pc board defects

I can't give you itemized values for each defect category because we don't ask operators to distinguish between them, and we haven't done any serious research to do so either.

I do wonder this, though. I read data somewhere from 14 years ago stating something similar. Fourteen years later the improvements in printer and stencil design hasn't reduced that to less than half? I don't believe it.

We reduced our standard pitch defect numbers (that we attributed directly to print problems) by about 25%, our fine pitch stuff about 75% when we got a new printer. Some of the reduction was based on repeatability of the printer, some on vision finding the defects, and some on reducing the requirement for operator intervention. The point is, I'll bet that 60% figure is invalid these days EXCEPT for those that still are struggling with the unenviable task of setting up a worn out clamshell to print 2400 pins worth of 20 mil pitch on a 10"x20" dsp board.

reply »

We need to remember here that as printer manufacturers were enhancing their product lines, miniaturization was quickly being developed in the background. In the last 2 years I was involved in developing and conducting a D.O.E./F.M.E.A in determining Print CPK. An internal Team was identified will the first task of identifying high level root cause, then a fault tree diagram was developed to Indentify print characteristics and variables associated with the process. A very detailed set of print tests were developed and conducted. The long and the short of it is we determined that the printers themselves are extrememly accurate and repeatable, mechanically, print to print etc., it is ALL other variables that affect printing performance. Our tests indicated many short falls associated with: Operator to Operator repeatablity Paste relax and recovery (Stencil life) Under wipe solvents Handling of boards Cyber Centry coverage and equipments inability to inspect 100% of sites. (escape factors) Stencil Design Stencil cleaning between runs The highest attributor was clogging apertures in fine pitch, resistor networks, and BGA, Micro BGA areas

Obviously there was a lot more "Meat and Potatoes" to this study.. But, yes...60-70% of SMT defects are print related. As the component manufacturers continue to develop smaller and smaller footprints the challenge will remain in keeping up with technologies to support accurate and repeatable paste deposition, especially on High Mix boards with an average of 45,000+ opportunites. I remember a well known trainer and industry expert (Phil Z.) about 5 years ago stating that 95% of all SMT related defects are printer related. So we have come a long way. Scott

reply »

That 60+% print defects number has been bantied around for so long in the electronic assembly business for so long that people believe that it's true. One starting point for this myth and lore is: "Where Quality Is Lost On SMT Boards" C Mangin, 'Circuits Assembly', 2/91 breakes-down process defects as follows: * Solder paste screening: 63.8% * Component placement: 15.3% * Reflow solder and cleaning: 15.2% * Incoming components: 5.7%

The industry is doing things differently now.

We has a similar discussion on SMTnet. Look here: http://www.smtnet.com//forums/index.cfm?fuseaction=view_thread&CFApp=1&Thread_ID=4934&#Message19096&

reply »