I agree with Wolfgang that a check list, like he proposes, is very useful in evaluating your ability to assemble products. No sense designing a 16 inch board if the throat of your wave is 15 inches.
Taking a different tact � what if you define your capability to do something (ie, place a component, maintain a temperature, bowl strikes, etc.) in terms of how well you do that thing time and time again? The SMTnet "Terms And Definitions" states:
Process Capability. Competence. A measure of the process variation about a defined target value. Cp and Cpk are common process characterization indexes.
If you place an QFP208 on the spot (maintain the heat on a QFP208 for a certain time over liquidous) [throw a ball down a lane] and then compare the actual position (time) [pins knocked down] with the intended position (period) [300], you�d get a measure of your capability to hit the target. If you repeated that many times, you get a good idea of just how good you are. That�s your capability.
From that, measuring process capability doesn�t sound like something you want to do very often or for very many things, but it sure would give you an good idea of how well your process will perform. And it would be a baseline for determining if the process was still in control. And it might be one tool to use to determine if the fancyschmancy equipment you purchase will meet your expectations.
That�s what I was babbling about last week when DEK was googooing their top-end printer with a Cpk of 2. Until recently, no equipment supplier wanted to talk about Cpk. Then Siemens and UIC began chirping about their 3s, but the printer and oven houses kept mum and tried to pass it off as a "placement machine measure." Yano why? Their Cpk are 1.67. From a big picture standpoint, that�s like the word for drawing water through a straw. Relative to similar equipment it�s pretty good.
If this is what you�re talking about, there is a great body knowledge on this that builds on basic SPC concepts. Look to AAIG, ASQC, and authors like Harry and Breyfogle for help in learning more. This is a subset of six sigma stuff. Consider that the SMTnet archives may be a excellent source, also.
reply »