Industries

Root Cause Analysis on Return Merchandise Authorization (RMA) Boards

2021-11-04  |  5 min read 

Every manufacturing company strives to produce 100% passing boards with good quality.  With Industry 4.0 transformation, there are now more readily available data analytics to accelerate the effort towards this ultimate goal.

It’s common understanding that only passing boards in the production line get shipped out from factory to the market.  However, in some rare occasions, products or boards failed earlier than expected in the market.  And these early failure boards are returned to the factory for diagnostics; these boards are also often known as Return Merchandise Authorization (RMA) boards.

Within PathWave Manufacturing Analytics application, there is an RMA module which allows user to retrieve back test history of a board based on its serial number.  Identifying the issue for this early failure is important, but even more importantly is to identify potential quality risk of other similar boards out there in the market and to prevent such quality gaps in future.  As this may have diverse impact on company reputation.       

Once the serial number of the early failure board is entered into the RMA module, test history on this board will be retrieved and displayed as shown in Figure 1.
The test history will provide information like how many times this board had been tested with details including the Start and End Time, which fixture and equipment the board had tested on, together with its Pass/Fail status.  This allows traceability of the early failure boards. 
 

Figure 1: RMA - Measurement

Depending on which run is selected, the corresponding measurement anomaly, limit change and failure information (if any) will be displayed in the Anomaly Information section (see Figure 1).

In this example, there are multiple measurement anomalies detected during the run.  And for each measurement anomaly, its corresponding measurement trend is displayed in the scatter plot on the right.  The serial number board is highlighted in yellow.  Noted that a measurement anomaly is not a failing measurement.  In fact, it is a passing measurement within the limits; however, it is detected as an outlier from the population based on PMA anomaly detection model.  A measurement anomaly (which is a passing measurement) is definitely going to be overlook as compared to a failed measurement. 

Highlighting measurement anomaly here will definitely point user to a reasonable starting point of root cause analysis.  Being an outlier from the population may have higher chances of causing board functionality to go out of specifications.  From the opposite angle, not finding any measurement anomaly on highly possible test (like components related directly to the early failure), will indicate to user that other external factors may need to be considered.  

At the same time, further tracking and investigation can be performed on other boards which exhibit similar measurement anomaly as highlighted in red box in Figure 1; to determine if they pose potential risk out there in the market.
 

Figure 2: RMA - Limit Change

In Figure 2, under the Limit Change tab, PMA will highlight any limit change detected during the period when board was tested as unauthorized limit change may pose potential risk to product quality.  One can imagine the catastrophic impact when test limits have been altered incorrectly and a supposing failed board escaped out into the market, especially for live threating products in medical and automotive markets.  

Figure 3: RMA - Failure

In Figure 3, under the Failure tab, PMA will display test failures associated to the board for the run.  This helps user understand what failures and repairs have occurred previously for the board.   

These information allow user to perform root cause analysis on this early failure board holistically.

With this, I hope you get to know how PMA can help you on your next root cause analysis for RMA boards.
Do visit PathWave Manufacturing Analytics page for more information.

That’s all for now.  Stay safe.