Re: PPAP for Beginners - Where do I start?
This is somewhat dated, but I think the essentials are the same: https://elsmar.com/APQP/
This is somewhat dated, but I think the essentials are the same: https://elsmar.com/APQP/
N
ncwalker
Re: PPAP for Beginners - Where do I start?
Continuing on.....
#7: Control Plan - The process flow diagram list the steps you are going to do. This begets the PFMEA, where those steps are then analyzed for potential failure, which then begets the Control Plan which details HOW you are going to run the process to not have any of the bad things in the PFMEA happen. It's a table with columns indicating the step, what the operation is, is it making a key characteristic. How are you checking it? With what? How often? Here's where the minefields are: SOME people put everything in here. Every machine setting, readout, everything. I've seen these things get longer than 100 pages. The problem with this is - if in 6 months you bring on a new chiller, and suddenly your shop water changes temperature and you have to recalibrate all your settings - guess what? You have to re-PPAP because you changed the process. So don't overdo this. It it OK to put something in there like "refer to work instruction XXXX" for settings so you don't paint yourself into a corner. BUT ... what your SQ will require is sadly very inconsistent from customer to customer. So this one is always a bit of a game. Another mine field - it is very tempting to put in your control plan "I check one part out of every lot." I advise against that, though it is common and widely accepted as OK. Instead, put how many parts that is. If you are making cookies and you say early on in the process you are checking a 40 lb sack of flour for moths as part of your control plan, how many cookies is that? 100? 1000? It will help YOU flush out if your checks are statistically sound for frequency by doing this. And give you a better idea of what percentage is checked in each shipment. Important stuff if you have a quality spill. And, of course, every step must match the PFD and the PFMEA. Also - don't be afraid to try and go with a "family" control plan. Say you are making potato chips and the process is the same for BBQ and Sour Cream chips except for the bag of seasoning you dump in. You don't need to to a control plan for both. WARNING: Not every SQ sees the sense in this. Again - a lot will depend on what will fly with your SQ but this is a good question to ask.
#8: Measurement Systems Analysis: If there is one thing that bites people in manufacturing in the butt, it is this. Everyone assumes that this purchased gage works. You may go buy a wonderful digital micrometer that can spit out a half a micron. But if the nature of your part requires your tech to balance on one foot and reach around measure in an awkward position - guess what? That 0.0005 claim on the micrometer box won't mean diddly. Even worse - you enlist some special automated gage company to make you a whiz bang automated gage..... You better put words in that contract that it can pass a Gage R&R on ACTUAL PARTS. Not mock ups, not masters, REAL parts. I've seen this burn people over and over. Do the Gage R&Rs. Do them right. You should do them when a process seems unstable - it may be the gage and not the process that has the problem. Be VERY careful of Gage R&R forms you find on the 'net. Many of them don't work because someone copy/pasted a value where a formula should have been. Type in the example from the MSA Book or other known, good source into your software. If you don't get the same results - guess what? Your software is broken. I have seen professional software NOT calculate this correctly. No lie.
#9: Dimensionals. This is where your bubbled print from above comes into play. This is where you fully lay out n parts and say "look man, they're all in spec." How many is "n"? usually 3 to 5, but these days, people are asking for 6. Also, when you submit your PPAP to your customer, you must number and indicate in that PPAP shipment that THESE are the parts you did your dimensional checks on. It may be worth it to ship them separately. More often than I would care to admit PPAP parts get mistakenly built with normal production. Why send them special? So your customer can verify whatever measurements he's going to take in incoming inspection and check them against what you got. This is CRITICAL. Because guess what? HIS measurement system may be totally whack. And he may say a part is out of spec that you say is in spec. If that day comes in production, he is the customer. And that means that YOU are wrong until you prove yourself right which takes time. That whole time he will be billing you for sorting. And do you think when it comes out that you were right all along he will say "sorry, here, let me pay that sort bill for my mistake...." Ha ha ha. Cost of doing business, buddy. The lesson? These layout parts are GOLDEN. Insist he verifies them if he can. The other minefield? If this is a PPAP for a replacement tool, or duplicate process, or you are taking over business, expect to ALSO do an A to B comparison. So that's 12 parts. 6 old, 6 new with the differences highlighted. You should expect to have to do this and have that ready.
11 Capability Studes: This is Cpk/Ppk. The stats and they are usually only needed on the key characteristics. Work this out with your customer as to what they are. Minefields? Number one is pass through characteristics. You need to identify as early as you can what these are. And they are features YOU are putting in that your customer does not use, but HIS customer does. He "passes them through". You are making the brass ends of a garden hose, he is making the hose and putting your ends on it. He's not going to check if the fitting fits a spigot. That will be discovered when the consumer buys the hose and it doesn't fit his spigot at home. And you will be responsible for making your customer look bad in front of HIS customer. That's not a fun day for you, so push to get the pass throughs identified and if it is at all possible, you should poka-yoke them. They SHOULD be in the DFMEA he made that he should give you laying out the risks of the design. But he never gives you the DFMEA because he waits until the last minute himself to do it. Doesn't matter, when that part doesn't work at his customer, the only thing that will matter on that day is how badly you stink. Second minefield - if you do NOT have any key characteristics, PICK SOME. Pick a few dimensions that are going to be hard to make, or useful in knowing the process is behaving. If he doesn't ask, you don't have to submit a capability study, but DO IT for YOURSELF. Why? So you really know what your process is doing.
#12: Qualified Lab Documents - Man, this one is annoying. I'm not sure anyone REALLY understands what this means. But here's what works and will get you PPAP'd. Put together a list of all your measuring equipment and what it is in your lab. Toss in your ISO, TS, or QS certs. Maybe a blurb about your trick calibration database and what software you are using and your latest CMM certification and this will sail right through. It's not really enforced. (Don't want to argue this point with anyone, I am stating what I have seen work).
Continuing on.....
