HALT is a method aimed at discovering and then improving weak links in products in the design phase. HASS is a means of finding and fixing process flaws during production. These methods are discovery testing in which you find problems by testing to failure. The old paradigm of qualification testing was one of trying to pass. Any failure that occurred was usually declared to be unusual and therefore not relevant. This latter approach is called success or compliance testing by the author. The HALT and HASS techniques represent a paradigm shift of major proportions. Companies using these evolving techniques correctly have obtained outstanding reliability.
HALT is an acronym for highly accelerated life tests that was coined by the author in 1988 after having used the term ‘design ruggedisation’ for several years. In these tests, every stimulus of potential value is used to find the weak links in the design and fabrication processes of a product during the design phase. These stimuli may include vibration, thermal cycling, burn-in, voltage, humidity and whatever else might expose relevant weaknesses.
The stresses are not meant to simulate the field environments at all but to find the weak links in the design and processes using only a few units and in a very short period of time. The stresses are stepped up to well beyond the expected field environments in order to obtain time compression in finding design weaknesses. HALT has, on many occasions, provided substantial (five to 1000 times) MTBF gains even when used without production screening, reduced the time to market substantially and also reduced the total development costs. The basic philosophy of HALT has been in use by the author since 1969 and he has used it on hundreds of products and his seminar attendees have used them on thousands more.
HASS is an acronym for highly accelerated stress screens that was also coined in 1988 after using the term ‘enhanced ESS’ for some years. These screens use the highest possible stresses (frequently well beyond the ‘QUAL’ level) in order to attain time compression in the screens. Note that many stimuli exhibit an exponential acceleration of fatigue damage accumulated with stress level1, and so a drastic reduction in screening equipment and manpower is obtained by the use of higher stress levels. The screens must be, and are proven to be, of acceptable fatigue damage accumulation or lifetime degradation using Safety of HASS techniques1.
HASS is generally not possible unless a comprehensive HALT has been performed as, without HALT, fundamental design limitations will restrict the acceptable stress levels to a very large degree and will prevent the large accelerations that are possible with a very robust product. It has been proven that HASS generates large savings in screening costs because much less equipment (shakers, chambers, monitoring systems, power and liquid nitrogen) is necessary due to time compression in the screens. HASS, too, is discovery testing as compared to success testing.
The phenomena involved
Several phenomena are involved when screening occurs. Among these are electromigration, chemical reactions and mechanical fatigue damage. Each of these has a different mathematical description and responds to different stimuli.
Chemical reactions and some migration effects proceed to completion according to the Arrhenius model or some derivative of it. It is noted that many misguided screening attempts assume that the Arrhenius Equation always applies; that is, that higher temperatures lead to higher failure rates, but this is simply not an accurate assumption. MIL-HDBK 217 is based on these concepts and therefore is quite invalid for predicting the field reliability of the products built today. MIL-HDBK 217 is even less valid and completely misleading when used as a reverse engineering tool to improve reliability, as it will lead one to make changes such as cooling that may well reduce reliability due to the introduction of new failure modes in the cooling system.
The fatigue damage done by mechanical stresses due to temperature, rate of change of temperature, vibration, or some combination of them can be modelled in many ways, the least complex of which is the Miner’s Criterion. This criterion states that fatigue damage is cumulative, is non-reversible, and accumulates on a simple linear basis; to wit ‘The damage accumulated under each stress condition taken as a percentage of the total life expended can be summed over all stress conditions. When the sum reaches unity, the end of fatigue life has been reached and failure occurs’. The data for percentage of life expended is obtained from S-N (number of cycles to fail versus stress level) diagrams for the material in question. A general relationship based on the Miner’s Criterion follows:
D ≈ nσβ, where:
D is the fatigue damage accumulated.
n is the number of cycles of stress.
σ is the mechanical stress (in pounds per square inch, for example).
β is an exponent derived from the S-N diagram for the material. β ranges from eight up to 12 for most materials in high cycle fatigue (low stress and many cycles to failure).
The flaws (design or process) that will cause field failures usually, if not almost always, will cause a much higher than normal stress to exist at the flaw than at a position without the flaw.
