SMT007 Magazine

SMT-Sept2016

Issue link: http://iconnect007.uberflip.com/i/721932

Contents of this Issue

Navigation

Page 49 of 121

50 SMT Magazine • September 2016 line. Asset utilization looks broadly at the time for which the machine was "operational," com- pared to when it was "available," and is often used to compare the performance of manufac- turing sites. However, the way of measurement can po- tentially be different depending on the specific focus or purpose of the report. Let's take a look at the following scenarios to illustrate how this might work. • The machine is present in the factory, ful- ly installed, and ready for operation. It is avail- able 24 hours per day, seven days a week. The asset was paid for, and, in theory, it should be adding value at all times, the more the better; so no excuses, measure the operational time pure- ly against the physical availability. • Perhaps though, the manufacturing site has vacation periods. Different sites have differ- ent vacation times, so why penalize a factory and its assets because of something that is be- yond the control of factory management? If the site is on vacation, then we could argue that the machines are not available during that period. • Some times in the factory not all lines are utilized. Scheduled downtime could also be re- moved from the calculation because it is not a reflection on the machine performance that the factory is not fully loaded. In many cases, knowing the performance of the machine while actually active in production is more valuable. • Given these scenarios, what about the changeover time? It is not the fault of the ma- chine that manual verification of materials has to take place. The machine can only be opera- tional during actual production, so perhaps ex- cluding the changeover time from the availabil- ity calculation adds more value in some cases. • In our example, machines "Y" and "Z" were in production, but stopped because of an- other machine, "X," earlier in the line. From the equipment performance perspective, this is not a fault of either "Y" or "Z", and so it can be argued that this external stop time should also be removed from the availability calculation. • How about machine "X" though? The reason for the stop of "X" was not related to the machine itself. Moisture contamination of a material is a reason that is completely exter- nal to the machine. If we want to compare ma- chine performance, model-to-model or site-to- site, then perhaps all external reasons for stop- page should be eliminated from the machine availability time. • There are also times that the machine is operating, but, in a way that is defective. Per- haps this time should also be removed because the machine operation was not adding value. The measurement of performance should be al- tered according to the quality data, so that it is only included when the machine is working perfectly, which would be then the definition of it being operational. • Whenever materials need to be replen- ished, the SMT machine may need to stop. Should this also be regarded as a machine re- sponsibility, or external? • How about the cases for more complex machines, where there are two or even three conveyors for different PCBs through the ma- chine simultaneously, with multiple modules of machine operation each with multiple heads? Rarely are all parts of the machine working at the same time, and it is equally rare during op- eration that they are all stopped. Different areas and elements of the machine will start and stop as required, but how can this be measured and accounted as run time or stop time, availability, and operation? • Finally, what about the loss time within the machine while executing a program that is not 100% efficient. Excess movement of heads should be avoided as far as possible during the programming sequence creation. But how to ac- count for the different losses when you're com- paring a program that is optimized for a single product versus that which is optimized for a range of products so that additional distance for travel to pick up parts is required which slows the cycle time of the machines? How about the case where feeders are not placed on the ma- chines in optimized positions (random feed- er setup), and the machine has to compensate accordingly with significant increase of cycle time? These are all simplified arguments that have been a part of industrial engineering for many years. Because of the complexity of the issues, a SMART FOR SMART'S SAKE, PART 1

Articles in this issue

Archives of this issue

view archives of SMT007 Magazine - SMT-Sept2016