Anyone else? This is an interresting question. With 1000PPM won't your AOI be worse then the process it is supposed to measure? Won't your AOI become more of a bling thing to impress customers than an actual usefull tool to improve process quality?

reply »

Stephen

#55773

AOI false fail PPM | 31 July, 2008

Aren't you worried about false passes? With that low of false fails, you must be only inspecting easy stuff like presence and not checking for correct parts.

reply »

Jdd

#55781

AOI false fail PPM | 31 July, 2008

We are in the process of determine our target.

We are inspecting with all the options Presence, Joints, OCV, Polarity, etc.

What would be an average of ppm on AOI? 1000, 5000, 10000? Low volume - High mix.

our machines are Agilent SJ50 Series 3

Thanks

reply »

Stephen

#55791

AOI false fail PPM | 1 August, 2008

I don't know about that machine but 5000 wouldn't suprise me.

And make sure your targets is less than X false fails with zero false passes.

I can probably get zero false fails with any machine but I"m sure a lot of defects would get past that machine.

Just make sure you keep your eye on the real goals.

reply »

GRViper

#55892

AOI false fail PPM | 7 August, 2008

First all, you must define the specific defects that you wanted to screen out by your AOI system based on its capability, so you can allocate the undetected defects on your other means of inspection methodology. Then optimize your AOI inspection parameters balancing the escape and the false failure rate. Keep in mind that as the false failure rate becomes lower, the defect escape rate is getting higher.

Currently our AOI goals were at: <100 PPM escape rate and <2000 PPM false failure rate. Low and high volume goals should be the same.

reply »

R.Powers

#55906

AOI false fail PPM | 8 August, 2008

With AOI it really depends on what you are testing. Not only what criteria, but what different types of components. We run a high volume low mix application for automotive checking for presense, positional accuracy, and all solder joints. We currently see about 20ppm false fail rates and 1-2 ppm escape rate. Although our application is extremely high volume, we have very few differnt types of components. We are also not looking for polarity, bridging, or OCR. With today's machines I think that 1000ppm false fail and 100ppm escape rates are reasonable for normal applications. Again, adjust that number up as the component mix increases.

reply »