 6 million bone screws a year of various types. Every process was validated to a CpK of 1.33 or less. This allowed us to check every 25th part or every 50th part depending on the process. We averaged less then 2% scrap on a year to year basis. No recalls in over 10 years and the FDA actually quit bugging us. Our internal regulatory group? Now that is a different story. Get your processes down pat and you need not worry so much.

Paul
 
I worked in an auto parts plant that made hydraulic valve lifters. In the 80's we started making them for Toyota. The quality standards were very high and we had to do a lot of work on our processes to meet the quality standards that Toyota demanded. It also helped the quality of our OEM parts.

Many years later our plant was the first in the USA to receive the Toyota Quality Award. Management sent a group of executives to Japan to receive the award. We had achieved a defect rate of two PPM, which was unheard of for a product like we made. Our executive team was so proud of the low PPM #. When they received the award the Toyota people asked our execs what we were going to do about the two parts that got through. They said it meant that two Toyotas would be sitting on the road broke down. Our execs said they were speechless. Nobody in the USA had achieved the quality levels that we had.

I have been retired from that plant for almost twelve years now and don't know what the PPM #'s are but our plant is the only one in the corporation that is still open in our area so I'm pretty sure it is good.
 
We run very high volumes. Some jobs are 2 million per year through vertical machining centers. What is a typical defect per million in a machining environment?

Thanks,

Back in "six sigma" days the goal was 99.9997% good parts. Frankly, you're not going to achieve that off a machining operation short of spending zillions. Inserts break, material has defects, and so on. But it would be achievable on an outgoing quality basis -- you'd have to detect the various failure modes, eliminate what you could, and scrap those few bad parts before they reach the customer.

On edit: BigB's plant actually achieved that 99.9997% Probably a pretty good story there of tracking down failure modes; as well as scrapping (or reworking) a few parts (like the ones where an insert broke prematurely) before Toyota ever saw them.
 
What you are talking about is the six sigma type of PRODUCTION. Buzz words are Cp and Cpk. Cp is a measure of how well the part conforms to the tolerance and Cpk is the same as Cp but also how well the part is in tolerance and centered about the nominal value. Cp is not normally used, instead Cpk is used. Example, a Cpk of 2 represents 3.4 defective parts per million. Note, defective means the part doesn't meet the tolerance specification, not that it necessarily is scrap.

The whole concept is rooted on having a design and a manufacturing environment CAPABLE of producing parts or assemblies consistently to engineering specifications. Note cost is not addressed directly but is inferred to in the lower scrap, rework customer complaint costs. The whole process is statistically based.

Step one is a design that is capable of being manufactured consistently to specifications. Engineers often find this very constraining. The design has to undergo a tolerance analysis to arrive at realistic tolerances. Critical dimensions are identified and others, such as those that fit air are very open. The critical dimensions are then those that are monitored. Standard shop run tolerances have no meaning. Also part of the design process is to select materials that readily available and can be consistently manufactured from. For example, a special alloy available from only one off shore supplier may not be realistic.

The design also has to be capable of being manufactured on available machinery and processes. As an example, a flat spring with many forms that have to be held to very tight tolerances that has to be heat treated after forming is not likely to meet capability requirements. Even one made from prehard can be difficult to make due to variable spring back. On the other hand, a 1/2" diameter brass rod 1" long from alloy 360 can be easily held to a tolerance of +-.005 on a Swiss lathe.

So now we have a design that can be made to achievable tolerances and we have machinery and processes that can product the parts and assemblies, now we need a way of monitoring the process. Often times these are referred to as Xbar and R charts. This is where control of the production is monitored. Samples of production are randomly selected from current production and checked for the control feature, be it dimensions, alloy, functional values such as spring rate or whatever; that is the agreed upon critical feature. These values are then plotted on the Xbar and R charts. Xbar refers to the average value and R is the range of values. The critical value is plotted on the charts and on both sides of the critical value are lines that represent half(usually) of the tolerance. The two additional lines are drawn the are the maximum tolerances.

As the data is plotted, the average deviation of center value will show up on the Xbar chart and drifting on the R chart.

More information on the net.

Six Sigma - Wikipedia
Process capability index - Wikipedia

Tom
 
Back in "six sigma" days the goal was 99.9997% good parts. Frankly, you're not going to achieve that off a machining operation short of spending zillions. Inserts break, material has defects, and so on. But it would be achievable on an outgoing quality basis -- you'd have to detect the various failure modes, eliminate what you could, and scrap those few bad parts before they reach the customer.

On edit: BigB's plant actually achieved that 99.9997% Probably a pretty good story there of tracking down failure modes; as well as scrapping (or reworking) a few parts (like the ones where an insert broke prematurely) before Toyota ever saw them.

Our PPM numbers were for parts leaving the plant. Of course there were the occasional bad parts that needed to be reworked or scrapped. When you are working in the millionths of an inch it is very difficult to have first time quality on every part you make.
 
Most normal people would think 99% a good target.
The classic poster is shown in the middle here.
Why a 99% Yield Is Not Good Enough - The Right Approach Consulting
Some things you just have to nail spot on.

Toyota did and continues to do a good job of educating US manufacturing facilities.
And this from a company that once built in the late 50's,,,,,well import junk which caused them to withdraw from the US market and regroup.
If they had not come back and "cleaned our clock" would we be where we are now?
We where the world masters of mass production, Japan became the master of mass production done right.
More strangely it was US citizens we sent there who led or lit the change.

I work in a place where 5S means "sort", "set in order", "shine", "standardize", and "sustain".
I get into arguments when I try to explain it means "seiri, seiton, seiso, seiketsu, and shitsuke" which is not quite the same.
Bob
 
Essentially, one American is responsible for changing the course of Japan's manufacturing: William Edwards Deming.

And this was after the Big 3 scoffed at him, even to the point of ridiculing his principles.

It wasn't but a few years later the Big 3 were beating Deming's door down needing help to improve their half-ass quality.

ToolCat


Sent from my iPhone using Tapatalk
 
Essentially, one American is responsible for changing the course of Japan's manufacturing: William Edwards Deming.

And this was after the Big 3 scoffed at him, even to the point of ridiculing his principles. . .

Years ago I did annual "Manufacturing Leadership Summits" as a session leader with Jim Womack -- who was probably the leading proponent of Toyota production methods at the time. If memory serves, he noted that the Chicago meat packing industry was the inspiration for Henry Ford's production line. And Henry Ford's production line the inspiration for "flow manufacturing" and the Toyota production system. So, basically we needed the Japanese to re-teach us lessons (Chicago/Ford and Deming) we ourselves had originated.

The few years I spent in the auto industry (mid 70's, some clients into the 80's), the worry was accountants being in charge. It only got worse with Wall St., hedge funds, and the like in charge.

We're not back to the glory days of dominating global auto manufacturing, but it does seem that Ford and GM make better products today than before. And if we're smart and stick to it, maybe we'll be a dominant factor in the next generations of personal transport; where computing and controls are one of the main values added.
 
...... but it does seem that Ford and GM make better products today than before......

That question posed by the Toyota guys above "They said it meant that two Toyotas would be sitting on the road broke down"
Now you will get the same from GM and Ford and you had better have a good answer ready.
If not they will send a team into your plant to "help" you until they are happy.
Bob
 








 
Back
Top