Several factors affect the reliability of a product or the reasons for its failures searched. In this post we go through some of these factors.
1. Poor design
Sometimes the cause for failure may be due to poor design. The error may have occurred for example due to too small margins of safety or due to too optimistic design. Such mistakes may occur if the use environments and requirements are not adequately known. Too little knowledge of materials and components used or pressure to cut down the costs may also lead to selection of improper materials/components for the needed use conditions or use life. Sometimes reliability considerations are not properly analysed in design, and they are found only in later stages of production or in use.
Chosen plastic material has not withstood the heat caused by internal heating of the cable.
2. Manufacturing defects and mistakes
For complex products manufacturing may require lots of different steps and processes. These can significantly stress the products and some failures may be caused already during manufacturing of the product. For example, processing may cause internal stresses or thermal damage to the product, which results in a failure immediately or later in the hands of customer. Humanerrors may also happen in the manufacturing line, especially, if the process control is poor, which may cause poor reliability to the product.
One typical cause of failure is the change of some material/components used in the product. This can happen for example when material is no longer available or somewhere in the supply chain it is changed to a cheaper one. On paper the materials may look the same but in real life the new one may have much poorer properties than were originally designed for the product. For example, plastics with same name may have very different properties.
3. Quality problems
Sometimes certain batches of products have large number of failures even though most of the products manufactured are functioning fine. This may be due to a quality problem. Quality problems may be caused due to manufacturing errors, but also due to differences between supply batches. For example, the properties of materials may vary markedly, and this may cause problems for some batches. Therefore, properties of materials should be occasionally measured for quality control.
In addition to materials, quality problems also apply to components. Quality of components may be poor for certain batch or for certain supplier. Components may also be fake. In this case the insides of a component are not what they should be, and the manufacturer is different from the one stated in the papers.
4. Environmental conditions
The environment in which the product is used or stored has naturally a considerable effect on the reliability. For example, high temperature, temperature changes, high humidity or corrosive environments tend to cause failures. These are important factors to consider, when reliability testing method is chosen or when reason for failure is investigated. When environmental factors are considered, often use conditions are thought. However, the determination of use conditions can be tricky especially, if the product is used internationally and can see any conditions from tropical heat and humidity to cold seashore or highly polluted big cities.
It is important to notice that the use environment is not the only environment which should be considered. Products are also exposed to environments which are present during their transportation and storage. This might mean hot temperatures due to being stuck in a shipping container in direct sunlight, resulting in much higher thermal exposure that was designed for the use environment. Or in the worst case the products are stored outside in rain and polluted areas for prolonged time while waiting for installation.
Improper storage of packages in winter conditions.
Overstress often causes rapid and surprising failures. In overstress a critical limit of material, structure or component is exceeded which leads to failure. Overstress can be any kind of stress, such as too high temperature, too quick temperature change, too humid or dry environment, exposure to chemicals which can break for example polymer chains in plastics, exposure to highly corrosive environments etc.
Overstress may occur in manufacturing if for example chosen processing temperature exceed the thermal limits of some material present in the product. Another cause for overstress is misuse. The product may be stored in wrong conditions or the customer may be handling the product incorrectly, for example using it in conditions exceeding the specifications of the product.
Wear means that a failure occurs because products, their materials, and components age and therefore lose their original properties. All products wear in time. The process is slow, and therefore failures due to wear develop slowly over time. Wear causes lots of different failure mechanisms depending on the materials, structures, and conditions. For example, solder joints tend to crack due to fatigue or plastic casings become fragile and crack due to prolonged exposure to increased temperatures or UV. When a failure due to wear occurs in the end of useful lifetime of a product, it is not reliability issue but a part of the normal lifecycle of a product.
Corrosion is one the most common reason for failures and poor reliability.
One particular figure is very commonly used when reliability is discussed: a bathtub curve. The bathtub curve shows the failure rate of a product as a function of time. Thus, it is a description of various failure rates for products over their lifetime. An example of a bathtub curve is shown below.
As can be seen, a bathtub curve consists of three different failure rates: decreasing failure rate, constant failure rate and increasing failure rate, which together form the curve shaped as a bathtub.
The first part of the curve describes early failures. At this stage a high number of failures is seen due to errors in design or manufacturing. The failure rate, however, is decreasing, because the products which have defects and are therefore failing fast are removed from the population.
The middle part of the curve describes the useful lifetime of a product. During this period the failure rate is constant. The failures seen are random failures, which can be caused for example due to random external stresses or mishandling of a product.
The last part of the curve describes the wearout failures of a product. At this stage the failure rate is increasing, as the aging of components and materials is accelerating the occurrence of failures. At this stage the failures can be caused for example due to corrosion, oxidation, or fatigue.
The first part of the curve is often called the period of infant mortality. This is because in addition to describing the failures in hardware population, the bathtub curve can also effectively describe the mortality of human population. Another name for the first part of the curve is burn-in, which brings us to the question of how reliability testing and the different parts of the bathtub curve are related?
Failures occurring in the first part of the curve can be at least partially found with burn-in or screening tests, in which a product is for example switched on to ensure its functionality or shortly tested with mild stress levels. The middle part of the curve should typically be rather long to ensure adequate use life. Various reliability test methods can be used to improve the performance of products and their components which prolongs the period of useful lifetime and postpones the occurrence of the wearout phase. Acceleration and harsher stress levels than those present at normal use environment are required in testing. Otherwise the failures will not occur in a reasonable test time.
Burn-in tests do not tell much about the reliability of the population of a product. They merely find the number of weak products in a population. They do not tell anything about the failures seen during the last part of the bathtub curve. When the reliability of products is tested, we recommend continuing testing until the majority of the tested samples have failed. This is especially important if the population has lots of early failures. Studying only the early failures may give misleading information of the actual wearout failures. Similarly, using only warranty data may cause similar problems.