#7: Control Plan - The process flow diagram list the steps you are going to do. This begets the PFMEA, where those steps are then analyzed for potential failure, which then begets the Control Plan which details HOW you are going to run the process to not have any of the bad things in the PFMEA happen. It's a table with columns indicating the step, what the operation is, is it making a key characteristic. How are you checking it? With what? How often? Here's where the minefields are: SOME people put everything in here. Every machine setting, readout, everything. I've seen these things get longer than 100 pages. The problem with this is - if in 6 months you bring on a new chiller, and suddenly your shop water changes temperature and you have to recalibrate all your settings - guess what? You have to re-PPAP because you changed the process. So don't overdo this. It it OK to put something in there like "refer to work instruction XXXX" for settings so you don't paint yourself into a corner. BUT ... what your SQ will require is sadly very inconsistent from customer to customer. So this one is always a bit of a game. Another mine field - it is very tempting to put in your control plan "I check one part out of every lot." I advise against that, though it is common and widely accepted as OK. Instead, put how many parts that is. If you are making cookies and you say early on in the process you are checking a 40 lb sack of flour for moths as part of your control plan, how many cookies is that? 100? 1000? It will help YOU flush out if your checks are statistically sound for frequency by doing this. And give you a better idea of what percentage is checked in each shipment. Important stuff if you have a quality spill. And, of course, every step must match the PFD and the PFMEA. Also - don't be afraid to try and go with a "family" control plan. Say you are making potato chips and the process is the same for BBQ and Sour Cream chips except for the bag of seasoning you dump in. You don't need to to a control plan for both. WARNING: Not every SQ sees the sense in this. Again - a lot will depend on what will fly with your SQ but this is a good question to ask.
#8: Measurement Systems Analysis: If there is one thing that bites people in manufacturing in the butt, it is this. Everyone assumes that this purchased gage works. You may go buy a wonderful digital micrometer that can spit out a half a micron. But if the nature of your part requires your tech to balance on one foot and reach around measure in an awkward position - guess what? That 0.0005 claim on the micrometer box won't mean diddly. Even worse - you enlist some special automated gage company to make you a whiz bang automated gage..... You better put words in that contract that it can pass a Gage R&R on ACTUAL PARTS. Not mock ups, not masters, REAL parts. I've seen this burn people over and over. Do the Gage R&Rs. Do them right. You should do them when a process seems unstable - it may be the gage and not the process that has the problem. Be VERY careful of Gage R&R forms you find on the 'net. Many of them don't work because someone copy/pasted a value where a formula should have been. Type in the example from the MSA Book or other known, good source into your software. If you don't get the same results - guess what? Your software is broken. I have seen professional software NOT calculate this correctly. No lie.
#9: Dimensionals. This is where your bubbled print from above comes into play. This is where you fully lay out n parts and say "look man, they're all in spec." How many is "n"? usually 3 to 5, but these days, people are asking for 6. Also, when you submit your PPAP to your customer, you must number and indicate in that PPAP shipment that THESE are the parts you did your dimensional checks on. It may be worth it to ship them separately. More often than I would care to admit PPAP parts get mistakenly built with normal production. Why send them special? So your customer can verify whatever measurements he's going to take in incoming inspection and check them against what you got. This is CRITICAL. Because guess what? HIS measurement system may be totally whack. And he may say a part is out of spec that you say is in spec. If that day comes in production, he is the customer. And that means that YOU are wrong until you prove yourself right which takes time. That whole time he will be billing you for sorting. And do you think when it comes out that you were right all along he will say "sorry, here, let me pay that sort bill for my mistake...." Ha ha ha. Cost of doing business, buddy. The lesson? These layout parts are GOLDEN. Insist he verifies them if he can. The other minefield? If this is a PPAP for a replacement tool, or duplicate process, or you are taking over business, expect to ALSO do an A to B comparison. So that's 12 parts. 6 old, 6 new with the differences highlighted. You should expect to have to do this and have that ready.
11 Capability Studes: This is Cpk/Ppk. The stats and they are usually only needed on the key characteristics. Work this out with your customer as to what they are. Minefields? Number one is pass through characteristics. You need to identify as early as you can what these are. And they are features YOU are putting in that your customer does not use, but HIS customer does. He "passes them through". You are making the brass ends of a garden hose, he is making the hose and putting your ends on it. He's not going to check if the fitting fits a spigot. That will be discovered when the consumer buys the hose and it doesn't fit his spigot at home. And you will be responsible for making your customer look bad in front of HIS customer. That's not a fun day for you, so push to get the pass throughs identified and if it is at all possible, you should poka-yoke them. They SHOULD be in the DFMEA he made that he should give you laying out the risks of the design. But he never gives you the DFMEA because he waits until the last minute himself to do it. Doesn't matter, when that part doesn't work at his customer, the only thing that will matter on that day is how badly you stink. Second minefield - if you do NOT have any key characteristics, PICK SOME. Pick a few dimensions that are going to be hard to make, or useful in knowing the process is behaving. If he doesn't ask, you don't have to submit a capability study, but DO IT for YOURSELF. Why? So you really know what your process is doing.
#12: Qualified Lab Documents - Man, this one is annoying. I'm not sure anyone REALLY understands what this means. But here's what works and will get you PPAP'd. Put together a list of all your measuring equipment and what it is in your lab. Toss in your ISO, TS, or QS certs. Maybe a blurb about your trick calibration database and what software you are using and your latest CMM certification and this will sail right through. It's not really enforced. (Don't want to argue this point with anyone, I am stating what I have seen work).
Last edited:
N
ncwalker
Re: PPAP for Beginners - Where do I start?
No problem. The "blue books" are a good resource. But they are a LOT to swallow. I'm trying to keep it light, and yet point out where what is supposed to happen and what really happens differ. (6 more elements to go, coming right up....)
No problem. The "blue books" are a good resource. But they are a LOT to swallow. I'm trying to keep it light, and yet point out where what is supposed to happen and what really happens differ. (6 more elements to go, coming right up....)
N
ncwalker
Re: PPAP for Beginners - Where do I start?
