Accelerate your DFMEA with Sherlock ADA Software: A Delphi Case Study

Cheryl Tulkoff, DfR Solutions
Michael Varnau, Delphi

 

Download PDF

 

Abstract

This paper describes how the DFMEA module in the DfR Solutions’ Sherlock Automated Design Analysis software can accelerate and standardize a Component Level DFMEA based on the electrical schematic or parts list of an electronic module. The required user data definitions and general process flow are described, and the quality and flexibility of the results are demonstrated and reviewed.

Design Failure Mode and Effects Analysis (DFMEA) is the application of the Failure Mode and Effects Analysis method specifically to product design. It is an analysis method used in engineering to document and explore ways that a product design might fail in real-world applications. A DFMEA documents the key functions of a design, the primary potential failure modes relative to each function and the potential causes of each failure mode. The DFMEA method allows the design team to record and analyze what they know and suspect about a product's failure modes prior to completing the design, and then use this information to design out or mitigate the causes of failure.

What is DFMEA?

DFMEA is a systematic group of activities designed to recognize and evaluate potential failures of systems, products, or processes, identify the effects and outcomes of the failures, identify actions that could eliminate or mitigate the failures, and provide a historical written record of the work performed. DFMEA is just one of many available tools in the design for risk and reliability tool box. A schematic of the steps is shown is Figure 1 and shows the iterative and ongoing nature intended for the process.

Accelerate Figure 1

Contrary to common belief, DFMEA be used across a wide range of fields and disciplines including:

  • Manufacturing
  • Software
  • Business Processes
  • Healthcare
  • Service Industries
  • Regulated industries – automotive, medical devices, NASA

Basically, DFMEAs can be used in any situation where there is a need to reduce risk and prevent failure. Sherlock focuses on the use of DFMEA in electronics products.

The purpose of a DFMEA study is to analyze and mitigate what might go wrong, how bad the effects might be, and how they might it be prevented, mitigated or detected at the earliest possible moment with the lowest cost, impact, safety risk. Unfortunately, the typical DFMEA creation process is time-consuming and plagued by inconsistency, underestimated risk, and outright inaccuracy. A robust DFMEA requires an extremely comprehensive, element by element review of the design. For each design component or element, designers must list how failures can occur (Failure Mode), what could happen (Failure Effect), how processes or the system itself can detect & prevent the problem, and then generate recommendations for improvements. DFMEA is designed to be a cross-functional, team-based process. It is vitally important that employees with the correct skills and qualities are brought to the core team and that their roles are clearly defined. To maximize use of resources, support team members can be used to provide specific expertise as needed and team composition may change during the ongoing DFEMA process. Typical DFMEA team members include representatives from development, production, supplier quality, purchasing, and validation.

Key DFMEA Terms

A failure is defined as the loss of expected or intended function under stated conditions. The failure mode is the way in which a failure is observed; it generally describes the way the failure occurs. The failure effect is the immediate consequences of a failure on operation, function or functionality. The failure cause can be defects in the design, system, process, quality, or part application. It is the underlying cause of the failure or things which initiate a process which leads to the failure. Severity is defined as the consequences of a failure mode. Severity considers the worst case outcome of a failure as determined by the degree of injury, property damage, or harm that could ultimately occur. Probability or Frequency of Occurrence is the likelihood that a specific failure cause/mechanism will occur during the intended design life or usage situation. The likelihood of occurrence ranking number is a relative rather than an absolute value. Detectability describes the cause/mechanism of failure or the failure mode can be detected either by analytical or physical methods. It is another relative ranking of the intended design controls.

Once these elements have been defined for a design, a Risk Priority Number (RPN) for each element can be calculated using three criteria:

  • Severity of the Failure Effect “S” (Scale of Low - High)
  • Frequency of Failure Occurrence “O” (Scale of Infrequent -Frequent)
  • Detectability/Preventability/Warning “D” (Scale Very Detectable – Not Detectable)

Where RPN = S x O x D with a range from “Good” to “High Risk.”

