reject percentage of assembled circuits after initial test?

This is a question for those of you who actually assemble components onto a board. Any idea what the "industry standard" is for fall out on assembled PCB's? I realize that the controls put in place in your processes will effect your overall quality. But I'm trying for a general answer. On pre-tested assemblies, should you expect to find defects in 1, 5, 10% of your assemblies? What's a common percent? If I find that 10 of out 100 tested need minor rework to function properly, is that high or about right? I would like to know if anybody else has a "watermark" on assemblies functioning after initial visual inspection. Any input you could offer would be appreciated. Thanks for your input.

ps. I have a customer who insists we ship assembled circuits un-tested. Then when he gets them, he finds some defects, reworks them and tries to bill me. My contention is that we should be running the boards through a functional test, otherwise some defects are to be expected. I'm trying to determine what the industry "norm" is, if there is such a thing. Our boards are assembled to IPC Class II standards.

reply »

| This is a question for those of you who actually assemble components onto a board. | Any idea what the "industry standard" is for fall out on assembled PCB's? I realize that the controls put in place in your processes will effect your overall quality. But I'm trying for a general answer. On pre-tested assemblies, should you expect to find defects in 1, 5, 10% of your assemblies? What's a common percent? If I find that 10 of out 100 tested need minor rework to function properly, is that high or about right? | I would like to know if anybody else has a "watermark" on assemblies functioning after initial visual inspection. | Any input you could offer would be appreciated. | Thanks for your input. | | ps. I have a customer who insists we ship assembled circuits un-tested. Then when he gets them, he finds some defects, reworks them and tries to bill me. My contention is that we should be running the boards through a functional test, otherwise some defects are to be expected. I'm trying to determine what the industry "norm" is, if there is such a thing. Our boards are assembled to IPC Class II standards. | You pose a vital concern. All too often the customer (you and me) suffer the consequences of poor process management and/or poor testing. One of my most important rules is never let the customer find defect - presupposing your contract prevents the situation you're in.

First: I sure agree with your contention. The supplier must be responsible for the quality effected and detected. How else to assume liability other than taking it in the shorts with the bill back BS you discuss.

Second: Your board suppliers must assume the same responsibility. They must test (more than electrical) their product to reduce defect introduction farther down the line. Whether using net list extraction from Gerber data (much preferred) or "golden board" testing (nearly obsolete), they must provide quality product.

Third: As far as industry goes, most ICT and functional test capabilities find a reasonable defect percentage. Often the hoped for number is above 95% with a sought after goal of higher than 97%. Please note I said defect found. This does not mean acceptability percentage.

Really, these numbers are expressed without establishing cause and effect. Often, a solder joint or component failure is blamed when a defective PCB plated hole wall (as but one example) exists. No matter, I readily agree you and not your customer should be finding and correcting problems.

Earl Moon

reply »

| | This is a question for those of you who actually assemble components onto a board. | | Any idea what the "industry standard" is for fall out on assembled PCB's? I realize that the controls put in place in your processes will effect your overall quality. But I'm trying for a general answer. On pre-tested assemblies, should you expect to find defects in 1, 5, 10% of your assemblies? What's a common percent? If I find that 10 of out 100 tested need minor rework to function properly, is that high or about right? | | I would like to know if anybody else has a "watermark" on assemblies functioning after initial visual inspection. | | Any input you could offer would be appreciated. | | Thanks for your input. | | | | ps. I have a customer who insists we ship assembled circuits un-tested. Then when he gets them, he finds some defects, reworks them and tries to bill me. My contention is that we should be running the boards through a functional test, otherwise some defects are to be expected. I'm trying to determine what the industry "norm" is, if there is such a thing. Our boards are assembled to IPC Class II standards. | | | You pose a vital concern. All too often the customer (you and me) suffer the consequences of poor process management and/or poor testing. One of my most important rules is never let the customer find defect - presupposing your contract prevents the situation you're in. | | First: I sure agree with your contention. The supplier must be responsible for the quality effected and detected. How else to assume liability other than taking it in the shorts with the bill back BS you discuss. | | Second: Your board suppliers must assume the same responsibility. They must test (more than electrical) their product to reduce defect introduction farther down the line. Whether using net list extraction from Gerber data (much preferred) or "golden board" testing (nearly obsolete), they must provide quality product. | | Third: As far as industry goes, most ICT and functional test capabilities find a reasonable defect percentage. Often the hoped for number is above 95% with a sought after goal of higher than 97%. Please note I said defect found. This does not mean acceptability percentage. | | Really, these numbers are expressed without establishing cause and effect. Often, a solder joint or component failure is blamed when a defective PCB plated hole wall (as but one example) exists. No matter, I readily agree you and not your customer should be finding and correcting problems. | | Earl Moon | | In the world of contract manufacturing why should the end customer care HOW you deliver a quality product at a competitive price. You, the CM, are the "expert" in your field. Otherwise why would the customer need your services? Outsourcing rules!

reply »

| | | This is a question for those of you who actually assemble components onto a board. | | | Any idea what the "industry standard" is for fall out on assembled PCB's? I realize that the controls put in place in your processes will effect your overall quality. But I'm trying for a general answer. On pre-tested assemblies, should you expect to find defects in 1, 5, 10% of your assemblies? What's a common percent? If I find that 10 of out 100 tested need minor rework to function properly, is that high or about right? | | | I would like to know if anybody else has a "watermark" on assemblies functioning after initial visual inspection. | | | Any input you could offer would be appreciated. | | | Thanks for your input. | | | | | | ps. I have a customer who insists we ship assembled circuits un-tested. Then when he gets them, he finds some defects, reworks them and tries to bill me. My contention is that we should be running the boards through a functional test, otherwise some defects are to be expected. I'm trying to determine what the industry "norm" is, if there is such a thing. Our boards are assembled to IPC Class II standards. | | | | | You pose a vital concern. All too often the customer (you and me) suffer the consequences of poor process management and/or poor testing. One of my most important rules is never let the customer find defect - presupposing your contract prevents the situation you're in. | | | | First: I sure agree with your contention. The supplier must be responsible for the quality effected and detected. How else to assume liability other than taking it in the shorts with the bill back BS you discuss. | | | | Second: Your board suppliers must assume the same responsibility. They must test (more than electrical) their product to reduce defect introduction farther down the line. Whether using net list extraction from Gerber data (much preferred) or "golden board" testing (nearly obsolete), they must provide quality product. | | | | Third: As far as industry goes, most ICT and functional test capabilities find a reasonable defect percentage. Often the hoped for number is above 95% with a sought after goal of higher than 97%. Please note I said defect found. This does not mean acceptability percentage. | | | | Really, these numbers are expressed without establishing cause and effect. Often, a solder joint or component failure is blamed when a defective PCB plated hole wall (as but one example) exists. No matter, I readily agree you and not your customer should be finding and correcting problems. | | | | Earl Moon | | | | In the world of contract manufacturing why should the end customer care HOW you deliver a quality product at a competitive price. You, the CM, are the "expert" in your field. Otherwise why would the customer need your services? | Outsourcing rules! | Gentlemen, One of the fundamental questions have I ever met was: how the yield of your EOL (end of line testing) reflects the real quality of your product. Just consider a hypothethical situation that the product is 100% 'faulty' before testing, then with 100% effective testing you are able to repair all the products and ship them. Question: do you belive your product is a quality product(of course this depend on how you define quality as such)? Please note that testing do not improve quality, it is just eliminating faulty items from the process. This is the process that determines quality in the first place. Regards, Lech

reply »