#13: Appearance Approval Report - This is like arguing with your spouse over what to wear, except there's a FORM! Seriously, if you are lucky, you are producing something that doesn't even need this (like a muffler) and if you are unlucky, you are producing something that does, (like a seat). And the kiss of death is if you are producing something that is visible, and some other guy is producing the thing that mates up to it - like you are making the seat and this other guy is making the door panel and he's on the other side of the world and your stuff has to match. Good luck. This is VERY customer specific and there are two pitfalls. The first is how much definition do you get? So the customer wants your product "to match." That's what he wants. Well, DEFINE this? Suddenly it's not so easy. You start asking questions in good conscience to get more clarity. And that's the trap. MOST LIKELY you are a subject matter expert on the object and the customer is not. Again I will go with car seats as an example. You probably know all about weave, and dye curing, and thread count. Your customer is an automotive engineer and doesn't know the ins and outs. As you push him for definition you have to be careful of your tone and how hard you push. Because if you start making the customer feel ignorant, there is a risk he will start making stuff up that "sounds good to him" so he doesn't feel ignorant. Yeah, we are supposed to be data driven, but this is a real risk. A LOT of design engineers are bright young men and women out of college who haven't made an object in production ONCE. So their knowledge of expected variability comes from what they believe to be true with little experience and what is copy/pasted from earlier designs which may stink. Gently steer this one, folks. If you have some measurement method that works for controlling appearance on whatever you make and you can sway this into "the way it's done" at your customer - BONUS !! Second pitfall - and this applies to any characteristic, really, but especially to these touchy/feely ones. This is the key point: in the event of a problem at your customer that may involve your product, it is less important that you be right and more important that you be not wrong first. That SOUNDS the same, but it isn't. You get a call from your customer saying "Hey man, I think I have a problem. Could be your thing, but it could be this other thing over here ....." You do NOT wait for him to make up his mind. You go full battle rattle and put a pretty document together that (truthfully) shows why you have your act together and it is not likely your thing. And you get it in your customers hands ASAP. Outgoing inspection data, traceability data, you show him you know about all your parts in the supply chain. Because if there are several suppliers who MAY be responsible for the problem, the focus is going to be on the LAST guy to do this professionally. And THAT guy is going to paying for all the sorts and expedites while the root causing is going on. Be not wrong FIRST. THEN, do your homework and check your stuff, because if you ARE the problem, it will be found. We (the customers) are pretty good at root cause. But usually, when the spill first hits, there's some days of panic where all the crazy demands that aren't data driven are made. And the key to avoiding that is being not wrong first. THEN get your ducks in a row. Don't make stuff up to avoid blame. The secret is have your act together and be able to show this in a neat, organized, PROMPT way.
#14: Layout Samples - Remember those 6 fully measured and marked parts from 9 Dimensionals above? Yep. These are those. They have their own section.
#15: Master Sample - Someone needs to hang on to the "golden part" when the process was new so that a year from now when you are asked "was it always like this?" You can say "Yep, here's an early part." Many times, your customer will want one at his location. But the fact is, storage space costs money. Most likely, they will say "you keep it." One can argue that with all the measuring, we have effectively digitized the "master part" and the physical part itself is redundant. Yeah, no. Keep it. Keep 3 if you have room. It could mean the difference between a back charge or not.
#16: Checking Aids - Here is your opportunity to shine. What is this? This is the non-standard gaging stuff. Hard gages, CMM fixtures. I have seen this mostly be helpful with gears and splines. But it should be considered with any kind of hard gage - like one for checking hard tubing that's bent. Mearuements where without the special clamping fixture or nest or whatever, it can become an arguing point over what the dimension actually is due to difficulties in measuring. If this is a new launch, odds are you are getting the tooling (and gaging) paid for by your customer. And if you are making one of these special gages because it is the best way to measure conglomerate dimensions quickly and efficiently, you go to you customer and say "Hey, you should fund us for two of these and we will build them at the same time. Because it really helps with the measuring and then you can have one at your location should you need to check our stuff." That shows foresight and it protects YOU. Why? Because if your customer does have trouble with your part and starts checking it the best way he can - a questionable measurement without this check fixture and he finds a problem, the customer's gage is always right. Let me say it again - the customers gage is always right. Yes, you will send in your engineer with this fixture in tow, and sit at his location and say "See, it's ok, you just need this fixture." However long that takes, you will be sorting and expediting. And yep, you may get stuck with the bill for that, your fault or not.
#17: Customer specific requirements - Blue books are no help here, this is up to your customer. A couple of pitfalls, at least in automotive, the CQI audits. Especially 9, 11 and 12. These are for heat treat, coatings and platings. If you do these things, get an audit done. In general, you need one once a year and MOST customers are OK if you have one within the year, even if it is not on their part. We won't go in to the fact that really only approved auditors may conduct these. It SHOULD be that way. In reality it's OK if you have a current one. Another pitfall is Early Containment. Called many things by many customers. In words, it means you are going to be extra careful with the new process until you get some miles on it. Some customers want an entire separate Control Plan for this. It is the same control plan as the NORMAL one, except the frequency of inspection is jacked up. Sometimes to 100% on key characteristics, so get ready for that. If you are really clever, when you create your Control Plan and you are filling out the frequency of inspection you put BOTH entries here. For example: you plan to normally check a diameter once an hour, but 100% during Early Containment. In your frequency of inspection box you put "1/hr (EC 100%)" and cover both with the same plan. The early containment inspection cannot be on the line. You have to have a special area away from the line where these checks are done. Set aside some floor space so you can do this in an organized way, and get used to it because this requirement is not going away as far as I can see in automotive. How long do you do this Early Containment? Depends on your customer, but typical is 30 days without a breach. What I mean is: expect to check 100% your KPCs for 30 days. And if the EC process catches ONE bad one, the clock resets. It's really the same thing as Controlled Shipping 1, except you are doing it because you have a new process instead of because you goofed. The mechanics are the same. If you have a special area properly equipped, this is easier.
