Issue link: https://iconnect007.uberflip.com/i/1466505
MAY 2022 I SMT007 MAGAZINE 23 The Labor Economics for Programming Placement Johnson: From the time you start program- ming offline for a typical board until you've completed debugging on the machine, what's the timeframe? Bennett: Typically, it takes about eight hours to program a medium complexity board offline and the debug time on the machine is 30 to 60 minutes. Johnson: at's quite a bit of time when you're in a high-mix environment. Bennett: at's correct. With small job order quantities, we're able to perform three change- overs within an eight-hour shi on any given SMT line. Johnson: You said earlier that good program- ming up front improves the backend. Talk about that inter-relationship, especially from the perspective of working with the high-den- sity boards. Bennett: Programming a high-density board is labor intensive for placement and AOI. e data provided to the AOI programmer comes directly from the placement program. Data from the same source will reduce inconsis- tencies between the components placed on a board and AOI verification. If the placement programmer and AOI programmer each start from scratch with the data provided by the customer, there is more risk of mixed results between the placement program and the AOI program. High Density and Profit Margins Matties: Do you think that the more profitable work is in the higher density kind of work, even though you have to put more effort into the front end? Rowland: Generally, the more complex the board is, the more we need to charge because more effort goes into programming and man- ufacturing it. Generally, the profit margin is higher on jobs that are more complex because they are more difficult to build. Matties: How do you gauge this without giv- ing away your secrets? When you're looking at your factory, are you at a dollar throughput per station or does that act as a guide for the type of work that you're taking? Rowland: We've developed a pretty compre- hensive module that we use to quote the labor it takes to build and test a product. It contains all the processes that we have here. If it's a good fit for our services and processes we will quote it, otherwise we would no-bid it. Over the years, we've learned it doesn't make good business sense to try to do every- thing for everybody. We were trying to do that years ago and it ended up being a no- win model, especially if there's a sizable capi- tal investment that's required. In the last cou- ple of years, we've taken a different approach, looking at the services and processes we want to offer and excel at. If it doesn't fit our model, Short tapes of components can be problematic to feed into today's automatic feeder mechanisms.