acceptable defect rates for smt process, in ppm.

Subscribe to Thread

O.K., folks, I know this is a loaded question, but I've bee... - Jan 14, 2000 by Steve Thomas

Steve, Basically in answer to your question, there is no r... - Jan 15, 2000 by Mark Wiegold

Thanks, Mark. Actually we already have an established mark... - Jan 15, 2000 by Steve Thomas

My old company (CM) ran SMT to 50ppm including some very ... - Jan 17, 2000 by

It makes a difference what the normalizer is. We calculate... - Jan 18, 2000 by

Steve Thomas

#7842

acceptable defect rates for smt process, in ppm. | 14 January, 2000

O.K., folks, I know this is a loaded question, but I've been asked to find the answer, sooooo:

What is an acceptable defect rate in ppm for a surface mount process, assuming that each component has the potential for one defect. This would include defects caught at all stages, iow during the process, during touch up, and at test.

Thanks.

reply »

Mark Wiegold

#7843

Re: acceptable defect rates for smt process, in ppm. | 15 January, 2000

Steve,

Basically in answer to your question, there is no real set number for defects. Defect rates will vary between products and companies. If my company was running the same product as yourself then there is no reason to suggest that the defect rates should be the same.

The main idea behind defect rates is to highlight problems with your process with one simple value. I suggest you use SPC charts and Test print-outs and find what you current defect rate is. Then set yourself a defect target and use a quality tool i.e. Parito analysis and improve your process where possible.

Hope this gives you a better idea.

Mark.

reply »

Steve Thomas

#7844

Re: acceptable defect rates for smt process, in ppm. | 15 January, 2000

Thanks, Mark. Actually we already have an established mark. We use 500ppm (99.95%) as our acceptable level. Problem is, someone (another manufacturer) told someone else (our pres.) that THEY build to 50ppm. Soooooo, someone else told my boss that WE should be building at 50ppm. We are not even CLOSE to having the process controls, the process, the staff, or the money to achieve that level of quality. I don't think the "someone" that told "someone else" they were running at 50 ppm are building anything with much complexity, but we dont' really know. Hutchins says that 1000 ppm (99.7%, or +/-3 sigma) will give good results. I figure we're performing at a reasonable good level, although we have a ways to go before I'm satisfied with it. My goal is to cut it in half, but not by a factor of 10!

reply »

Brian W.

#7845

Re: acceptable defect rates for smt process, in ppm. | 17 January, 2000

My old company (CM) ran SMT to 50ppm including some very complex boards. We established the normalizer number by: #components + #solder joints. As was stated earlier, the ppm for any given product is the result of many factors. You may get different results from different lines. Design also plays a major factor. We had boards that ran higher than the overall process average due to poor design from the customer. You should track each assembly. Get the process into control, then calculate control limits for the ppm. 99.7% is 3000 ppm. +/-3 sigma means your cpk is 1, and it really needs to be better than that. 3000 (or 1000) ppm meqans a lot of rework. Reducing the defect level can be done, but it takes lots of effort and commitment. You must define ALL the parameters that affect the process, including design, and figure out how to control each one. 50ppm is attainable, but EVERYONE must be commited to reaching it, and understand that it might mean some cost to get there.

If you have any further questions, please feel free to e-mail me.

Brian W.

reply »

WDavidson

#7846

Re: acceptable defect rates for smt process, in ppm. | 18 January, 2000

It makes a difference what the normalizer is. We calculate solder ppm and placement ppm separately for each assembly. Solder ppm = #solder defects*1E6/(Qty boards*#solder joints per board). For us this number is easily less than 50 and sometimes <10. Now do the same calculation for placement defects using the number of components per board in the denominator. For us this number varies much more than the solder ppm. Our best numbers are around 200-300ppm on a board with 900 components and 3000 solder joints. I have seen bad runs where ppms jump to 2000. All of our data comes from AOI. I'd be interested in what your thoughts are on SPC charts and process capability.

reply »