ACCELERATED TEST DEVELOPMENT FOR COMBINED STRESSES

Project Number : C95-13
Point of Contact : dasgupta@calce.umd.edu
T. Rothman, K. Upadhayula, P. Dujari,
A. Plummer, B. Balachandran, A. Dasgupta

Objective

Investigate cost-effective schemes for conducting accelerated stress testing, based on physics- of-failure (POF) considerations, for design/manufacturing qualification and for process verification. This is intended to be a three-year task and will focus primarily on interconnect failure mechanisms at the circuit card assembly (CCA) level. This includes solder joints, leads, PTHs/vias, traces, connectors, PWB substrates, attachments to heat sinks, etc. Specific issues to be addressed include the effectiveness of (i) simultaneous versus sequential application of multiple stresses; and (ii) electrodynamic (ED) shakers versus repetitive-shock (RS) shakers, for generating vibrational loads. In objective (i), we will focus our attention on combinations of temperature cycling and mechanical vibration. Other stress types are deferred to future studies. The objectives for each year are as follows:

Year 1: Qualitative experimental phase to determine (i) effects of simultaneous vs. sequential application of combined temperature and vibration stress parameters, and to compare the effectiveness of RS vs. ED shakers, for accelerated vibration stress application.

Year 2: Quantitative phase when POF models will be combined with the results of phase I to present a comprehensive methodology for deriving acceleration transforms for different field environments such as dual-use automotive, aerospace, and telecommunication environments.

Year 3: Documentation phase to incorporate the findings of the first two phases into standards, software tools, and guidelines for design and testing.

Background

Accelerated stress testing has been recognized to be a necessary activity to ensure the reliability of high-reliability electronics. The application of enhanced stresses is usually for the purpose of (i) ruggedizing the design and manufacturing process of the package through systematic step-stress and increasing the stress margins by corrective action (reliability enhancement testing); (ii) conducting highly compressed/accelerated life tests in the laboratory to verify in-service reliability (Accelerated Life Tests); and (iii) eliminating weak or defective populations from the main population (screening or infant mortality reduction testing).

There is significant confusion in the literature, in the standards, and in industry, regarding the most cost-effective and scientific way to conduct accelerated stress testing for electronic packages exposed to multiple stress environments. There are many unanswered questions about the advantages of applying sequential stresses vs simultaneous combined stresses. There is also significant controversy regarding the use of RS vs. ED shakers for accelerated vibration testing. Perhaps the most critical issue is a rational method to relate the test results quantitatively to in-service reliability, using a scientific acceleration transform. In other words, the amount of test-time compression achieved in the accelerated test must be determined quantitatively.

CALCE EPRC has demonstrated in previous years (C94-14: Task 4), how experimental data can be combined with POF models to obtain maximum information from a minimum amount of test data, and to derive acceleration transforms for single stress environments. These methodologies will be systematically explored in this project to address combinations of stresses.

Approach

TEST SAMPLES: Test specimen CCAs have been solicited from members. Test specimens are either fully functional or consist of daisy-chained components to facilitate monitoring of electrical failure modes. The primary focus is on interconnect failures but other failure modes are also be monitored. Specimens will be sought with mixed insertion-mount and surface mount technologies. Peripheral I/O format (DIP, SIP, axial-leads, gull-wing, J-lead and leadless) and area-array I/O format (BGA, CGA) interconnects have been examined. PTHs, vias, connectors and metallization traces also form an integral parts of the daisy-chained circuits. Mounting fixtures have been designed and fabricated with the help of member companies.

INSTRUMENTATION: During the first year, we acquired new data acquisition and post- processing capabilities. Both commercially available tool-kits and indigeniously developed tool-kits have been utilized. A GHI RS-CAT analysis system has been acquired and setup for controlling and quantifying the damage due to vibration testing in the 6 DOF RS shaker. This tool-kit has several time-domain, frequency-domain and fatigue analysis tools. The fatigue tools such as Accumulated fatigue damage function (AFDF), Peak probability Distribution Function (PPDF), and Rainflow Analysis for cycle counting will all be used during the following year to quantify the random vibration spectrum used in our vibration testing. A GENRAD controller has been acquired and installed for similar control of the 1 DOF ED shaker. An in-house data-acquisition facility consisting of 3 triaxial assemblies of accelerometers has been designed and implemented to monitor the distribution of acceleration histories across the sample CCA. Additionally, strain gages are used to obtain similar distributions of board curvature histories. Thermocouples are used to obtain distributions of temperature histories. These measurements are collected in digital formats and will be used as inputs to POF models during the following year, for assessing the damage accumulation and for computing acceleration transforms. Clearly, the transform will depend upon the failure mechanism being addressed.

Electrically functional boards are be monitored for failures using hardware/ software provided by the sponsoring company providing the specimen. Specimens with daisy-chained components are monitored for interconnect failures using Anatech event detectors.

TEST PLAN: A fractional factorial pseudo-design-of-experiment matrix consisting of 5 load cases has been identified to address the objectives defined above. Both test plans are illustrated in the enclosed overheads. The load cases consist of several accelerated stress tests involving simultaneous and sequential applications of 6 DOF repetitive-shock vibration (or 1 DOF electrodynamic vibration) and temperature cycling. The tests are run long enough to collect adequate failure data. Failure modes and times to failure are carefully documented for each occurrence. The GHI data analysis system and the indigeniously developed analysis system are used for on-line monitoring of damage metrics using the AFDF, PPDF and rainflow approaches. Since stress levels are different for each component on each sample CCA, POF methods will be used in following year to extract meaningful information at the individual failure sites, as illustrated in project C94-14. Results are used for qualitative assessment of the effectiveness of ED vs RS shakers for accelerated vibration testing, and the effectiveness of sequential vs.simultaneous application of vibration and temperature cycling.

In the second year, we will conduct detailed POF simulations using PPDF, cycle-counting and AFDF metrics to analyze all the test data. The analysis will reveal the pros and cons of running combined stress tests vs. sequential stress tests, and provide information regarding the relative performance of the different vibration methods used in the program. Most importantly, the analysis will provide quantitative acceleration transforms.

In the third year, we will prepare detailed documentation incorporating the accelerated testing guidelines. We will also update the CALCE software to help packaging engineers design cost-effective accelerated stress tests and to compute associated acceleration transforms from POF considerations, to interpret the test results.

Work Accomplished

Continued last year's accelerated test program on Hamilton Standard specimens (characterized CTEs of boards and updated IPC POF model to illustrate need for better POF model).
Identified 6 sets of candidate test samples for L5 test matrices for this year's program.
Acquired, setup and calibrated the new GHI controller and damage-monitoring tool-kit for 6 DOF RS shaker.
Acquired, setup, and calibrated new GENRAD controller for 1 DOF ED shaker.
Acquired samples based on the overall test plan, from all participants.
Completed fixturing and preliminary destruct level testing.
Setup in-house instrumentation for quantifying 3-D acceleration histories.
Conducted accelerated stress test program as per defined L5 test matrix on several samples, using simultaneous/sequential vibration and temperature stresses in RS and ED test facilities.
Collected test data for POF analysis in following year, and performed failure analysis on all failed samples.
Documented failure data for dissemination among members.

Note: Due to limitations of test sample availability and test equipment availability, the entire test matrix is not complete yet. We expect to complete these tasks by Dec 95. No addition funds are required to complete these tasks. A complete detailed report will be placed on the World Wide Web at that time.

Internal Reports

Accelerated Test Development for Combined Stresses