Organization: GHI Systems, Inc.

Describes vibration fatigue spectrums and their relevance to evaluating the fatigue producing ability of HALT/HASS (6DOF) machines. Explains why gRMS is not a valid metric for comparison of fatiguing uniformity of machine tables, fixtures, or between different machines. The paper proposes a vibration fatigue spectrum metric to measure the baseline ability of any machine to perform at a predetermined standard effectiveness in order to improve screening uniformity. The paper also describes adjustment factors that may be used to compensate low machine effectiveness and thus provide for machine operation at a standard level. A significant portion of the paper is allocated to describing examples of fatigue comparison between different machines and gives two specific cases:

1) Between two machines of same manufacture but different configuration, and

2) between two machines of similar configuration but different manufacture.

     Title: Managing Partner

     Organization: Ops A La Carte LLC



On-Going Reliability Testing (ORT) is currently one of the tools of choice for reliability engineers in the manufacturing phase of a products life cycle. But how was it chosen and is it really an effective tool? These days, most companies are finding that tools such as ORT and burn-in, once effective tools, are starting to become much more ineffective as component failure rates decrease and manufacturing variability increases.
Highly Accelerated Stress Auditing (HASA) is a better alternative for three reasons: 1) the stresses being applied have been tailored specifically for the product; 2) the stresses are accelerated enough so that information learned is timely and can affect current shipments; and 3) the audit program is being measured and adjusted to fit the current defect rate for the product.
OUTLINE
• What is ORT
• Why is ORT currently being used
• Is ORT working for anyone?
• What is HASA
• How to move from ORT to HASA and achieve much better results
• Success stories

     Organization: Bose



This tutorial introduces the use of the well-known technique of Fault Tree Analysis as a tool in reliability modeling and analysis of an electronics of mechanical design (including software), identification of potential failure modes that are high contributors to product unreliability, tradeoffs and mitigation of those failure modes. Applied early in the product design phase, this activity allows for relatively inexpensive and easy design and manufacturing process improvements and, in that manner, achieving considerable improvement of the product reliability before the design is completed or the product is manufactured. Reliability improvement of the product in this manner can be measured (monitored), as reliability is continuously estimated by simultaneous reliability modeling using the Fault Tree Analysis techniques. When analytical methods are not possible or are not practical, quantitative accelerated testing is used to determine probability of occurrence of particular failure modes of some components or assemblies necessary for estimation of the overall system probability of failure. Also, some qualitative tests such as HALT or other failure mode identification methodologies may be used to confirm presence of a potential failure mode considered a high contributor to the product probability of failure in its use environment. A real example of this analysis as applied to audio products is discussed along with the achieved reliability improvement.
 

     Title: Engineering Development Manager

     Organization: Entela Inc



Accelerating time to failure is the primary focus of an Accelerated Stress Test. The key method is to increase stress on a part. Common sources of stress are vibration, temperature and voltage. But there are many others. This paper will explore obscure and unique stress sources such as biological material and hexadecimal data words from over 80 accelerated tests. How they were applied, and what was learned from them.

     Title: Testing Engineer

     Organization: Entela Inc



There are many questions arising about the validation and verification of the FMVT or HALT testing process. In the past many companies have had disappointments with these testing methodologies, largely due to the improper technique and lack of a good test plan. Thus, validating the test is of great importance, as it will confirm that the HALT or FMVT setup and execution as a viable and valuable tool for identifying and reproducing warranty, design and production failure modes. Validating the test requires demonstrating that the test setup will accomplish the goal of the test.
A validation test can be done using two samples (Sample 1 & Sample 2) for which field data is available. Both go thru the each level of testing described in the test plan separately. The first run with sample 1 is the Pilot run and the second run with sample 2 is the Validation run. Several points for validation can be drawn from running the test: 1) Repeatability of the times to failures and the types of failure between the two runs; Reproduction of warranty failures in the both runs of the test.
This paper explains in detail the validation procedure along with a case study example to give a better insight on the process and methodology.

