Highly Accelerated Life Test

from Wikipedia, the free encyclopedia

A Highly Accelerated Life Test (abbreviated to HALT , English for strongly accelerated load limit test ) is a qualitative test procedure with the aim of exposing electronic and electromechanical assemblies to accelerated aging while they are still in the development stage in order to be able to detect weak points and design errors . The knowledge gained in this way is incorporated into the device design and the production process, which improves the robustness and thus the reliability of the product. The procedure has been used since the 1980s.

history

The HALT process has its origins in Environmental Stress Screening (ESS) from the 1960s. The further development to a test procedure in which tests are carried out with loads above the product specification initially met with rejection and represented a paradigm shift. In the following time, various test procedures that followed this principle, such as the Stress for Life (STRIFE) or the Accelerated Stress Test (AST), described and applied. To differentiate themselves from other methods, Gregg K. Hobbs and his wife Virginia created the term Highly Accelerated Life Test and its acronym HALT in 1988. Hobbs had previously worked on the development of this process and the required equipment for years. Today the HALT process is mainly used by American companies to improve the reliability of products. It has also been established in various standards and norms such as IPC 9592A or DIN EN 62506: 2014-03.

Goal setting

Figure 1: SN diagram

The aim of a HALT is primarily to uncover weak points in the device design and manufacturing process within the shortest possible time and using a small number of prototypes. This is achieved by means of an extreme acceleration of the aging and damage process through exposure to temperature , vibration and rapid temperature changes. This means that failures occur early. Each vulnerability discovered in this way offers the opportunity to improve the device design or the production process. As a result of the improvement measures, the robustness of the product is increased and thus its reliability. Furthermore, the HALT shortens the development time and ultimately the costs.

Basic idea

The basic idea of ​​the HALT is based on the assumption that the error patterns and types of failure occurring in the HALT are the same as those later observed in the expected service life of the product. In addition, it is assumed that loads occur cyclically; for example, the ambient temperature outside is higher during the day than at night. The assumed relationship between the number of load cycles and failure in the field is shown schematically in Figure 1. The number of cycles and the load increase in the representation from the coordinate origin. Under the influence of the field stress S 0 , a product fails after the number of cycles N 0 . In the case of HALT, the load is increased to S 2 , for example , in order to reduce the number of cycles until failure to N 2 and thus to accelerate aging. This fact applies to all components of a product, there may be differences in the slope, and the relationship is often non-linear. However, this does not change this basic assumption for the product. In turn, stress cannot be increased indefinitely without changing the failure mechanism. This point is called the fundamental technological limit.

Equipment

Table 1: Technical parameters of the Star Galaxy 36 Deluxe test chamber
parameter size
Temperature range −100 ... +200 ° C
Heating rate Max. 70 K / min
Cooling speed Max. 70 K / min
Table dimensions 915 mm × 915 mm
Vibration bandwidth (fixed) 5 ... 10,000 Hz
Vibration bandwidth 2… 60 G rms
Chamber dimensions (W × H × D) 122 cm × 106.5 cm × 129.5 cm

For a HALT test, special HALT chambers are used in which the test item can be fixed on a vibration table. The test table can be randomly excited in six degrees of freedom using compressed air-operated hammers . The load energy of the vibrations generated in this way is given in G rms . The chambers have a temperature range of −100 ° C to +200 ° C, with possible temperature gradients of up to 80  K / min. This is achieved with powerful electrical heating elements and nitrogen cooling. So that the thermal energy can act as effectively as possible on the test object, it is conducted using air conduction hoses. Typical technical parameters of a HALT chamber are given in Table 1.

Test items

The test items that are to be used in the HALT must be prepared before the test so that their operation and function monitoring is possible from outside the HALT chamber. In addition, temperature and acceleration sensors can be attached to the prototypes used in order to be able to document the exerting stress. Protective mechanisms implemented in the DUTs may have to be disabled for the HALT, because otherwise the DUTs protect themselves and error images cannot be stimulated.

method

Figure 2: 360 ° STOP process

The procedure and the individual tests for the HALT are described below. The entire procedure and events that occur during the implementation of the individual test steps must be carefully documented. Any damage to the test item occurring during the tests must be subjected to a precise error analysis and result in improvement measures on the product. The damage can then be repaired so that the test can be continued with the corresponding test item. A defect is not a termination criterion, but represents a desired intermediate result. As shown in Figure 2, a HALT is a 360 ° process that ideally is carried out up to the fundamental technological limits.