#18: The Part Submission Warrant (PSW). It's a bad idea to stray from the one in the blue books. This is a pretty universal form. And it has it's pitfalls. Primarily, fill it out correctly, but here are common problems:
a) Not a problem, but a pet peeve. The books say the weight is to be completed to 4 decimal places. What they mean is 4 significant figures, which isn't the same thing. You sell me the front of a bulldozer, and tell me it is 14,230 kg. That's fine. I don't need to know it is 14,234.8831 kg.
b) Comments. If there is a dimension out that is not critical, like a fillet radius is supposed to be R20 max and you are a little generous and it is actually R21. Make a note in the comments that there is a need for deviation. This may start a lot of arguments here, because many believe the PPAP parts shouldn't be deviated. But there are lots of components and processes where correcting that initial tooling can do more harm than good. You may get lucky and the customer may say "eh, it's ok." But if he changes his mind in 6 months and you don't have a blurb on your warrant, you may get hit for it later.
c) There is a blurb about 3/4s of the way down the warrant in the DECLARATION that says "I further affirm that these samples were produced at the production rate of blank/blank hours." That is exactly what it reads like. You put in the number of parts and the number of hours it took for your PPAP run. Example: 400 pcs/8 hrs. It doesn't have to be per 8 hrs, it just has to divide out to the (rough) production rate you made your samples at. This is to certify that you did not make jewelers samples and baby them through your process. But that they (reasonably) went through the process at intended production speed so that they actually DO represent what we expect from the process. THIS IS NOT A RUN AT RATE It is also not an agreement to any kind of rate at all. If you think this, you are wrong. But be warned, there are lots of people out there who DO and will try and stick you with the number you put in here. Don't fall for it. Fight it if a customer comes at you with that. READ the sentence, it could not be any more clear, and yet .... ARRRGG.
d) IMDS. This goes on the warrant also. And if your industry does not partake of this, count yourself lucky. In reality, it is a really cool system. But too often done as a pencil whipping exercise, or not given diligence. It's really pretty easy to do, if you take the time to learn what the system wants. If you don't, well, your boss will be asking you again and again why all that tooling money is being held up. Different customers have different levels of what they are willing to accept. But here are some pitfalls:
- Just because you submitted it, does NOT mean your customer approved it. They have to log in and approve it. And many times it's an additional duty for someone who really doesn't want it but got stuck with it. Usually resulting in a panicked flurry at the last moment to get it all right. So here is what will help...
- Semicomponent vs Component - if what you sell, what you are PPAPing, is going to have some follow on operation done to it that will CHANGE THE WEIGHT of the part, like machining, you are selling a SEMICOMPONENT. If you don't submit it this way, your customer cannot adjust the weight for HIS submission and it will be rejected.
- Also, IN the IMDS database there are two weights. One is called "calculated" and the other "measured." Customers are getting picky about this and checking. The calculated weight needs to match what is on the print, calculated from the 3D math and the density. The measured weight needs to be what the part weighed when you made it and THIS MUST MATCH THE WARRANT.
- And last, not only is the part number in there, but the revision is as well. THAT needs to match the revision you submitted to and the warrant.
And that's it. It seems daunting, the first time you put a PPAP together. But if you really are doing the right stuff from a quality perspective, there shouldn't be any "new work" other than the packaging.
#13: Appearance Approval Report - This is like arguing with your spouse over what to wear, except there's a FORM! Seriously, if you are lucky, you are producing something that doesn't even need this (like a muffler) and if you are unlucky, you are producing something that does, (like a seat). And the kiss of death is if you are producing something that is visible, and some other guy is producing the thing that mates up to it - like you are making the seat and this other guy is making the door panel and he's on the other side of the world and your stuff has to match. Good luck. This is VERY customer specific and there are two pitfalls. The first is how much definition do you get? So the customer wants your product "to match." That's what he wants. Well, DEFINE this? Suddenly it's not so easy. You start asking questions in good conscience to get more clarity. And that's the trap. MOST LIKELY you are a subject matter expert on the object and the customer is not. Again I will go with car seats as an example. You probably know all about weave, and dye curing, and thread count. Your customer is an automotive engineer and doesn't know the ins and outs. As you push him for definition you have to be careful of your tone and how hard you push. Because if you start making the customer feel ignorant, there is a risk he will start making stuff up that "sounds good to him" so he doesn't feel ignorant. Yeah, we are supposed to be data driven, but this is a real risk. A LOT of design engineers are bright young men and women out of college who haven't made an object in production ONCE. So their knowledge of expected variability comes from what they believe to be true with little experience and what is copy/pasted from earlier designs which may stink. Gently steer this one, folks. If you have some measurement method that works for controlling appearance on whatever you make and you can sway this into "the way it's done" at your customer - BONUS !! Second pitfall - and this applies to any characteristic, really, but especially to these touchy/feely ones. This is the key point: in the event of a problem at your customer that may involve your product, it is less important that you be right and more important that you be not wrong first. That SOUNDS the same, but it isn't. You get a call from your customer saying "Hey man, I think I have a problem. Could be your thing, but it could be this other thing over here ....." You do NOT wait for him to make up his mind. You go full battle rattle and put a pretty document together that (truthfully) shows why you have your act together and it is not likely your thing. And you get it in your customers hands ASAP. Outgoing inspection data, traceability data, you show him you know about all your parts in the supply chain. Because if there are several suppliers who MAY be responsible for the problem, the focus is going to be on the LAST guy to do this professionally. And THAT guy is going to paying for all the sorts and expedites while the root causing is going on. Be not wrong FIRST. THEN, do your homework and check your stuff, because if you ARE the problem, it will be found. We (the customers) are pretty good at root cause. But usually, when the spill first hits, there's some days of panic where all the crazy demands that aren't data driven are made. And the key to avoiding that is being not wrong first. THEN get your ducks in a row. Don't make stuff up to avoid blame. The secret is have your act together and be able to show this in a neat, organized, PROMPT way.
#14: Layout Samples - Remember those 6 fully measured and marked parts from 9 Dimensionals above? Yep. These are those. They have their own section.
#15: Master Sample - Someone needs to hang on to the "golden part" when the process was new so that a year from now when you are asked "was it always like this?" You can say "Yep, here's an early part." Many times, your customer will want one at his location. But the fact is, storage space costs money. Most likely, they will say "you keep it." One can argue that with all the measuring, we have effectively digitized the "master part" and the physical part itself is redundant. Yeah, no. Keep it. Keep 3 if you have room. It could mean the difference between a back charge or not.
