Inspection Plans - Subcontracted Electronics Assembly

[Andy Bassett]

Can anyone help me out with a small problem. I am working as a consultant for a small (6 People) computer manufacturer near Dusseldorf. The company has had to learn from zero all about in-coming inspection, final inspection, inspection at the supplier etc.
As assembly is actually sub-contracted they need a very good Assembly/Inspection Plan, what they have done is a fair effort, but to help them does anybody have a sample/example of a Assembly/Inspection Plan for a peice of electonic equipment, ie computer, printer etc.
Regards

------------------
Andy B

[Kevin Mader]

Andy,

Have your clients performed FMEAs or done other Risk Analysis? These processes point to where the need for inspection at Incoming Inspection are. For instance, if during Risk Analysis, your clients had determined that each circuit board must be individually wrapped, part of your incoming inspection plan may be to visually observe for proper packaging.

Regards,

Kevin

[Andy Bassett]

Hello Kevin

No they have not done any FMEA Analysis. On another thread i managed to ascertain that FMEA Analyis cannot really be outsourced, and i have found a company to train them with a 1 day FMEA course in August.

However this will come a little late for their current product. If i can show them a Inspection Plan for a similar product it would help to shape their thoughts. Typical questions that they are asking are.
1. Do they have to define the assembly process themselves(Consider assembly is sub-contracted).
2. Just how detailed do the inpsection steps have to be (ie isnt 'check screws after tightening' a little over the top).
3. They have designed some quite detailed inspection plans, but the employees seem to be prone to 'Tick Syndrome' without really checking.
I should just mention that the Assembler is ISO 9001 certified and currently applying for QS 9000.

By the way, i have never heard of Risk Analysis in this how it sounds? Is it a special Technology, or just another name for FMEA?

Regards
Andy

[Kevin Mader]

Andy,

You guessed right, FMEA is a form of Risk Analysis. Generally speaking, FMEAs live in the automotive world. This is a great tool, however, for any industry. When you get into Medical Devices and Electronic equipment, often times other directives apply where Risk Analysis is expanded on (FDA, CE, etc.). Here, additional requirements must be completed as part of the process.

Taking your questions one by one:

1 What controls are excersized over your Clients suppliers? Does the supplier have a control plan they can send you? Is it satisfactory? Are proper percautions taken to ensure only conforming product is shipped to you? Do you need to perform any Incoming Inspection?

2 What are the Critical Product Characteristics (ie. what characteristics are critical to the operation or functioning of the unit)? Checking screws, unless determined a Critical Characteristic, is probably overkill.

3 Inspection is not a value added operation. It may be a nonvalue added but NECESSARY function however. Which do you have? Inspectors not inspecting might suggest that inspection is overkill, especially if proper precaution were taken when the item was created. It may also point to a deeper problem, so beware.

My suggestion would be to determine what precautions are taken by the supplier and what controls over the supplier are in place. From here, identify the critical characteristics, create an inspection plan using a statistically sound sampling plan. Based on past history, you establish the sampling level that best covers the organization and the customer (alpha/beta risk). Review the plan after a period of record making to see if you were too aggresive or too liberal in your planning.

Got to run......

Kevin

[Andy Bassett]

Thanks for your comments. As we are about to lauch a second new product i will take your comments on baord TODAY, as next week we will start interviewing the main suppliers for it. When i read the posts in this forum i get the impression that the Auotmotive boys all speak the same language.

When we start to discuss with the suppliers next week i know that they will hide behind the 'We are just a retailer' or 'Nobody has ever asked for that before' or 'Inspection...you will have to pay more for that'.etc etc.

As we are only on our second product, which is completely different to the first we dont have much experience to start setting sampling levels, we will have to shoot in the dark a little.

Reading your comments i had the idea to do the planned FMEA course using the new product with the target of producing a Process FMEA for the actual assembly of the final product which should solve one problem.

Could you tell me anything about the Alpha/Beta that you mentioned, is it something i can use, if its heavily into stats i will reitre hurt.

Regards

Andy Bassett

[Kevin Mader]

Andy,

I think it would be a good idea to use the new project in your FMEA training. While selecting your source for training, speak with the trainer about your idea to use this new product. I would have to believe that the trainer will agree that this is a good idea.

When preparing an Inspection plan, you should use a statistically proven sampling plan (perhaps ANSI Z1.4). If you arbitrarily select a sampling plan (i.e. 5 because I have 5 fingers), it becomes difficult to prove to a registrar that your plan is proven effective. This does not mean that you can't do that. It means you need to do additional work to prove your method works, which most folks don't want to bother with especially in light of having plans already in existence and proven.

Alpha/Beta risk refers to the Producers/Consumers risk to rejecting a "good" lot or accepting a "bad" lot of product. You need to establish as an organization the level of risk you take (what is tolerable?). Generally, 5% risk is used as the industry norm. This isn't heavy stats, just basic stuff really. You can read more about it in most Sampling books (reference Jurans Quality Handbook, 25.9).

Inspection is a funny topic. Is it necessary? When is it necessary? Different for any given situation. Picking a starting point for Sampling Plan and Sample Size is difficult for most folks. Where do you start? The only way to answer this is to review historical data (how many lots rejected vs. lot received) proportion defective and know the risks encountered. While not very analytical, it can give you the basis for drawing the line in the sand. After collecting history for a period (again difficult to say, depends on your volume) you will need to review the Plan to determine its effectiveness. Not effective, increase your sampling size, frequency, tighten your sampling plan, many, many, considerations. But eventually you get to a plan that will work well in your system. I am giving you the brief on this, but you should read more about sampling plans to draw stronger conclusions for yourself. I hope this helps a bit. Don, are you out there? You tend to explain the OC pretty well.

Regards,

Kevin

Back to the group....

[Don Winton]

Don, are you out there? You tend to explain the OC pretty well.

Yea, Kevin I am here. Have been real busy the last few weeks and will be so for the next many (I hope). I have started consulting part-time and have two clients thus far (in addition to my full time job). So, things are a bit hectic. Anyway, on with the show:

You have addressed this subject well, to which I will only add-Rather than discuss in detail here the various aspects of sampling, I suggest the following:

For a discussion of sampling plans in general and OC curves in particular http://www.samplingplans.com

There is also an OC curve simulator at my site in the Statistical section. Hope this helps.

Regards,
Don

[Andy Bassett]

Thanks for the Statistics Link, i have gone through it (along with a bottle of wine) and will defintely use it for reference for the future.

For the moment it is a bit further advanced than either my knowledge or my clients experience.

Maybe i am unlucky, but i always seem to be involved with clients who have something less than regular mass production, which seems to make statistics difficult to use (Or maybe i dont know how to use statistics properly?),

For example the current client wants every product to come off the production line to be correct, which means a 100% Final Inspection is planned.

On the supplier side as many goods are coming through retailers, they give you the cold shoulder if you start to question them about their quality levels.

Historical evidence is nearly impossible to come by as this client is only on its second product which is completely different form the first.

I can only assume that we need a little more experience with the products before creating sampling plans with any accuracy.



------------------
Andy B