In discrete manufacturing today, when product design engineering and quality teams discover units in their manufacturing lines that are failing functional tests, it can take days and sometimes weeks to simply find the root cause of the issue, let alone start engineering a solution. A recent survey conducted by Instrumental revealed that, in an average week, engineers can spend up to 75% of their time on tasks that are not engineering – things like hunting down data stuck in siloes or endless email back-and-forths across time zones to try to understand what’s happening.
Engineers often need to dedicate hours to communicating with factories in different time zones, soliciting data and images from their product lines. The process is often inefficient and typically falls to the most experienced engineers within teams.
All of this slows down the problem-solving process and is a major hidden cause of schedule delays.
Engineers want to engineer, but a survey of 100 teams revealed that much of their time is spent collecting data, communicating, traveling or waiting for samples to arrive, and creating decks and status updates.
Instrumental’s mission is to continuously improve how companies manufacture physical products with data, and the process of getting to root cause quickly seems ripe for optimization.
With the introduction of Discover Relationships, we’re extending our artificial intelligence to not only help identify new issues when they crop up but actually automatically suggest possible root causes. Instrumental does this by automatically scanning terabytes of your data to find the visual anomalies that correlate most tightly with anomalous test populations, sorting them in real-time, and providing you with a ranked list of possible root causes.
The tool automates away hours of engineering time spent gathering and analyzing data and helps quickly narrow down the possible sources of issues by where in the product they appear, and when in your process they occur.
Discover Relationships is one more way Instrumental takes the guesswork out of finding and fixing issues. Your builds are more likely to keep on schedule when issues are discovered faster and earlier, accelerating time to market.
One example of a tricky issue that many product design engineers will know well is partial frequency sweep failures on microphones. Many wearable electronics include microphones, and anomalous frequency responses can be particularly challenging to understand because of the broad variety of possible root causes – the microphones themselves could be defective, or they could be interacting in a variety of ways with parts around them, like glues and seals, which may produce intermittent issues.
Because the cause is often related to glues or interactions between parts, mic issues can be particularly resistant to teardown analysis. Once you open up the unit, you’ve often obscured the root cause.
In test data sets, Instrumental’s Discover Relationships AI was able to pinpoint multiple root causes of partial microphone failures within a defective sample population, surfacing anomalies related to seals, glue intrusion, and a molding issue – giving engineers three clear paths to follow to design solutions.
It was able to do this with its machine learning capabilities that can analyze thousands of images on a product line. An engineer scrolling through data or zooming into an image would have been unlikely to catch and determine this as the root cause, and in production, this sort of issue could have plagued yield on a line for months.
With the addition of Relationship Explorer, Instrumental empowers engineering teams to do more:
- Navigate through product data more quickly to get context about what is occurring;
- Find what is causing product defects and issues;
- Dive deep into collected data with histograms and intelligence at their fingertips; and
- Access functional data for a failure test faster; and
- Filter data down to specific test metrics.
To find out more about Discover Relationships or get a demo, contact the Instrumental team here.