#16: Checking Aids - Here is your opportunity to shine. What is this? This is the non-standard gaging stuff. Hard gages, CMM fixtures. I have seen this mostly be helpful with gears and splines. But it should be considered with any kind of hard gage - like one for checking hard tubing that's bent. Mearuements where without the special clamping fixture or nest or whatever, it can become an arguing point over what the dimension actually is due to difficulties in measuring. If this is a new launch, odds are you are getting the tooling (and gaging) paid for by your customer. And if you are making one of these special gages because it is the best way to measure conglomerate dimensions quickly and efficiently, you go to you customer and say "Hey, you should fund us for two of these and we will build them at the same time. Because it really helps with the measuring and then you can have one at your location should you need to check our stuff." That shows foresight and it protects YOU. Why? Because if your customer does have trouble with your part and starts checking it the best way he can - a questionable measurement without this check fixture and he finds a problem, the customer's gage is always right. Let me say it again - the customers gage is always right. Yes, you will send in your engineer with this fixture in tow, and sit at his location and say "See, it's ok, you just need this fixture." However long that takes, you will be sorting and expediting. And yep, you may get stuck with the bill for that, your fault or not.
#17: Customer specific requirements - Blue books are no help here, this is up to your customer. A couple of pitfalls, at least in automotive, the CQI audits. Especially 9, 11 and 12. These are for heat treat, coatings and platings. If you do these things, get an audit done. In general, you need one once a year and MOST customers are OK if you have one within the year, even if it is not on their part. We won't go in to the fact that really only approved auditors may conduct these. It SHOULD be that way. In reality it's OK if you have a current one. Another pitfall is Early Containment. Called many things by many customers. In words, it means you are going to be extra careful with the new process until you get some miles on it. Some customers want an entire separate Control Plan for this. It is the same control plan as the NORMAL one, except the frequency of inspection is jacked up. Sometimes to 100% on key characteristics, so get ready for that. If you are really clever, when you create your Control Plan and you are filling out the frequency of inspection you put BOTH entries here. For example: you plan to normally check a diameter once an hour, but 100% during Early Containment. In your frequency of inspection box you put "1/hr (EC 100%)" and cover both with the same plan. The early containment inspection cannot be on the line. You have to have a special area away from the line where these checks are done. Set aside some floor space so you can do this in an organized way, and get used to it because this requirement is not going away as far as I can see in automotive. How long do you do this Early Containment? Depends on your customer, but typical is 30 days without a breach. What I mean is: expect to check 100% your KPCs for 30 days. And if the EC process catches ONE bad one, the clock resets. It's really the same thing as Controlled Shipping 1, except you are doing it because you have a new process instead of because you goofed. The mechanics are the same. If you have a special area properly equipped, this is easier.
#18: The Part Submission Warrant (PSW). It's a bad idea to stray from the one in the blue books. This is a pretty universal form. And it has it's pitfalls. Primarily, fill it out correctly, but here are common problems:
a) Not a problem, but a pet peeve. The books say the weight is to be completed to 4 decimal places. What they mean is 4 significant figures, which isn't the same thing. You sell me the front of a bulldozer, and tell me it is 14,230 kg. That's fine. I don't need to know it is 14,234.8831 kg.
b) Comments. If there is a dimension out that is not critical, like a fillet radius is supposed to be R20 max and you are a little generous and it is actually R21. Make a note in the comments that there is a need for deviation. This may start a lot of arguments here, because many believe the PPAP parts shouldn't be deviated. But there are lots of components and processes where correcting that initial tooling can do more harm than good. You may get lucky and the customer may say "eh, it's ok." But if he changes his mind in 6 months and you don't have a blurb on your warrant, you may get hit for it later.
c) There is a blurb about 3/4s of the way down the warrant in the DECLARATION that says "I further affirm that these samples were produced at the production rate of blank/blank hours." That is exactly what it reads like. You put in the number of parts and the number of hours it took for your PPAP run. Example: 400 pcs/8 hrs. It doesn't have to be per 8 hrs, it just has to divide out to the (rough) production rate you made your samples at. This is to certify that you did not make jewelers samples and baby them through your process. But that they (reasonably) went through the process at intended production speed so that they actually DO represent what we expect from the process. THIS IS NOT A RUN AT RATE It is also not an agreement to any kind of rate at all. If you think this, you are wrong. But be warned, there are lots of people out there who DO and will try and stick you with the number you put in here. Don't fall for it. Fight it if a customer comes at you with that. READ the sentence, it could not be any more clear, and yet .... ARRRGG.
d) IMDS. This goes on the warrant also. And if your industry does not partake of this, count yourself lucky. In reality, it is a really cool system. But too often done as a pencil whipping exercise, or not given diligence. It's really pretty easy to do, if you take the time to learn what the system wants. If you don't, well, your boss will be asking you again and again why all that tooling money is being held up. Different customers have different levels of what they are willing to accept. But here are some pitfalls:
- Just because you submitted it, does NOT mean your customer approved it. They have to log in and approve it. And many times it's an additional duty for someone who really doesn't want it but got stuck with it. Usually resulting in a panicked flurry at the last moment to get it all right. So here is what will help...
- Semicomponent vs Component - if what you sell, what you are PPAPing, is going to have some follow on operation done to it that will CHANGE THE WEIGHT of the part, like machining, you are selling a SEMICOMPONENT. If you don't submit it this way, your customer cannot adjust the weight for HIS submission and it will be rejected.
- Also, IN the IMDS database there are two weights. One is called "calculated" and the other "measured." Customers are getting picky about this and checking. The calculated weight needs to match what is on the print, calculated from the 3D math and the density. The measured weight needs to be what the part weighed when you made it and THIS MUST MATCH THE WARRANT.
- And last, not only is the part number in there, but the revision is as well. THAT needs to match the revision you submitted to and the warrant.
And that's it. It seems daunting, the first time you put a PPAP together. But if you really are doing the right stuff from a quality perspective, there shouldn't be any "new work" other than the packaging.
Last edited:
N
ncwalker
Re: PPAP for Beginners - Where do I start?
And one closing comment, because a lot of this "PPAP stuff" is about the measurement systems.
If you are using a CMM, then the calibration certificate of your CMM is not, IS NOT, IS NOT as good as a Gage R&R.
What I mean is if your CMM calibration guy has rolled through your shop and done his thing and given you a piece of paper that says "the volumetric accuracy of this CMM is 3 microns" that does not mean the CMM is measuring everything to an accuracy of 3 microns. And NOT a Gage R&R.