     Title: VP of Quality

     Organization: C&D Technologies



Power supply reliability is crucial to overall system reliability, and lack thereof is sometimes the top cause of system failure. There are key steps to ensure success here. An introduction to this is made by defining reliability and different ways of calculating MTBF.
A list and explanation of 49 steps for controlling the design and manufacturing process for high reliability follows. This includes comprehensive spec development, good design processes, and proper component selection. Thermal aging in magnetics with powdered iron cores is discussed, with steps to prevent this. SMT capacitors require special consideration to prevent cracking; there is clear methodology defined to prevent problems here.
Testing processes and their impact on reliability are discussed. Whiskers including zinc, cadmium, and tin may affect reliability. This tells where they occur, what problems they cause, and how to best deal with them. A brief discussion of corona and how it is a phenomenon not only in high voltage supplies, but also with progressively higher power densities is becoming a factor in lower voltage supplies. A number of steps to prevent this are included.

     Organization: Entela Inc



Accelerated Stress Testing is often used as a means of causing latent product defects to emerge where they can subsequently be eliminated through product redesign. However, one of the fundamental principals of this technique is that defects will only emerge in response to conditions imposed. Therefore, for those latent defects that can be activated by imposed vibration or a combination of vibration and other stress sources, it is imperative to use the broadest range of input frequencies possible. This presentation will show the effects of natural frequency and resonance on product displacement and strain. Examples will also be given of how the use of low frequency input generated by the FMVT machine resulted in repeatable failure modes in electro-mechanical products, while the higher frequency input generated by air hammer tables while the higher frequency input generated by air hammer tables did not result in the same failures. Likewise examples of how high frequency energy generated by air hammer tables produced failures in surface mount components, while the FMVT machine with low frequency was unable to produce the same failures.

     Title: President

     Organization: Smithson & Associates



The imperatives and justifications for considering the fatigue domain for HALT/HASS and end-use environments are presented. Technical, economic, competitive and strategic issues and potential are addressed.

     Title: CTO

     Organization: Data Flare Inc



Pneumatic vibration test systems are increasingly being adopted as part of companies accelerated testing programs. All the manufacturers have a basic problem in that the failure of one or more of the pneumatic actuators will affect the test system vibratory output without any notification to the user that a problem exists. The failure of actuators directly affects the vibratory energy spectrum and in turn the efficacy and repeatability of the test. This presentation will explore the following:

• The causes of actuator failure
• The reasons why there is no user notification
• If I do not know about it is it really a problem?
• The effect on the vibratory output of the system
• The effect on testing and screening
• Costs to users of actuator failure
• Methods to identify individual actuator failure
• Summary and conclusions

The goal of an AST (Accelerated Stress Test, or Overstress Test) program is to make cost effective improvements in the field reliability of the hardware being tested by selective use of stresses beyond the design limit stress of the product. This paper sets out a framework for determining the ROI (Return on Investment) for HALT tests, using practical examples from the computer and telecommunications industries.
 

Reliability programs that rely on reliability experience with a mature technology often cannot respond fast enough to the needs of many new products, where the product cycles are short. Environmental stress screening methods had been developed in military and space programs to achieve high reliability without waiting for field data. For industrial electronics, these methods were often too expensive and had been modified into accelerated stress testing to guide product design and development to achieve product robustness even though products are developed with new processes under short product-cycles. It was also called HALT but had been empirical. Studies had been conducted by various people in computer and telecommunication manufacturers, and a theoretical explanation was also added by the author. The use of stress testing to improve product design, component quality and manufacturing process has now spread quite widely to numerous manufacturers in network, computer, health electronics, and auto-mobile. Such product quality improvement is a cost effective business investment.
This tutorial provides a fundamental understanding of concepts and techniques that are applicable to the current and future market, and a high level view of the equipment, processes and results involved in accelerated stress testing.

     Title: Reliability Engineer & Director (respectively)

