blog banner temp.jpg

Blog

3 keys to efficient new product failure analysis

On Jan 31, 2019

In the test environment, your ability to find product failures and make necessary improvements efficiently is paramount. The quicker the feedback loop, the faster appropriate action is taken, which in turn, speeds up the product development process.

As a test engineer, you plan for a variety of errors in the process: however, many will overlook the key preparation areas needed to discover malfunctions quickly and determine the most efficient way forward in the improvement process. Some won’t even have the right test system that can provide them the clear and concise data needed to accurately characterize a new or existing product line.

This course of action can be especially crucial if you’re working with rotating electrics. During a test, motors are rotating, valves are actuating, coolant is getting pumped, software is bouncing messages to and from sensors, and real-time measurement devices must be reliably updated - all in a matter of milliseconds. Problems can arise quickly. This is what we anticipated when we designed and built durability test stations for a new PO HEV Motor Generator Device (you can read our full case study here). To ensure an efficient new product failure analysis, here are three essential aspects of your process:

Accurate failure timing records

This is where it’s essential to start with a test system that’s able to acquire data at a high-speed and store it for review. It’s also equally important that this system can shut down quickly once a device under test (DUT) failure occurs.

Here was our approach with our durability test station: we designed a “Flight Recorder” software system – modeled after the black box crash data collection unit in airplanes – that would stop within 400 milliseconds after any specified system failure mode occurred. Then, we ensured our system captured data for every 20-milliseconds over a 60-second timespan around the failure. This includes both the 30 seconds prior to DUT failure and the subsequent 30 seconds after the DUT failure. By isolating this timespan, a test engineer can take a more systematic approach when doing data analysis. Evaluating the data presented both before and after the DUT failure sets your team up to identify the critical problems throughout the qualification process.

No matter the system, your data acquisition and recording needs to be fast, accurate, easy-to-access and remarkably informative so problems can be spotted quickly, and fast improvements can be implemented.

Failure data traced to root cause

The first sign of failure is rarely the root cause of the problem. Once you’ve identified the time of failure and isolated the data, dig down to find the source of the issue. What are all the abnormalities you can find? Which came first? You want to be able to spot the first data point that was out of range and determine how that might have caused a chain reaction.

For example, with the DUT in our durability test station, the root cause could be an output device in the inverter shorting, which forced it to drive the DUT to full torque and it overstressed a belt and caused it to break. Our first sign of failure will be the obvious broken belt or an over-RPM measurement and alert. But by monitoring the inverter current, we would be able spot the abnormality in the output device and realize that’s what caused the real issue. Our “Flight Recorder” software was able to deliver real-time data so that test engineers could characterize performance of the DUT and compare to both the ideal units and other failed units.

Golden unit comparisons

Once you’ve got a test system running and you’ve found success some units, it’s time to determine your “golden unit,” the out-of-the-box product that has worked well and its data falls within specified limits through every test. Once you have that identified as a golden unit, that defines the product signature that you can measure all other units against. As you move forward through the test process, being able to compare the data of your best-performing unit will speed up your product development process.

With the golden unit in place, your product validation engineer then can answer the following question: Did the data for the current unit match up with the data for the golden unit?

Ensure you have the right test system

With the right test system, you move through the new product evolution process more quickly and with fewer misdirected efforts.

If you’re ready to design a new test solution and you’d like to walk through every step of the process you envision, get in contact with us here. We’d love to design a custom solution for you and ensure you have all the data you need to make your perfect product.

Contact Us

Blog Comments

Related Post