COMPANY CONFIDENTIAL. DO NOT DISCLOSE.
For Samsung, working smart means exploring, analyzing, automating, innovating
Samsung Austin Semiconductor offers rich training and mentorship opportunities designed to expand data-driven creativity
_____________________________________________________________________________________
Challenge: Meet the demands of an evolving technological marketplace. Maintain quality and reliability standards even as wafer dimensions shrink and new applications – like driverless cars and artificial intelligence – introduce new requirements.
Solution: Samsung’s Work Smart group partners with other departments to help them tackle a variety of data challenges. A range of training opportunities encourages engineers and management alike to use JMP® to deploy the Work Smart philosophy on their own projects.
Results: JMP equips engineers at Samsung to accelerate processes, optimize quality and drive new innovations that keep the company at the forefront of a highly competitive marketplace.
_____________________________________________________________________________________
It’s no exaggeration to say that the entire digital economy is today built on the microchip. Consumer electronics, mobile devices, network equipment, IoT sensors and automotive components define the landscape of global commerce and communications; all depend on ever-evolving high-precision wafer technology.
Samsung Austin Semiconductor is among the world’s most advanced facilities for semiconductor fabrication. Part of the Samsung Group, South Korea’s largest business conglomerate, the operation represents one of the most significant single foreign investments in the United States. In step with its motto, “Where technology meets creativity,” Samsung continually pushes the envelope in wafer technology – most recently, though a 3D chip structure that achieves higher performance without increasing chip size.
But maintaining a leading edge is far from easy. Samsung trains a laser focus on minimizing costs, accelerating manufacturing processes and driving R&D innovation. “We try to always be in the lead,” says Dan Sutton, an engineering statistician in Samsung’s Innovation Group. “That takes more than just squeezing pennies out of old processes. We’re trying to make products that are faster, more reliable, more innovative.”
Innovate smart, not hard
The most important innovation, of course, is the creativity that goes into the company’s products, though the design target shifts with a product’s intended use. For a smartphone with a short product lifecycle, speed might trump longevity. For an automotive component that must last a decade or longer, quality trumps other characteristics.
To meet these innovation cycle challenges, Samsung has implemented an initiative called Work Smart. The concept shares a similar philosophy to the Lean Six Sigma movement in manufacturing; data-driven decision making is a core principle. When it comes to data streams, Sutton explains, “You can look at data in a spreadsheet, show some signals, look at some dashboards. But to work smarter, you need to analyze that data and reach a conclusion.
“To get the innovation and quality we need, we analyze data, pick out the signals and make the best chips we can…. And instead of spending hours getting the data together in an Excel pivot table, we provide tools like JMP to speed up that work analysis. That’s what we call Work Smart.”
One tool, many applications
JMP statistical discovery software from SAS, Sutton says, helps engineers across the organization deploy the Work Smart philosophy. With its wide-ranging applications, JMP has become a go-to tool for helping Samsung to achieve its goals in several key areas:
- Speeding engineering cycles. Samsung is all about innovation, but creativity has to occur at the speed of the marketplace. “If a customer requires an improvement, we need to achieve that quickly,” Sutton says. “Design some experiments. Find the signals. Deliver to the customer as soon as possible.”
- Deploying more rigorous quality and reliability measures. In wafer fabrication, quality is paramount. “A good example is automotive,” Sutton says. “If the future is driverless cars, all those systems will have to be very reliable for obvious reasons.”
- Turning process and product data into actionable insight. From product design to materials testing to fabrication monitoring, Samsung captures, manages and analyzes huge volumes of data. “There’s so much data that everyone complains there’s not enough time to look at it all,” Sutton concedes. “We use tools like JMP to speed up analysis and understanding.”
From breaking ground to groundbreaking
“Both engineers and managers like to be able to drill down on data,” Sutton says. “For instance, if you’re doing a measurement system analysis, you can look at the residuals versus the time sequence. And you might see a little signal, like the first couple of wafers are a little higher than the rest. That might tell us we have to warm up the equipment longer.”
But the ability to explore data effectively and recognize trends when they appear begins with a basic statistical proficiency – one that not all engineers possess when they first join Samsung. As someone who devotes a large part of his time to training and mentorship, Sutton works with engineers across the organization to develop and hone these skills. The goal? Invest time and resources in building an engineering workforce that can use data to inspire create ideas.
“When I teach basic statistics,” Sutton says, “I get to compare what Excel can’t do to what JMP can do.” And it’s often the JMP Graph Builder that really makes a difference for newcomers to statistics, he says. The ability to explore data interactively through dynamic visualizations breeds the kind of creativity Samsung relies on. “People start using Graph Builder more and more…. They try things out in Graph Builder first to see how they want to look at the data.”
At the slightly more advanced side of the scale, Sutton encourages all engineers at Samsung to embrace design of experiments (DOE) – the ubiquitous approach to exploring multifactor opportunity spaces – and to learn the specialized DOE capabilities JMP offers. From split-plot designs to definitive screening designs to custom DOE, JMP helps engineers implement the Work Smart philosophy by constraining experiments to the factors and interactions that matter, saving time and ultimately, experimentation costs. “Custom DOE is very important to us,” Sutton emphasizes. “When I’m teaching, I show all the time how custom DOE reduces the number of wafers we have to test.”
Automation extends the benefits
With offerings ranging from introductory statistics through Six Sigma certification, Samsung’s robust training program is testament to the organization’s desire to cultivate a data literate workforce. One of the most essential – and perhaps also most creative – skills Sutton supports is automation via custom scripting.
“The point of data analysis is to get to a decision point,” Sutton notes. “If you spend four days getting the data into a graph, then you don’t have time to think about what it means. But if you can automate the display of data, then you can spend more time on new ideas.”
Scripts enable the automation of repetitive data access and reporting, allowing teams to scale up or across faster and reduce the amount of time they spend on repeat tasks. “If you write a script in JMP to automate a process, you won’t have to drag and drop data every time,” Sutton advises. In a more advanced case, Sutton wrote a series of JMP scripts to assist with a metrology site reduction analysis. This resource now enables engineering teams to better understand process targeting and sources of variation and allows them to easily pick the best reduced site map.
While scripting may sound daunting to some, Sutton’s training courses help practitioners make sense of the tools they already have available. The script editor in JMP, he says, provides a user-friendly environment for writing and reading JSL scripts: “I like that you can just highlight a little bit, show this or that, and troubleshoot your own little program…. We also love the ability to save scripts at any point of analysis. If you’re working on a file but you won’t be done today, you can just save your analysis at that point and pick it up again tomorrow.”
New ideas, new success
For Samsung Austin Semiconductor, new ideas are what drive new success. The company recently announced America’s first “innovation zone” for manufacturing-focused 5G communications. And they are now exploring how fifth-generation cellular technology – with its high data rate and low latency and energy usage – can enable massive connectivity in manufacturing environments. This example is no outlier; everything Samsung does involves innovation. “We have cost innovation,” Sutton says. “We have manufacturing equipment innovation. We have process innovation. And now we have data analysis innovation.”