     Organization: MBtech E/E North America



Accelerated tests use shorter time to failure at elevated stress levels to shorten the testing time. However accelerated testing does not demonstrate reliability and provides no reliability metric.
A PoF (physics-of-failure) methodology can assess the failure risk for a proposed product when it is subjected to its anticipated life cycle and determine whether the anticipated reliability is achievable. Attributes for such methodology include dominant failure mechanisms, stress drivers for failure, and severity ranking of the times to failure of the dominant failure mechanisms.
The PoF methodology involves identifying potential failure mechanisms, failure sites, and failure modes and identifying and applying the appropriate failure mechanism models to compute the effective time-dependent reliability function.
We present in this paper, how a methodology for reliability assessment based on Physics-of-failure can incorporate a probabilistic approach to enable failure assessment under uncertain life cycle loads, physical dimensions and material properties by modeling the distribution of failure free operating period (FFOP). Mean time between failures (MTBF) is often used as reliability metric but it is usually based on exponential distribution and assumes constant failure rate. FFOP is a more useful measure of reliability in contrast. True value of parameters can only be known if we have infinite samples, infinite test time, and infinite number of tests – all of which are impossible in real life. Therefore, the approach presented uses a distribution for the purpose of estimating FFOP.
Inputs such as material and geometric properties of the electronic control module are given as a distribution and a Monte Carlo simulation is used to generate time to failure data. This data is used to determine the FFOP and the steps are repeated for various confidence level values. Based on the frequency of occurrence of various values of FFOP, a distribution for FFOP is obtained and FFOP values can be finally determined with their confidence level.
This probabilistic PoF approach provides information about the ability of the product to meet the required performance specifications in its life cycle application environment for a specified period of time. It allows an optimal selection of the component parameters (e.g., dimensions, materials) so that the minimum time-to-failure of any component is greater than or equal to the expected product life.

     Organization: QualMark



Not all pneumatic vibration tables are the same. Energy levels can vary across the table surface and from machine to machine for the same control levels. Presently there is no standard method of measuring these differences. The purpose of a “Standardized Method” is to allow the practitioner to evaluate the consistency of stresses applied to a product from one table location to another and from one vibration table to another. This knowledge enables the practitioner to make rational judgments regarding moving product and fixtures from one location on a table to another and from one machine to another and understanding the ramifications of the stress levels applied by different machines.

      Organization: Screening Systems, Inc.



The following presentation shows how a HALT test surfaced some important design weaknesses and defects in a new OEM RF Electronic Product. One of the unusual features of the RF design is that the stresses that were found to be killing the product could not be stress relieve in the usual sense. Analysis and HALT testing together had to be used to find alternate solutions.

                By:E. Suhir

In a simple demonstration experiment, a 500 kHz oscillator circuit was designed using an inexpensive timer and passive components. The functionality of this circuit is easily monitored with a frequency counter.
Ten oscillators were fabricated from components with minimal operating temperature ranges and high thermal sensitivities. Without changing the fundamental circuit layout or fabrication techniques, ten additional oscillators were fabricated from components with extended operating temperature ranges and reduced thermal sensitivities.
Both populations were subjected to the HALT process, and it has been demonstrated that there is a statistically significant difference in the operating limits and destruct limits of the two designs.

     Title:Reliability Engineer

     Organization: General Motors



An acceleration factor relates an increase in applied stress to a reduction in test time. The magnitude of a typical acceleration factor allows for a dramatic reduction in test time, sample size, and testing costs. This concept also permits faster and more frequent “learning cycles” during product development yielding mature and reliable products upon entry to the market place.
This discussion will focus on the acceleration factors most useful for electronic products. Thermal cycling with increased ramp rates along with accelerated vibration testing will be discussed. Examples using real life conditions will be explained during the presentation with the audience given the opportunity to participate. How to combine acceleration factors will also be explained through practical example.
Do you have a special case where no acceleration factor is currently available? No problem…. I will show you the process for generating your own acceleration factors quickly and easily. Explanation of the Calibrated Accelerated Life Testing (CALT) method will be given for the special situation where one is trying to validate to a requirement that is multi-dimensional. A multi-dimensional requirement would typically require different levels of a stress with varying frequencies of application. The CALT method does what traditional Weibull analysis can never do, and when combined with Miners Rule, is a “must learn” for the reliability engineer with a passion for the discipline and technical excellence.

 

     Organization: Teradyne



The high air flow rates in modern HALT chambers increases the risk of a negative pressure in the chamber cavity. A negative pressure translates to a condition known as sucking. Sucking translates to moisture on your test sample. Moisture translates to a test unit failure that is considered chamber damage and it reflects very badly on the HALT testing staff. It is wonderful when HALT finds a failure but it is not wonderful when HALT was the root cause of invalid failures.
This paper will explain:

• Why high air-flow rates automatically cause negative pressures in the chamber system
• How to tell why some chambers have more significant negative pressure problems than other chamber designs
• Clues to tell you if your chamber test cavity has a negative pressure
• Operational advice to minimize moisture damage
• Evidence of moisture damage
• Possible modifications to your chamber to minimize negative pressure
• Proof that the recommended chamber modifications work