Reader Question: A longtime customer of ours is pushing for larger orders from us but also requires a PPAP to certify production. Can you explain this process and what it means for our machining process? 

PPAP (production part approval process) is the package you submit to prove, before the first production order ships, that your process can make the part not just once, but every time you press “cycle start.” Part of that package is capability, which is a number that says how well your process fits inside the tolerance. The reason a customer asks for it is simple: Capability lets them buy down risk. If you can show the process holds the feature with room to spare, they can stop inspecting every part and trust the process instead. As volume increases, the ability to scale quality control to match machine output is strained. Therefore, the tradeoff for larger, steadier work requires a high level of trust, which is why PPAP becomes a hurdle to prove a process will hold up with low risk of escapes.

Featured Content

Metalworking Fluid Tackles the Challenges of Titanium Machining READ MORE

35.BI-MU: Plenty of New Features for an Edition Rich in Thematic Insights READ MORE

WEBINAR: From Machine Data to Guided Action: How Modern Shops Are Closing the Execution Gap READ MORE

Statistical Process Control

The capability numbers, Cpk and Ppk, are built on the standard deviation of your process, which is just a measure of how much your parts vary from one to the next. We call this statistical process control (SPC). There are different methods to calculate each, which I will touch on another time, but understand that both represent repeatability and centering of a feature relative to its tolerance. One represents the short term (back-to-back pieces), and the other the long term (a population run over days or weeks). It means the math is a direct readout of part-to-part consistency. As you improve a process and limit variation, the number improves.

The PPAP requirement from your customer may say 1.00, 1.33, 1.67 or even 2.00. This is an indicator of how likely a scrap part could be produced and can be correlated to parts per million.

Normal process distribution and defect (parts per million). Source: Way of the Mill LLC

A customer may also state different capability requirements for different levels of features like “normal” versus “critical” features. As it turns out, manufacturers are usually presented with a double whammy because a feature critical to part function likely has a tight tolerance already and is then stacked with a capability requirement.

The change in machining between making a good part, and making a “capable” process, is how much of a tolerance you are permitted to consume. The capability requirement quietly takes a cut of your band before you make a chip. Perfectly centered in the tolerance, a Cpk of 1.33 leaves you about three quarters of the band to work in. 1.67 leaves you about 60%. In practice, you plan on losing a third to a half of the tolerance right off the top. As an example, a print may still read ±0.0020", but the number you get to machine to is closer to ±0.0012" because the process needs room to absorb deviation above and below the average to still be within spec. That reframes the entire job. Once capability takes half your band, you stop treating the tight number as a target to hit and start treating it as the tolerance the part always had to begin with.

On The Machine

At the spindle, that reduction in tolerance shows up everywhere. As an example, you may need to bore or ream a hole where you used to interpolate it with an end mill, because the interpolation’s variation is fine against the full tolerance and may be fatal against half of it. Capability is a property of the whole process; therefore, all sources of variation will require more care. The part must locate the same way every time, so you probe the fixture instead of trusting it. Tools get changed preemptively, on a schedule, rather than run until they are cooked. Runout you would have shrugged off now stacks up and matters. The machine gets a real warmup, because a capability run on a cold machine is a capability run you will fail. Every one of those is the same move: put bounds on a step you used to leave open, because anything you leave open shows up in the standard deviation, and the standard deviation is the number.

The place this gets most difficult is a one-sided tolerance, like a flatness callout. Say the print allows 0.025 mm. You may hold 0.020 mm, with a spread of maybe 0.004 mm across the run — in spec and reasonably well controlled. Every part you make would ship and function without complaint. Despite this, you can still fail capability, because 0.020 mm sits too close to the 0.025 mm limit for the math to pass you, even though nothing is wrong with the parts. That is the moment “good enough to ship” and “good enough to run day to day” stop being the same thing. The parts are fine. The process, measured honestly, is not centered where it needs to be to prove itself reliably. Reality says ship it. The math says not yet. Both are right and closing that gap is the work.

That customer dangling the bigger order isn’t betting on a new skill you have to go learn. They already know your parts are good, or they wouldn’t be asking. What PPAP changes is the burden of proof. You now have to demonstrate, in numbers, the consistency you’ve been running all along. The good news is the work is familiar. Control the variation, hold your center, put bounds on every step that feeds the tolerance. Do that, and the capability number takes care of itself.

Do you have a machining question? Ask the expert. John Miller leans on more than a decade of industry experience to answer machining questions from MMS readers. Submit your question online at mmsonline.com/MillersEdge.