Just for illustrative purposes, let us assume that there is a stress that is twice as high at a particular spot that is flawed due to an inclusion or void in a solder joint. According to the equation above, the fatigue damage would accumulate about 1000 times as fast at the position with the flaw as it would at a non-flawed position. This means that we can fatigue and break the flawed area and still leave 99,9% of the life in the non-flawed areas.
Our goal in environmental stress screening is to do fatigue damage to the point of failure at the flawed areas of the structure. With the proper application of HALT, the design will have several, if not many, of the required lifetimes built into it and so an inconsequential portion of the life would be removed in a HASS. Note that the relevant question is ‘How much life is left after HASS?’ not ‘How much did we remove in HASS?’ Also note that all screens remove life from the product. This is a fundamental fact that is frequently not understood by those unfamiliar with the correct underlying concepts of screening.
The application of the techniques mentioned generally is very much enhanced by, if not impossible without, the use of environmental equipment of the latest design such as all-axis broadband random vibration and very high rate thermal chambers (80°C/minute or more product rate). Both of these techniques, HALT and HASS, have been in use by some of the author’s consulting clients for several decades, using the early all-axis shakers for about 25 years and the more modern and effective systems in later years.
Any of the pneumatically driven shakers do fatigue damage much more rapidly at the same GRMS level than do ‘classical’ shakers which usually are set to clip acceleration peaks at 3 sigma and therefore prevent cost effective screening. The repeated impact shakers such as the Modular Vibration System by HALT&HASS Systems Corporation have a peak to RMS ratio of about 10, whereas the classical electrodynamic shakers have a ratio of about 3 (when set to clip at 3 sigma).
Note that we are trying to do fatigue damage in a screen, and the more rapidly we do it, the sooner we can stop and the less equipment we need to do the job. It is not unusual to reduce equipment costs by orders of magnitude by using the correct stresses and accelerated techniques. This comment applies to all environmental stimulation and not just to vibration. An example given in the seminar, Comprehensive HALT and HASS, shows a decrease in cost from $22 million to $50 thousand on thermal/vibration chambers alone (not counting power requirements, monitoring equipment and personnel) by simply increasing the rate of change of temperature from 5°C/minute to 40°C/minute.
Another example shows that increasing the RMS vibration level by a factor of two would decrease the vibration system cost from $100 million to only $100 thousand for the same throughput of product. The use of an all-axis shaker would further reduce the cost ratio. With these examples, it becomes clear that HALT and HASS, when combined with modern screening equipment, provide quantum leaps in cost effectiveness, which is precisely why most of the leaders in screening techniques are not publishing.
Some typical results of these screening techniques applied to product design and manufacturing are as follows:
1. An electromechanical product’s MTBF was increased approximately 1000 times when HALT was applied. A total of 340 design and process problems were identified in the several HALTs that were run, and all of these identified problems were removed from the product before production began, resulting in an initial production system MTBF of 55 years on a product that wore out in five years. This means that most products never had even one failure before wearout.
2. HALT found, using only four units in just a few weeks, 97% of the problems which were later found in an extended life test lasting 16 weeks and involving 12 units run 24 hours per day under normal conditions. The one problem not found in HALT was missed due to a technician reapplying grease to a lead screw every evening without my knowledge. No corrections were made to the product until after the life tests as the designers refused to believe that failures caused by HALT were relevant until these same failures were found under normal operational conditions. This reluctance to address identified problems because they were found by ‘over spec’ stresses is a typical tendency of those unfamiliar with the modern methods, and why a paradigm change through education is necessary for the methods to be effectively applied.
3. HALT, in a three hour demonstration associated with a seminar, detected and allowed solutions to three real design problems in three different pieces of equipment which had been fielded for years and which had had many field failures, two mission critical (safety of flight) and the other one disabling (grounding the aircraft). That equates to one major problem found per hour. The manufacturer had not been able to duplicate the field failures, although extensive classical testing had been done for several years, and therefore could not understand the failure mode and conceive the corresponding fix. All three failure modes were found ‘over spec’, two at temperatures slightly beyond spec and one in six-axis vibration in 10 minutes at four times the specified GRMS.
This is the first part of a two part article on HALT and HASS. More successful applications of this technology, as well as more information about the technology itself, will be published in a future edition of Dataweek.
1. HALT and HASS, Accelerated Reliability Engineering, available from Hobbs Engineering Corporation.
© Technews Publishing (Pty) Ltd | All Rights Reserved