Here's the proof - we have ALL looked at a CMM report and had someone say or agree to "Hey, the diameter looks a little funny, let's take more probe hits...."
Why would that make a difference if there wasn't something to do with the PROGRAM that could affect the result? Why would you be given acceleration and probe speed options if the only thing that mattered was the calibration? Why all the talk about using different tips?
Further, the more complicated measurements are subject to lots of internal calculations. Every one of those reduces the significant figures of the result. I don't care how many decimal places you have the thing spitting out, that's just a formatting statement, NOT an indication of significance.
So how do you do a Gage R&R? Like every other gage. Now, you CAN do a reduced check, and not run the numbers on EVERY measurement (often in the hundreds). With a well thought out scheme such as at least Gage R&R:
1) The most important dimension to the customer.
2) The dimension with the tightest tolerance.
3) The dimension that is the "most constructed" in the CMM. Like something involving the centerline of a probed cylinder that needs a lot of internal math to spit out the answer.
4) The biggest feature.
5) The smallest feature.
6) The true position that is reference to the datums that are "farthest" from the primary DRF.
You get the idea - exercise the thing.
Having said all that, the maddening thing is this....
I have seen WAY too often an SQ accept the calibration AS the Gage R&R. Blows me away. So.... if you want to try and see if that ship sails, I am letting you know it just may.
And one closing comment, because a lot of this "PPAP stuff" is about the measurement systems.
If you are using a CMM, then the calibration certificate of your CMM is not, IS NOT, IS NOT as good as a Gage R&R.
What I mean is if your CMM calibration guy has rolled through your shop and done his thing and given you a piece of paper that says "the volumetric accuracy of this CMM is 3 microns" that does not mean the CMM is measuring everything to an accuracy of 3 microns. And NOT a Gage R&R.
Here's the proof - we have ALL looked at a CMM report and had someone say or agree to "Hey, the diameter looks a little funny, let's take more probe hits...."
Why would that make a difference if there wasn't something to do with the PROGRAM that could affect the result? Why would you be given acceleration and probe speed options if the only thing that mattered was the calibration? Why all the talk about using different tips?
Further, the more complicated measurements are subject to lots of internal calculations. Every one of those reduces the significant figures of the result. I don't care how many decimal places you have the thing spitting out, that's just a formatting statement, NOT an indication of significance.
So how do you do a Gage R&R? Like every other gage. Now, you CAN do a reduced check, and not run the numbers on EVERY measurement (often in the hundreds). With a well thought out scheme such as at least Gage R&R:
1) The most important dimension to the customer.
2) The dimension with the tightest tolerance.
3) The dimension that is the "most constructed" in the CMM. Like something involving the centerline of a probed cylinder that needs a lot of internal math to spit out the answer.
4) The biggest feature.
5) The smallest feature.
6) The true position that is reference to the datums that are "farthest" from the primary DRF.
You get the idea - exercise the thing.
Having said all that, the maddening thing is this....
I have seen WAY too often an SQ accept the calibration AS the Gage R&R. Blows me away. So.... if you want to try and see if that ship sails, I am letting you know it just may.
N
ncwalker
Re: PPAP for Beginners - Where do I start?
Sorry for my ramble above. I also noted some of cdj100s questions weren't directly answered.
No. That means which elements of PPAP must be submitted. The 18 I mentioned above. If it is a brand new process, the will request most of them. If you are adding capacity - duplicating an existing process, they probably will only want a few items. Your customer SQ will tell you this.
The PPAP Levels are arbitrary levels someone assigned to what is in the PPAP. That's a little goofy. They are (paraphrased):
Level 1 - Warrant Only or Warrant With Appearance Approval.
Level 2 - Warrant with Samples and Limited Data
Level 3 - Warrant with Samples and Complete Data (this is the FULL Submission)
Level 4 - Warrant and Other Customer Requirements.
Level 5 - Warrant and Samples with Data Reviewed at the Supplier
I don't understand why the Full (most complete) is in the middle and not the end. And 99.999% of the submission will be Level 3 or Level 4.
You either, as a customer, want everything (Level 3) or you want a partial submission. And in the case of the partial, you are telling the supplier what you want. That's Level 4.
What's level 2? What does "Limited Data" mean? If I don't want a full, I'm going to tell you what I want, and that's Level 4. So to me, Level 2 is totally unnecessary.
And Level 1? The only time I want JUST a warrant is if everything else was OK and your warrant was bad. But that's more a replacement warrant. Level 1 is unnecessary.
Level 5? Only ONCE have I had a supplier not wish to send me information citing proprietary reasons. And he didn't want to send (understandably) his detailed prints. He was totally OK sending the data on my key dimensions, everything but his detailed component drawings. And we agreed to this. Why doesn't level 4 cover this too?
Here is what you do. If you are giving him a complete submission, mark it level 3. If you aren't, mark it level 4. I'd be stunned if were ever rejected following that convention. I've never seen it.
The ideal situation is for your PPAP to be complete and signed full before the next step. For example - you make pistons, you sell to an engine builder, who sells to an OEM. If everything went perfectly, you would be Fully PPAP'd - meaning your customer would sign your PSW to him and mark it FULL approval. Then, with your approved parts, he would put his PPAP together to HIS customer.
In reality because of the extreme pressures to get things to market quickly, there are always things that are not quite done. To further complicate things, when a customer says the requirements are met can be unclear. Some may not sign anything full until Run at Rates are completed, which is usually done slightly AFTER SORP (Start of Regular Production) when the system is being exercised. Some may wait for in vehicle testing to be complete. Because of this, you can be given "Interim" approval. This means the parts are usable, and shipping can being in limited quantities, but there are some items holding up PPAP. In most cases these items are non critical. Sometimes a critical thing like a Run at Rate can cause it. Frequently in a new launch, everyone is reluctant to order lots of parts until they are tested. And you CAN'T do the run at rate because you are making a couple hundred at a time, which isn't a good test. One would think the circumstances would be more clear for interim vs. full than they are, but the truth is financial aspects factor in to it. In quality we like to think all decisions are data driven, then we have to go live in the real world.