Cold level test

The HALT is started with the cold level test because this is usually the least damaging effect on the test item. It starts at an ambient temperature of 20 ° C, which is reduced in 10 K steps until the lower operating limit for the temperature is reached. After that, the Lower Destruct Limit is determined, if possible. The respective residence time in each stage is approximately 10 minutes. This time is required so that the test item settles thermally and the function can be checked.

Heat level test

The heat level test is started at an ambient temperature of 20 ° C, which is increased in 10 K steps until the upper operating limit of the temperature is reached. Then the Upper Destruct Limit is determined, if possible. The respective residence time in each stage is approximately 10 minutes. This time is required so that the test item settles thermally and the function can be checked.

Temperature change test

During the temperature change test, the maximum possible heating and cooling output is cycled to and fro between two corner temperatures. The corner temperatures used in this test are based on the operating load limits (lower and upper operating limit) determined in the temperature level tests. At least five cycles are run through. The test item must be permanently monitored and checked for functionality. In addition, the test item can be switched on and off at the corner temperatures in order to generate further stress. The residence time for this test is also approximately 10 minutes at the respective temperature.

Vibration test

The vibration test begins at around 5 G rms and is increased by 5 G rms per step until the operational load and destruction limit is reached. The function of the test object must be monitored during the entire test, and the vibration stress acting on it is recorded via an acceleration sensor. The holding time is around 10 minutes in each stage. If a vibration level of 30 G rms and higher is reached without an error pattern having occurred, the function of the test item should be checked again at lower stress levels. There is the possibility that error patterns caused by high vibration levels can only be detected at lower vibration levels.

Combined stress test

After the individual stress tests have been completed, the combined stress test, which is made up of the superimposition of the temperature change and vibration test, can be carried out. This combined test uses the temperature profile of the temperature cycle test. For the vibration levels, the value of the destruction limit from the vibration test divided by five cycles is used. The importance of this combined stress test is based on the fact that the vibration behavior of many products changes with the change in temperature. As a result, different error patterns occur in this stress test than in the individual tests.

Test result and end

A HALT provides the following parameters related to the thermal and mechanical loads in the individual tests for a product:

  • Lower operating load limit ( L ower O perating L imit)
  • Upper operating limit ( U pper O perating L imit)
  • Lower destruction limit ( L ower D estruct L imit)
  • Upper destruction limit ( U pper D estruct L imit)
  • Fundamental technological limits ( F undamental L imit of T echnology)

The operational load and destruction limits with the associated margins on the design specification are shown schematically in Figure 3. It is not always possible to find all limits on one test. In such a case, the test time must be measured in such a way that any weak points should have been identified.

Figure 3: Schematic representation of the operational load and destruction limits

The error patterns and damage that occurred during the HALT are documented and analyzed. The knowledge gained is incorporated as improvements in the product design and the manufacturing process, so that in the end the robustness of the product is improved, the service life is extended and the failure rate is reduced.

criticism

The HALT procedure has disadvantages and for this reason gives cause for criticism. On the one hand, there is no correspondence between the loads in the test and the real environmental conditions that will later occur in the field. For this reason, errors occurring in the test must be subjected to a careful and careful analysis of the cause. It is important to clarify whether the error that has occurred can generally occur later with the loads in the field or does not have to be taken into account. To be able to assess this, experience and specialist knowledge of physical and chemical failure mechanisms and failure models are required. Furthermore, there is no scientific basis for inferring the service life of the product from statistical acceleration models and test results. Another disadvantage is that, in contrast to other test methods, this test technique is not sufficiently standardized and therefore the test conditions cannot be independently reproduced.

See also

literature

  • Gregg K. Hobbs: HALT and HASS - Accelerated Reliability Engineering . Hobbs Engineering Corporation, Westminster, Colorado 2005, ISBN 0-615-12833-5 .
  • Harry W. McLean: HALT, HASS, and HASA Explained - Accelerated Reliability Techniques . Amer Society for Quality, 2009, ISBN 978-0-87389-766-2 (Revised Edition).
  • Arno Meyna, Bernhard Pauli: Reliability Technology - Quantitative Evaluation Methods . Carl Hanser Verlag, Munich 2010, ISBN 978-3-446-41966-7 .
  • Mark A. Levin, Ted T. Kalal: Improving Product Reliability: Strategies and Implementation . John Whiley & Sons, 2003, ISBN 0-470-85449-9 .
  • H. Anthony Chan: Accelerated Stress Testing Handbook: Guide for Achieving Quality Products . Whiley-IEEE Press, 2001, ISBN 0-7803-6025-7 .

