Issue link: https://iconnect007.uberflip.com/i/373246
28 SMT Magazine • September 2014 The lesson learned is a simple one. Optimization of processes extends beyond the actual machine itself, often all the way, as this SMT example shows, to shop-floor planning and scheduling. " " greater, as manufacturing trends toward lower volumes, higher mix, and the need to support sudden demand changes. This first lesson is a fundamental issue of au- tomation, but it was the second lesson that ar- guably has ultimately caused the failure of most attempts at automation. Going back to SMT, one of the key criteria for machine performance is placement accuracy and quality, so much so, that inspection is required in most production lines to confirm consistency of placement. This started out as simple human sight-based visual inspection, based on feedback from in-circuit testers (ICT) and subsequent repair processes. Issues that were detected on a regular basis were highlighted and specifically visually checked on every PCB, with no time to check ev- ery placement. Causes of defects ranged from changes in material's shapes or sizes between ven- dors, to variations in the PCB positioning, contamination of materials or PCB, nozzle wear, and feeder positioning and wear, among many oth- ers. These various causes fell into three main categories: ones that are setup-depen- dent, such as different mate- rials shapes; some "random" one-off defects; and some that are trending defects, which ap- pear and worsen over time. It is this latter kind of defect that is the most difficult to recognize with inspection be- cause the criteria of what is acceptable and what is not is hard to apply. The development of AOI machines brought a solution, allowing specific control criteria to be set and measured against by a computer. At least, that was the theory. In practice, most de- fects detected by the AOI machines were later dismissed as acceptable when inspected by a repair operator, which led to sophisticated net- works of AOI machine results being funnelled in real-time to a repair operator who could re- motely over-ride AOI detected defects with a simple push of a button. With a huge effort by AOI machine vendors, the visual defect detec- the trUe ImpACt OF AUtOmAtION continues feATure performing movements beyond their minimum cycle time. Optimisation software for the reduction of excess machine travel works on material posi- tioning and placement sequence order so that the maximum throughput of the machine for a specific product can be achieved. Although this may be relatively simple when considering one machine program for one product, the reality is that with the increase in the mix of products produced by SMT, the saving of time through program optimisation is almost insignificant compared to the time taken to change the ma - chine between products, even if the product is changed only once per day. Most of this change- over time is for the teardown of old materials for the previous product, followed by the setup and verification of the new material for the next product. When considering overall SMT productivity, the shift to creating material setups where locations were com- mon across sequentially run- ning products reduced the changeover time dramatical- ly. However, the compromise was a reduction in machine performance for each product because materials could no longer be located at optimum positions. The lesson learned is a simple one. Optimization of processes extends beyond the actual machine itself, often all the way, as this SMT example shows, to shop-floor planning and scheduling. This first lesson had not been learned in the 1990s, when technology advances brought the first real opportunity of automated assembly production lines into electronics. In most cas- es, lines of so-called robots were set up to make simple products, such as remote key fobs for cars. When the product model changed, how- ever, the line had to be retooled and reconfig- ured, which in some cases led to the teardown and reassembly of the whole line itself. Automa- tion at the time, although technically possible, faced significant challenges in flexibility. Today, the need for flexibility is an order of magnitude