Then, an unacceptable or actionable range/ high risk range must be defined. This allows a Pareto ranking of the RPN to be performed and used to prioritize corrective action efforts. As an example: RPNs greater than an identified value are deemed unacceptable and require a corrective action or redesign. The primary objectives of recommended actions are to reduce risks AND increase customer satisfaction by improving the design.

Common DMFEA Mistakes

Since creating a DFMEA is both time-consuming and complex, a number of mistakes are commonly seen in industry. One frequent mistake is using a “fill in the blanks” process where a designer quickly fills out a standard template as documented by a company or regulatory environment. This type of DFMEA offers little value. Developing a useful DFMEA requires a true cross-functional team. Even with a team though, some members don’t fully understand the scope and objective of a DFMEA. They cannot separate failure mode from cause or effect. Or, the group limits team members to direct designers who don’t rigorously challenge the design or its risks. Some teams create DFMEAs that simply identify the problems but neglect to document potential solutions or control plans. Oftentimes, mitigation becomes the default strategy when design changes should truly be considered. In some companies, a lack of consistency is seen between different design team who develop DFMEAs. The DFMEA process also fails when the ranking criteria are too loose or when risk is not adequately characterized. Finally, DFMEAs are meant to be “living” documents. They should be frequently updated, reviewed, and reused as appropriate rather than filed away.

How Automating DFMEAs Can Help

Automating the DFMEA creation process can result in significant engineering time savings. Default templates and entries can be created by experts allowing rapid, consistent, and rigorous development across team members, product lines, and company divisions. Automation allows the design engineering resources to focus on what truly matters – the critical few instead of the thousands of potential risks.

The Sherlock Automated Design Analysis Software has been developed with a DFMEA module that automates most of the time consuming DFMEA process potentially saving thousands of hours for a company every year. Sherlock is fast, smart and specializes in electronics design reliability. It takes standard design files, like bills of material, net lists, and Gerbers, and automatically prepopulates DFMEA templates with reference designators, component technology (capacitor, resistor, etc.), and failure modes (open, short). Sherlock also automatically performs a comprehensive review of pin location and net and removes non-relevant shorting events, such as electrical shorts within the same circuit net. It is not limited by the size of a design or the number of components. Complex board assessments can be dramatically simplified using a unique nesting capability that allows designers to more easily view the overall circuit as well as defined sub circuits and components.

Sherlock DFMEA Implementation

Sherlock automatically imports Net List data for ODB++ or Eagle CAD files. For other project types, users must supply a standard IPC-D-356 formatted file containing the Net List data. Once the file has been added to the list for a given circuit card, the “Net List (IPC)” file type is selected and Sherlock loads the net list data found in the file. Net List data associated with a given circuit card can be easily viewed by selecting the “Inputs >> Net List” entry in the project tree. An sample imported data file is shown in the picture below.

Accelerate Figure 2

Sherlock maintains DFMEA data as a hierarchical collection of the following types of data:

  • One or more subcircuits per circuit card
  • One or more parts per subcircuit
  • One or more failure modes per part

For maximum usability, DFMEA data is displayed and managed as a graphical data tree, allowing users to easily expand, collapse, browse and edit the data.

When creating a DFMEA in Sherlock, a dialog box appears allowing various processing options. Sherlock creates default failures based on part types, default “open” failures for all part pins, and default “short” failures for all pairs of pins. Simply select the “Update From Parts List” button within the software to create default DFMEA data entries for all parts in the parts list. The dialog box is shown in the next figure.

Accelerate Figure 3

Next, default failure values can be loaded using a pop-up menu. A capacitor example is shown in Figure 4.

Accelerate Figure 4

Custom failure modes can also be created by selecting a part and adding a failure mode. Once added, these custom modes become part of the user’s library and may be used again and again. The software also provides a number of DFMEA templates for designers to use. Unique templates exist for failure modes, parts, subcircuits, and upper level circuit card assemblies. These templates can be also be used to both export DFMEA data from Sherlock and import DFMEA data from an external spreadsheet into Sherlock. This allows templates to be shared between projects and allows company-developed criteria to be imported directly into the software.

