Most of us have accumulated a number of negative experiences with notebook computer failures that result in costly fees and a bad impression of a given company. The costs associated with the testing, maintenance, and repair of notebook computers are astronomical. Unfortunately rapid technological advancements in the field are adding to both the number and complexity of notebook failures and anomalies. A single notebook computer consists of a number of different components produced by different manufacturers with different qualification standards and methodologies. As the complexity of computer components increases and as the market for notebook computers – both personal and corporate – continues to expand, companies are fiercely struggling to reduce not only the costs, but also failure rates as to not lose market share or popularity with customers.
Computers used for business applications make up the majority of the market for notebook computers (between 80 and 90 percent). An analysis by J. Gold Associates estimates that companies experience at least 10 percent failures per year and that the cost of each failure is $865 USD. For a reasonably large company with 50,000 employees each using a notebook computer, that comes out to an annual repair cost of approximately $4.3 million USD. A study by Wipro Technologies shows similarly that the average annual maintenance cost of a mobile PC during its first year is approximately $716 USD. With 170 million notebooks sold worldwide annually in 2009, this equates to a total cost of approximately $121 billion USD of repair costs for computers sold in 2009 alone.
PHM is a methodology that can enable manufacturers to detect anomalies in notebook computer systems. Anomaly detection involves measuring deviation and degradation from a product’s normal operating state. It is a way to evaluate the product’s health and estimated lifetime.
Currently, anomalies in computers generally are not detected until the product finally fails to function correctly. At this point the customer typically returns the product to a retailer or manufacturer for inspection. The manufacturer will then test the product extensively to determine the location of the anomaly or fault, attempting to discover the root of the dysfunctional behavior. Unfortunately today’s notebook computers are extremely complicated, involving an integration of several sub-assemblies, components, and parts that are manufactured by different companies under different qualification standards. This means that retailers and manufacturers often fail to reproduce the exact anomalies or faults reported by the customers, making it nearly impossible to discover the root cause of failures. These faults are referred to as No Fault Found (NFF) failures, and cause much frustration in all areas of electronics.
If a PHM system could be implemented to continuously monitor different parameters (temperature, current leakage, turn-on voltages, etc.) of a number of components in the computer, this would free a lot of money, wasted time, and frustration involved in identifying the causes behind notebook computer failures. Not only would it make the repair process more straightforward, but it would provide useful data (such as correlations between component parameters and failure events) allowing manufacturers to quickly develop more robust, fail-safe designs in the future.
What are the first steps to developing a PHM system for notebook computers? Results of preliminary FMMEA indicate three high-risk failures within notebook computers:
Each of these failure modes is correlated with changes in a number of measurable parameters, which should be monitored to indicate the potential for failure. By running experiments where we monitor these variables and attempt to correlate them to system failures, we hope to develop a rudimentary PHM system for notebook computer health monitoring and maintenance.