So yes, your PPAP can be at Interim with some further actions to the extent that your customer is willing to agree, it is their call. Document things clearly.
GM, Ford and Chrysler got together and laid out the APQP system and named it because they realized that their different requirements we adding cost. That's the point - standardize things. What they laid out in the "Blue Books", the AIAG manuals isn't horrible. I have my complaints, but all in all it could be a lot worse. Every automotive OEM I deal with OUTSIDE of the "Big 3" has realized the sense in this except for one, who really goes their own way.
Some may CALL it by different things. But for 90% of the stuff, it matches the APQP Process. They all have their little tweaks they want. Many of them are reasonable.
The two biggest holes I see in what the AIAG has laid out is:
1) Early Containment - which really isn't enforced like you will be led to believe. It generates a LOT of data. And everyone wants it their way. That's something that REALLY needs standardization, because at the end of the day it's really just reporting scrap during the launch.
2) Run at Rates/Capacity - I have found in the automotive industry the general understanding of this and the metrics needed is pretty weak. A LONG time ago, when I was in the appliance industry, this was handled much better than what I see in the automotive industry, in general.
For these two things, it's really going to be customer specific. And until you get used to your customer, you are going to be redoing a lot of confusing sheets.
Sorry for my ramble above. I also noted some of cdj100s questions weren't directly answered.
No. That means which elements of PPAP must be submitted. The 18 I mentioned above. If it is a brand new process, the will request most of them. If you are adding capacity - duplicating an existing process, they probably will only want a few items. Your customer SQ will tell you this.
The PPAP Levels are arbitrary levels someone assigned to what is in the PPAP. That's a little goofy. They are (paraphrased):
Level 1 - Warrant Only or Warrant With Appearance Approval.
Level 2 - Warrant with Samples and Limited Data
Level 3 - Warrant with Samples and Complete Data (this is the FULL Submission)
Level 4 - Warrant and Other Customer Requirements.
Level 5 - Warrant and Samples with Data Reviewed at the Supplier
I don't understand why the Full (most complete) is in the middle and not the end. And 99.999% of the submission will be Level 3 or Level 4.
You either, as a customer, want everything (Level 3) or you want a partial submission. And in the case of the partial, you are telling the supplier what you want. That's Level 4.
What's level 2? What does "Limited Data" mean? If I don't want a full, I'm going to tell you what I want, and that's Level 4. So to me, Level 2 is totally unnecessary.
And Level 1? The only time I want JUST a warrant is if everything else was OK and your warrant was bad. But that's more a replacement warrant. Level 1 is unnecessary.
Level 5? Only ONCE have I had a supplier not wish to send me information citing proprietary reasons. And he didn't want to send (understandably) his detailed prints. He was totally OK sending the data on my key dimensions, everything but his detailed component drawings. And we agreed to this. Why doesn't level 4 cover this too?
Here is what you do. If you are giving him a complete submission, mark it level 3. If you aren't, mark it level 4. I'd be stunned if were ever rejected following that convention. I've never seen it.
The ideal situation is for your PPAP to be complete and signed full before the next step. For example - you make pistons, you sell to an engine builder, who sells to an OEM. If everything went perfectly, you would be Fully PPAP'd - meaning your customer would sign your PSW to him and mark it FULL approval. Then, with your approved parts, he would put his PPAP together to HIS customer.
In reality because of the extreme pressures to get things to market quickly, there are always things that are not quite done. To further complicate things, when a customer says the requirements are met can be unclear. Some may not sign anything full until Run at Rates are completed, which is usually done slightly AFTER SORP (Start of Regular Production) when the system is being exercised. Some may wait for in vehicle testing to be complete. Because of this, you can be given "Interim" approval. This means the parts are usable, and shipping can being in limited quantities, but there are some items holding up PPAP. In most cases these items are non critical. Sometimes a critical thing like a Run at Rate can cause it. Frequently in a new launch, everyone is reluctant to order lots of parts until they are tested. And you CAN'T do the run at rate because you are making a couple hundred at a time, which isn't a good test. One would think the circumstances would be more clear for interim vs. full than they are, but the truth is financial aspects factor in to it. In quality we like to think all decisions are data driven, then we have to go live in the real world.
So yes, your PPAP can be at Interim with some further actions to the extent that your customer is willing to agree, it is their call. Document things clearly.
GM, Ford and Chrysler got together and laid out the APQP system and named it because they realized that their different requirements we adding cost. That's the point - standardize things. What they laid out in the "Blue Books", the AIAG manuals isn't horrible. I have my complaints, but all in all it could be a lot worse. Every automotive OEM I deal with OUTSIDE of the "Big 3" has realized the sense in this except for one, who really goes their own way.
Some may CALL it by different things. But for 90% of the stuff, it matches the APQP Process. They all have their little tweaks they want. Many of them are reasonable.
The two biggest holes I see in what the AIAG has laid out is:
1) Early Containment - which really isn't enforced like you will be led to believe. It generates a LOT of data. And everyone wants it their way. That's something that REALLY needs standardization, because at the end of the day it's really just reporting scrap during the launch.
2) Run at Rates/Capacity - I have found in the automotive industry the general understanding of this and the metrics needed is pretty weak. A LONG time ago, when I was in the appliance industry, this was handled much better than what I see in the automotive industry, in general.
For these two things, it's really going to be customer specific. And until you get used to your customer, you are going to be redoing a lot of confusing sheets.
N
ncwalker
Re: PPAP for Beginners - Where do I start?
And I am sorry guys, one MORE closing comment.
Lots of people have been saying "It's really up to what the customer wants." Or some such, myself included.
That said - I do want to make this point. The customer is NOT always right.
Consider: There is A LOT of material in the blue books. We all here in the Cove won't agree on what they say. It stands to reason that the "customer" won't either. Heck, some of them are Cove members.
Consider: There many times when the guy from the customer who is reviewing your PPAP is NOT a subject matter expert. The guy maybe spent his career in frame and body and took this job as supplier quality for seats and interior. So he may not, because of inexperience, even know what the right questions to ask are. If he is willing to learn, he will ask. If he wants to prove he is smarter than everyone (yes, there are some like that out there) he will saddle you with requirements.
