Design for Yield (DfY) is nothing new under the sun, and it’s been heavily emphasized in various industries for some time now. However, it’s only recently that we’ve started to understand the intricate link between it and zero defects, and it turns out that there is a strong potential for synergy between the two concepts when they’re applied correctly.
However, this assumes that you have a solid fundamental understanding of design for yield, and that you also understand the implications of zero defects in your own manufacturing process. Zero defects is not a universal methodology that works for everyone, so it shouldn’t be your default choice but rather something you’ve reached after educating yourself on its implications.
Dealing with Randomness
When working with various types of production facilities, randomness can be a huge factor in the quality and volume of the final output. Sometimes it’s unavoidable, but design for yield can provide various techniques to minimize it, or at least its impact on your operations.
By maximizing the output potential of your production lines, you will bring your facility closer to a state where you’re producing quality products, with the number of failed tests brought down to a minimum. You’ll also be able to account for various types of variation in the production process, and you should be able to plan ahead much more efficiently.
This also applies to processes that tend to produce a large number of failing units, such as various kinds of production facilities in the modern IT world. There is often a huge potential for savings in this area by applying design for yield correctly.
Automating the Process
Another important aspect of design for yield is that it tends to promote automation in your production processes, which in turn makes them much less prone to errors. Of course, in most cases you can’t automate every single process within your facility, but even increasing the percentage of automated tasks with a few points can often bring significant improvements in the quality of your final output.
Zero defects, on the other hand, is something that naturally works well with automation, as this removes the element of human error to a large degree. Sure, machines can sometimes get things wrong too, especially when the specifications weren’t correct in the first place, but those errors can often be corrected more easily and don’t lead to significant problems down the road.
Keep in mind that process automation should be handled in a gradual, incremental manner rather than trying to deploy multiple systems all at once across the facility. While this may shorten the overall deployment period and make things easier later on, it will impose a severe risk on your production output for the time being, and may drag your whole operation down to a halt if you’re not careful.
You must make sure that you have good systems in place to measure the results of your efforts, and to collect historic data for future analysis. There are various ways to do that, and if you’re working in the IT field you’ll have a very good choice of options to pick from. Whatever you do choose, you have to make sure that it’s aligned with your current practices and that the data you’re collecting actually targets the points that are important for your own operation.
Design for yield can have some strong implications for zero defects, and combining the two correctly can lead to great results down the road, especially in some specific industries. However, this is not a universal approach that works for everyone, and you’ll want to carefully evaluate the potential impact on your own business before starting.