Determining the root cause of a failure is crucial so that the appropriate corrective actions can be taken to prevent reoccurrence or to remedy ongoing problems. The investigation of the failure should indicate if the failure was caused by materials, design, process, or conditions. There are a few tips that apply to engineers, designers, managers and technicians when evaluating failures. These tips apply to both internal investigations and also to situations where outside laboratories are contracted for analytical services.

Confirm Before beginning
The first step in an effective analysis of failures is to confirm that the failure has occurred. Confirmation may be as simple as reviewing test data or it may actually require re-testing of an assembly. Optimally, images of the sample should be obtained and recorded before the investigation begins. Isolating a failure helps the investigator by reducing time and the area of investigation. This increases the probability of identifying the root cause, increases the efficiency of the analysis, reduces cost, and prevents mis-diagnosis. One key theme to remember during the analysis is that documentation is critical.

Become a researcher
Determining the root cause of failures is like being a detective or investigative reporter. The key to a rapid and thorough investigation is gathering as much information as possible before disturbing the sample. Collect a sample history that includes component data sheets and the characteristics of the failure (short, open, etc.). Some key questions to ask when investigating component, board or assembly failures include:

• What is the suspected failure?
• What are the symptoms?
• Where is the failure located?
• What is in the area of the failure (next to and on the opposite side)?
• What was happening when the failure occurred?
• Was the device under power?
• Was it undergoing any stress testing (heat, cycling, vibration, etc.)?
• Did it work prior to assembly, prior to shipping, or did it ever work?
• How many failures are like this one?

If there are other similar failures, look for shared characteristics such as production lots, operators, manufacturing sites, process changes, a change in materials, a change in vendor or a change made by a vendor. Assumptions made about failure modes without gathering as much information as possible can result in incomplete or incorrect conclusions. The BGA failures like those observed from the board shown in Figure 4-1 were once thought to be a result of poor processing. Review of the reflow profile, storage conditions, component and solder paste data sheets, and BGA design guidelines revealed possible alternative causes of failure. The answers to the previous questions should help to build a chronicle that may then be used to focus the investigation.

Bring it into focus
Review all of the gathered information and form a hypothesis of what may have occurred. Many find it beneficial to use peers to brainstorm about theories of what could have occurred. The majority of problems that occur in electronics manufacturing are not new, and the chances of someone having experienced the same issue are high. Experiments can then be modeled from past experimental design. The experiments should involve proven techniques with methods and specifications that are clearly outlined.

Select test methods
At this stage, determining the conditions (possible events) that could have caused functional changes is a key objective. With this information, the investigator is now ready to choose the correct analytical techniques and test equipment to determine the root cause of the failure. There are usually multiple test methods available to test for a particular problem or defect. Complimentary test methods can be selected, but care should be taken to ensure that conflicting results are not produced. Table 2-1 outlines some of the most common test methods for BGA failures.

Perform testing to validate the hypothesis
Experiments should have a control or reference sample that can be used to validate the experiments. Design the testing to answer any questions created by the hypothesis.

Perform testing to disprove the hypothesis
Many overlook this step in experimental design. Utilize experiments that prove or disprove the counterpoint of the hypothesis.

Become a reporter
Once all the data is collected and determinations are made, communication is key. The use of written reports is the most effective method for relaying information. The reports should have data, images, graphs and bulleted conclusions that are easily explained with minimal wording. The purpose of the reports are not only to communicate the cause of failure, but also to record failure mode, cause and corrective action for future users. This allows for prevention methods to be established that avert future failures.