Web links

Individual evidence

  1. ^ Gregg K. Hobbs: Editorial - The History of HALT and HASS. (PDF) In: Sound & Vibration. October 2002, p. 5 , accessed August 10, 2014 .
  2. ↑ Arno Meyna, Bernhard Pauli: Reliability Technology - Quantitative Evaluation Methods . 2nd Edition. Carl Hanser Verlag, Munich 2010, ISBN 978-3-446-41966-7 , p. 587 .
  3. ↑ David Rahe: HALT and HASS in IPC9592A. (PDF) Qualmark & ​​DLi Labs, accessed August 15, 2014 .
  4. ↑ IPC-9592A - Requirements for Power Conversion Devices for the Computer and Telecommunications Industries . IPC, May 2010, p. 25-28 .
  5. ↑ DIN EN 62506: 2013-03 - Procedure for accelerated product testing (IEC 62506: 2013) . Beuth Verlag GmbH, Berlin 2014, p. 15-21 .
  6. ^ Gregg K. Hobbs: HALT and HASS - Accelerated Reliability Engineering . Hobbs Engineering Corporation, Westminster (Colorado) 2005, p. 4 .
  7. ↑ Mike Silvermann: How Reliable is Your Product? - 50 Ways to Improve Product Reliability . Super Star Press, Cupertino 2010, ISBN 978-1-60773-060-6 , pp. 177 .
  8. ↑ HALT and HASS - Test Equipment For Highly Accelerated Life Test and Stress Screen Applications. (PDF) Weiss Umwelttechnik GmbH, July 2, 2011, p. 11 , accessed on July 30, 2014 .
  9. ^ Neill Doertenbach: The Calculation of G rms . (PDF) QualMark Corp., accessed August 2, 2014 .
  10. ↑ Mike Silvermann: How Reliable is Your Product? - 50 Ways to Improve Product Reliability . Super Star Press, Cupertino 2010, ISBN 978-1-60773-060-6 , pp. 183 .
  11. ↑ a b c d e Mike Silverman: Summary of HALT and HASS Results at an Accelerated Reliability Test Center . In: IEEE - Reliability and Maintainability Symposium Proceedings . 1998, ISBN 0-7803-4362-X , pp. 30-36 .
  12. ↑ a b Harry W. McLean: HALT, HASS, and HASA Explained - Accelerated Reliability Techniques . Amer Society for Quality, 2009, ISBN 978-0-87389-766-2 , pp. 13-14 (Revised Edition).
  13. ↑ Harry W. McLean: HALT, HASS, and HASA Explained - Accelerated Reliability Techniques . Amer Society for Quality, 2009, ISBN 978-0-87389-766-2 , pp. 15 (Revised Edition).
  14. ↑ Harry W. McLean: HALT, HASS, and HASA Explained - Accelerated Reliability Techniques . Amer Society for Quality, 2009, ISBN 978-0-87389-766-2 , pp. 15-16 (Revised Edition).
  15. ↑ Harry W. McLean: HALT, HASS, and HASA Explained - Accelerated Reliability Techniques . Amer Society for Quality, 2009, ISBN 978-0-87389-766-2 , pp. 16-17 (Revised Edition).
  16. ↑ Luis A. Escobar, William Q. Meeker: A Review of Accelerated Test Models . Ed .: Institute of Mathematical Statistics. August 2, 2007, arxiv : 0708.0369 .
  17. ^ Wayne B. Nelson: Accelerated Testing - Statistical Models, Test Plans, and Data Analysis . John Wiley & Sons, Inc., New York 2004, ISBN 0-471-69736-2 , pp. 37-39 .
  18. ↑ Arno Meyna, Bernhard Pauli: Reliability Technology - Quantitative Evaluation Methods . 2nd Edition. Carl Hanser Verlag, Munich 2010, ISBN 978-3-446-41966-7 , p. 588 .
This version was added to the list of articles worth reading on September 8, 2014 .