The DFMEA Navigation panel is used to quickly find the data associated with any subcircuit, part or failure mode. As any entry is selected, data corresponding to that entry, and all higher-level entries, is displayed in the property panels. In addition to using the “Subcircuit Name(s)” property to organize parts in the DFMEA tree, it is easy to assign one or more parts to a specific subcircuit. The Navigation panel is shown in Figure 5.

Accelerate Figure 5

Once all DFMEA data has been input, DFMEA results can be generated by selecting the Analysis entry in the Sherlock project tree for a given circuit card and running it. RPN values greater than or equal to the defined “Error Cutoff Value” are colored as errors in red.

Risk Priority Numbers (RPNs) greater than or equal to the “Warning Cutoff Value” are colored as warnings in yellow. All other results are colored green. After the analysis completes, select the Analysis entry in the project tree to view the DFMEA results. Sherlock displays:

  • Summary Results
  • RPN Distribution Chart
  • DFMEA Ratings Table
  • Processing Issues

Typical summary results are shown below in Figure 6.

Accelerate Figure 6

The “RPN Dist” tab shows a graphical distribution of the maximum RPN values assigned to each part in the DFMEA data.

Accelerate Figure 7

The “Table” tab shows a list of all parts defined in the DFMEA tree and the maximum ratings assigned to the failure modes associated with those parts.

Accelerate Figure 8

The DFMEA module also generates graphical layers that display color-coded parts based on the maximum RPN value assigned in the DFMEA data. Figure 9 shows an example of the layers view.

Accelerate Figure 10

Case Study: Delphi

A Delphi DFMEA was created for a Gate Driver Board for a 3 Phase Inverter for an electric vehicle. The design bill of material contained over 521 components representing 1300 failure opportunities.

The data driving the Sherlock-generated DFMEA was set to represent a “Baseline” DFMEA that assumes that the circuits were typical and not safety critical. The basic Delphi DFMEA process was to perform the System/Subsystem Level DFMEAs and adjust the severity (SEV) rates of the “Baseline” Component Level DFMEA as appropriate to the System Level criticality, Severity and circuit function. The adjusted Baseline DFMEA then became the DFMEA of record.

Delphi has a well-established DFMEA process in place. They imported their pre-existing component level template and criteria into Sherlock as shown in figures 10-14. Figure 10 shows the basic template.

Accelerate Figure 10-1

Figure 11 shows the typical Delphi DFMEA definitions for the levels and properties.

Accleration Figure 11

Figure 12 shows the Delphi definitions for the part designator, type and material construction.

Acceleration Figure 12

Figure 13 shows a typical Delphi example of failure modes, causes, effects, and scale for severity, occurrence, and detection factors. Delphi used a combination of experience, historical data, and industry guidance to identify potential prevention and detection opportunities. Common component failure modes included opens, shorts, leakage resistance fails, and parametric shifts, Potential causes of the failures ranged from aging to damage to processing defects. Prevention tactics incorporated development of design standards, use of derating, stringent component qualification, and various forms of testing.

Accleration Figure 13

Figure 14 shows the Delphi subcircuit definitions for the design under evaluation. For purposes of this design, 70 subcircuits across 521 parts were identified.

Acceleration Figure 14

Finally, Figure 15 shows a completed DFMEA analysis for the Bus Bars Subcircuit.

Acceleration Figure 15

Once the initial templates and settings are established for a company, subsequent designs can take advantage of this to rapidly assess and perform detailed DFMEAs. At this point, the analysis is exported from Sherlock to a template which may then be used to assign actions, track status, and update the DFMEA as needed. Having an automated tool perform the initial data entry and calculations for the 1300+ failure opportunities allows the team to focus their efforts on prevention and/or mitigation of the prioritized risks.

Summary

Design Failure Mode and Effects Analysis (DFMEA) is a critical process in electronics design. It allows design engineers to identify and correct potential failure modes early in the design process to improve overall product reliability, decreases cost, and enhances customer satisfaction. With increasingly complex electronics components, the time required to complete a DFMEA can take weeks, adding to costs and product delays but the Sherlock automated design analysis software can significantly accelerate DFMEA efforts.