A Fortune 500 company uses Instrumental to solve a ramp-blocking issue. Where other methods failed to provide definitive root cause, Instrumental came through almost immediately.
With two months until the American holiday shopping season, Company B should have been ramping up production volume in order to meet demand. Instead, the discovery of a troubling product reliability issue during the final stages of development in early summer had thrown that timeline into chaos. Since they had already missed their initial shipment date, getting the product ramped up and shipping in time for Christmas became mission-critical.
Company B builds a new type of consumer electronic product that due to respecting confidentiality, we’ll refer to as Company B Product, a handheld consumer electronic device. The team working on Company B Product is innovative, highly-resourced, and well-experienced. Together, they worked hard to remedy the issue themselves for several months before engaging with Instrumental.
Eventually, they were able to leverage Instrumental’s manufacturing data system to effectively isolate a challenging design problem, identify its root cause, minimize the amount of experimentation required to arrive at a validated corrective action, and implement a change that would ensure highly reliable performance for their customers. All in time for Christmas.
Instrumental received an email from Company B on a Friday afternoon in September and engaged the product team first thing Monday morning.
Company B’s handheld device comes in two SKUs (stock-keeping units), each of which is built on two equivalent lines with a side-by-side setup. While the devices are nearly identical, they performed differently during a 72-hour heat soak test. On a subset of the test units, the SKU B top covers, which were adhered to their enclosures with double-sided, pressure-sensitive adhesive (PSA), popped off their housings, while the SKU A top plates were all fine. They didn’t know why the SKUs performed differently, but needed to find out quickly so they could ramp the product for launch.
Company B had already investigated several potential root causes. They had extensively measured all of the components to ensure they were made to the specification, had made multiple iterations of improvements to the pressure fixture that cures the PSA, had removed components to see if they could isolate any that were applying an upward force, and even attempted liquid glue.
We have got to ship this thing and don’t necessarily have the data we need to solve this problem — Company B product design engineer
“We have got to ship this thing and don't necessarily have the data we need to solve this problem,” said a Company B product design engineer. The team needed a better way to correlate data from before and after reliability testing. With their current method, it took them three to four days to build the DOE units, and then another three to four days to test and receive results from their factory counterparts. These iterative cycles were consuming resources on the ground and slow. With the planned launch date receding into the past, the team needed tools that could enable them to discover the root cause quickly, as guessing was costing them time.
When Company B came to Instrumental, they were seeking a solution to the frustration of running numerous, time-consuming DOEs that had not yet yielded success. At the point when Company B engaged Instrumental in September, the product team did not yet have a solution, but did have one key theory they had been unable to test: that there was an interference in the design that was not obvious from standard measurements. The team hoped to obtain more data about their processes and parts so they could eventually compare SKU A and SKU B before and after reliability testing.
Instrumental worked to devise a strategy that would quickly collect high-resolution images of the enclosure (pre-snap), the top cover, and the final assembly. Working in tandem with Company B’s engineers, Instrumental built and deployed four stations to Company B’s assembly lines in China in record time: a pre-snap and post-snap station for each SKU A and SKU B.
After a two-day period of calibration and setup at the final assembly site in China, data started flowing. Immediately, the team — both onsite in China and remotely at Company B headquarters in California — was able to access the data.
Using Instrumental software, and armed with pre-and post-snap data, Company B engineers were able to easily search and review pre-snap photographs of the units that later failed reliability testing using serial numbers. “It’s an X-ray view into the product, and is especially useful for the development phase,” said the product design lead.
I noticed right off... there’s tons of variation and we need to be paying attention. — Company B product design engineer
The Company B product team learned that Instrumental’s ability to provide insight via visual comparison was much more valuable in the assembly process than they initially thought. “I noticed right off, when flipping through 20 assemblies rapidly, what kind of variation there was. I had anecdotal evidence that parts were biased, but wow, there’s tons of variation and we need to be paying attention,” said a product design engineer. The product team used these measurements to discover that the top cover interfered with the enclosure 75% of the time. “We found value in the measurement data” explained the product design lead.
Timing is everything. Engineers were able to do all of these actions remotely, removing the distance and time barrier between their offices and the factory line. Best of all, the data is self-serve — all of these observations were made by the engineering team members, when they needed it, without physically touching a single unit.
Company B was able to use Instrumental to quickly obtain access to high-resolution data that moved the needle for their failure analysis process. Instrumental systems collected nearly 8,000 units worth of image data over the course of several months.
Within a few days, the Company B team used Instrumental to identify an interference issue that was preventing the top cover from obtaining a solid bond to the enclosure. While this ultimately underscored a design flaw, the team needed to implement a solution that would enable them to produce, ship, and deliver a reliable product before Christmas. They planned to fix the underlying design issue in future revisions of the product.
The Company B product team used Instrumental’s data to verify several new DOEs to overcome the design flaw and ensure a reliable assembly. In this case, they were able to select new adhesive and tune their process to increase adhesion. The resulting units for both SKU A and SKU B passed all reliability tests.
In tandem to solving the top cover issue, during PVT and early ramp, an engineer discovered that some of the buttons had a “mushy” feel when pressed. “Where do the two centers line up?” asked the product design lead. “The problem was not just that the diameters weren’t matching, but basically what you have is a smaller diameter and somewhere the larger one is going to seat. Where do the two centers end up.”
The Company B product design lead used Instrumental data to investigate. He found that the units with mushy buttons also had poor button-to-switch alignment. He was also easily able to verify that buttons with great click feel were well-aligned. The product design lead further discovered that there was a large amount of variation in the placement of the flexible PCB that had the switch mounted on it. The team was able to address the issue quickly by improving the fixtures that located the switch in the enclosure.
In the end, Company B was very pleased with how the Instrumental system was able to help them get to the bottom of a tricky failure, but also to assist with other issues that they initially did not anticipate.
As Company B moves into the next generation of their product development, they have a new tool in their toolbox that will enable them to understand variations in their process. The insights they gained from their work with Instrumental have already informed future product and design iterations.