Do not forget that the REAL POINT of APQP and PPAP (and I firmly believe in it when done right) is to ensure that design intent is made with a stable process that has the capacity to produce what is needed. All the "stuff" is just demonstration of that. It's just sound statistics and documentation that you are being careful.
What's the point? Well, many times the customer will come down with "his way of doing things" and suddenly you have a new form, or process. And if you have more than one customer (you should) you have to switch gears and it causes confusion on your floor. You have to do it THIS way for X and THAT way for Y. And every time you have to do things differently, the risk of making a bad part increases. And your costs go up.
Why do the customers do it to us? Some say it's because every customer thinks his way is the best way. I like to think we aren't that arrogant. I think what drives this is if the customer isn't clear on what you are doing, he will make a requirement that is FAMILIAR to him. Because that's easier on him. And "his stuff" becomes ANOTHER requirement.
Now, if you are in a little shop - this can be VERY intimidating. Here comes a dude from a major auto manufacturer, this is the big leagues. But most of them (there are a few bad apples, to be sure) are just regular folks just like you. You don't have to be frightened or intimidated. MOST of them really love what they do, and will share what they know. And MOST of them will also LISTEN.
I have found that if my shops "way" of doing this stuff meets the requirements, and I go to the customer and present "my shops way" in a clear and organized way. That at the very least:
1) Clearly shows I meet the requirements of what's needed.
2) The only real difference between "his stuff" and "my stuff" is formatting.
3) Bring up the point that every time we do something different moving parts through the shop, the risk of a mistake greatly increase.
I have found that the customer will say "Good points, approved to do it your way."
The bottom line is, if you have something in your shop THAT WORKS you absolutely CAN go to the customer and say "I'd really prefer to do it the way that works best in MY shop...." and you will find that they will listen. The blue books are a GUIDE, not a mandate. If you have your act together, you will find customer "requirements" melt away to "suggestions." And this leaves you to streamline, get efficient and cost effective.
Example: If your customer asks you about Early Containment and you give him a shaky, sketchy answer as to what it is, internally he's going to be thinking "what did my purchasing department get me into, better nip this in the bud real quick." And externally he is going to go "this is a REQUIREMENT."
But if he asks and you are johnny-on-the-spot and say. "Yes Sir! We are all over Early Containment. Come see our containment area on the floor. And here's our standard data collection system. And here's how we decide what goes on it." And it's organized and streamlined, your customer is going to say "You guys have it licked. Just do what you're doing."
And I am sorry guys, one MORE closing comment.
Lots of people have been saying "It's really up to what the customer wants." Or some such, myself included.
That said - I do want to make this point. The customer is NOT always right.
Consider: There is A LOT of material in the blue books. We all here in the Cove won't agree on what they say. It stands to reason that the "customer" won't either. Heck, some of them are Cove members.
Consider: There many times when the guy from the customer who is reviewing your PPAP is NOT a subject matter expert. The guy maybe spent his career in frame and body and took this job as supplier quality for seats and interior. So he may not, because of inexperience, even know what the right questions to ask are. If he is willing to learn, he will ask. If he wants to prove he is smarter than everyone (yes, there are some like that out there) he will saddle you with requirements.
Do not forget that the REAL POINT of APQP and PPAP (and I firmly believe in it when done right) is to ensure that design intent is made with a stable process that has the capacity to produce what is needed. All the "stuff" is just demonstration of that. It's just sound statistics and documentation that you are being careful.
What's the point? Well, many times the customer will come down with "his way of doing things" and suddenly you have a new form, or process. And if you have more than one customer (you should) you have to switch gears and it causes confusion on your floor. You have to do it THIS way for X and THAT way for Y. And every time you have to do things differently, the risk of making a bad part increases. And your costs go up.
Why do the customers do it to us? Some say it's because every customer thinks his way is the best way. I like to think we aren't that arrogant. I think what drives this is if the customer isn't clear on what you are doing, he will make a requirement that is FAMILIAR to him. Because that's easier on him. And "his stuff" becomes ANOTHER requirement.
Now, if you are in a little shop - this can be VERY intimidating. Here comes a dude from a major auto manufacturer, this is the big leagues. But most of them (there are a few bad apples, to be sure) are just regular folks just like you. You don't have to be frightened or intimidated. MOST of them really love what they do, and will share what they know. And MOST of them will also LISTEN.
I have found that if my shops "way" of doing this stuff meets the requirements, and I go to the customer and present "my shops way" in a clear and organized way. That at the very least:
1) Clearly shows I meet the requirements of what's needed.
2) The only real difference between "his stuff" and "my stuff" is formatting.
3) Bring up the point that every time we do something different moving parts through the shop, the risk of a mistake greatly increase.
I have found that the customer will say "Good points, approved to do it your way."
The bottom line is, if you have something in your shop THAT WORKS you absolutely CAN go to the customer and say "I'd really prefer to do it the way that works best in MY shop...." and you will find that they will listen. The blue books are a GUIDE, not a mandate. If you have your act together, you will find customer "requirements" melt away to "suggestions." And this leaves you to streamline, get efficient and cost effective.
Example: If your customer asks you about Early Containment and you give him a shaky, sketchy answer as to what it is, internally he's going to be thinking "what did my purchasing department get me into, better nip this in the bud real quick." And externally he is going to go "this is a REQUIREMENT."
But if he asks and you are johnny-on-the-spot and say. "Yes Sir! We are all over Early Containment. Come see our containment area on the floor. And here's our standard data collection system. And here's how we decide what goes on it." And it's organized and streamlined, your customer is going to say "You guys have it licked. Just do what you're doing."
Re: PPAP for Beginners - Where do I start?
I can take care of the Blog aspect. Actually these days posts in forums are more "noticeable".
I am really appreciative of the time ncwalker took to write all he did and the level detail in ncwalker's posts in this thread is exceptional.
I have marked this thread as a FAQ thread.
I can take care of the Blog aspect. Actually these days posts in forums are more "noticeable".
I am really appreciative of the time ncwalker took to write all he did and the level detail in ncwalker's posts in this thread is exceptional.
I have marked this thread as a FAQ thread.
