how to achieve 500ppm?

In that the machine times out and waits for a feeder refill and doesn't just skip that part and leave the space empty, it's not a defect.

I have found 500ppm *fairly* easily attained in SMT depending on the board mix and how you define (1) a defect, and (2), a defect opportunity. If you make every pad, component, and board one opportunity it's quite do-able.

My focal point is continuous improvement, so how I arrive at my ppm numbers is less important than reducing them consistantly, at least until the low fruit has been plucked.

reply »

From my understanding: Sigma is the same as one standard deviation. So 3 sigma is the same as 3 standard deviations. It all boils down to how repeatable your machine is. Another way to look at it is the probability (or confidence) that the machine will actually place within its specs. 1 sigma is equivalent to 68.2%. This means if you said your machines are rated to +/- 0.1mm at 1 sigma, you only expect 68.2% to actually be +/- 0.1mm. The other 31.8% could be worse. 3 sigma is 99.73% and 4 sigma is 99.99%. I know of some machines (Siemens and Universal) who rate at 4 sigma, but I think most others either don't say or use 3 sigma. So the lower the sigma, the lower the probability (or confidence) that the machine will work within specs. I'm not sure where you got the numbers per million above...

reply »