Your questions would require volumes of information to answer completely. This is a readers digest version to help you get started volumes Don't rely on inspection as a process indicator of your quality and defect levels. Unfortunately Inspection alone is ineffective (Inspection is only 70-80 % efficient when locating defects) as a defect indicator the information received is often to late to facilitate process improvement
The key to high yields in the shortest possible time is to identify those which have the most impact and concentrate on reducing and controlling the parameter that cause process variation. Where to start? A) Identify areas that are assumed critical. Some examples� Documentation, Screen Printing, Reflow oven, Placements, Wave solder and Hand load. B) Take each item and �. 1) Where possible use simple DOE to determine the optimum process and define areas that will require control, and areas that will not. 2) Implement procedures to reduce variation to maintain optimal process control. 3) Implement Monitoring (automated if possible) systems for process variation. The responsibilities to accomplish this would be� Engineers.. 1) Perform DOEs to determine the critical parameters. 2) Identify monitoring systems and controls. 3) Implementing the monitoring systems and controls Management 1) Support implementation of monitoring systems and controls. Operators 1) Monitor process variation and report out of spec conditions. This proposal would shift the process control from being reactive to a proactive approach. In other words, preventing defects before they occur rather than finding defects after they have occurred. Regards Michael Cox
| Good luck in your conquest. I have been struggling with the same issues for some time now. I have made the following conclusions. | First, it is nearly impossible to establish parameters in SMT to perform X-Bar and R charts. Electronics is not as clean cut as other processes such as metal stamping. There is not a dimension with dimensional tolerance attached. Usually it is a go no-go criteria. Your best bet in terms of charting is a p chart. It is going to be hard to establish 6 sigma. Standard deviation is calculated from normal methods of calculating x-bar and r charts. I think your best bet is to establish PPM defects. | I am very interested in your findings however, please e-mail or follow-up post any findings you have. | | I have recently been put in charge of our SMT line. We have to developed a SPC. My questions are: | | q1. | | What are the �most� important parameters to monitoring and which of the control charts do you usually use depending of the parameter..? | | Our line is basically composed by: dispensing of solder paste (Rotary pump), Component Placement and Insertion, Reflow soldering (air heat-transfer). Just one side. | | q2. | | About the capability, there is an idea to implement the 6-sigma criteria. I have a lack of knowledge about this criteria, and i am not sure if the best way is to apply just the 6-sigma or to mix it with 3-sigma criterium. Does anybody have suggestions in this area? | | q3. | | Does the 6-sigma criteria affect to the constants (A3, d2......) to calculate the control limits or the standard variation? or it just affects to determine the capability? | | Thanks
reply »
I have recently been put in charge of our SMT line. We hav...
- Jan 13, 1998
by