As a final note to the bathtub curve it is important to notice, that the curve for an actual product may look very different from the one shown in the picture. For example, there may be a very few early failures or the failures due to wear may start very early and happen slowly causing the curve to increase slowly already during the useful lifetime. There are endless possibilities, but the simple bathtub curve is still a good tool to describe typical failure rate behaviour.
Flexible electronics means applications build with flexible substrates and materials, enabling bending and stretching of the applications. Therefore products which would not be possible with traditional electronics can be accomplished. Application such as wearable and medical electronics and flexible displays have unique structures and features compared to traditional electronics, which makes them highly attractive for many products.
Since flexible and stretchable structures are still new, there is a limited number of test standards and methods available. However, there are a few important and useful factors to consider when testing these structures. For example we always recommend measuring the electrical behaviour of the tested structure in real-time during testing. This is the only way to actually learn how the tested structures are functioning in certain environments and under certain stresses.
Quite often samples are not functional during testing. However, when they are returned to ambient conditions, their behaviour reverts to normal. In these cases, if intermittent measurements done only before and after testing, they give misleading results, as the failures in the test conditions are not revealed. This applies especially to bendable and stretchable structures.
Below you can see the resistance behaviour under bending for a printed structure we have worked with in Smart2Go Eu-project. The sample consists of printed silver wires on PET film and dummy resistor components attached with silver paste (or isotropically conductive adhesive). The figure shows that during each cycle of the test, the sample has either very low resistance or extremely high resistance – basically an open circuit – depending on the severity of the bending. In this case each bending cycle was recorded with ten data points. When the structure is bent, it loses electrical contact, but when the sample is straightened, the resistance value is normal, in this case approximately 10 Ohms. This behaviour started at cycle one and lasted for the entire test. And, when the bending test was finished after 50.000 cycles and the sample was in straight position, the resistance was again 10 Ohms.
Why could this be a problem? Well, it may not be a problem for the sample, if it is not bent in its real use, or if bending is occasional and losing contact is acceptable. However, if the sample is repeatedly bent during its use, it is vital to be aware that the sample may not be functional at all positions. For example, if this would be a wearable electronics device, the electrical functionality may be lost when a person bends his or her arm. This means that the product abruptly stops functioning properly.
We provide testing services for the reliability testing of flexible and stretchable electronics. We have three testers designed especially for flexible and stretchable structures. Two of these testers are dynamic three-point-bending testers shown in the figure above. One is designed for dynamic twisting. We offer also a possibility to combine bending testing with high humidity and high temperature environments.
Fluctuation of temperature tends to cause lots of failures for electronics. This effect can be tested with two different tests: thermal cycling and thermal shock. These two tests are compared in this infographic.
Humidity testing is one of the most commonly used accelerated reliability test methods for electronics. Have you ever considered why? Here are 10 practical points about humidity testing and why it is so important and popular in electronics.
1. Humidity testing is easily available
Humidity testing is quite straight-forward type of testing. This means humidity tests are relatively easy to perform, with modest costs and the repeatability of results between various tests is good. Humidity testing is also performed in standard humidity test chambers which are readily available both in testing labs and R&D labs.
2. Very effective method for plastic parts
Even though humidity testing is a useful test for electronics, it is not the most efficient method to study all parts in electrical applications. For example, humidity testing often has a limited effect on solder joints. In general, humidity alone is not very effective test method for metals and ceramics unless corrosion processes are studied. But for plastic parts, such as printed circuit boards, adhesives, coatings, overmoulds and plastic casings, it is a very effective test. This is because humidity causes polymers for example to swell and change their dimensions. Furthermore, water may react with plastics or it can act as a plastizicer, causing the polymer to become soft and easily deformed.
3. Humidity and temperature accelerate each other
Usually humidity is combined with increased temperature in humidity testing because temperature and humidity accelerate the effect of one another. Increased temperature itself fastens reactions or cause new reactions to happen. However, when high temperature is combined with humidity, the test is commonly far more severe and efficient than the humidity or temperature alone would be.
4. Water may react with materials
Water from the environment may also react with material itself. Such reactions may be very slow at low temperatures and eventually cause unexpected failures in field conditions. Sometimes impurities are needed for the reaction to occur. Consequently, in reliability and durability testing elevated temperature or presence of impurities is often needed to accelerate the reactions. A good example of this is hydrolysis in which water attacks the polymer chains and breaks them. Hydrolysis eventually causes the material to crack. More information about hydrolysis can be found in this infographic.
5. It can be used to study the effect of impurities
In addition to humidity itself, impurities such as salt and chemicals tend to cause corrosion when combined with humidity. For example, flux residues or incorrect handling of parts with bare hands can cause severe corrosion in humidity testing or in humid environments. Salt and many impurities also conduct electricity, which increase the risk for leakage currents and electrical failures. As said, humidity alone is not effective for metals, but when impurities causing corrosion or conductive electrolytes are present, the situation changes.
6. Variety of humidity tests is available
Commonly used humidity tests combine a constant humidity level to a constant temperature level. 85°C /85% RH test is probably the most known example of this. In this test, a test temperature of 85°C and relative humidity of 85% are used. However, lots of other combinations are also widely used. Moreover, constanthumidity testing is not the only type of humidity testing available.
Testing can also be conducted as cyclic humidity testing to study the effect of varying humidity conditions on materials. Sometimes cyclic tests are very efficient for accelerated testing and reveal failure mechanisms which might otherwise be overlooked.
Usually condensation of water is avoided humidity tests. However, if condensation could happen in use environment, its effect can be studied with condensing testing. In this test the humidity and temperature levels are cycled in a manner that produce condensation on the tested surfaces. Such test may be very harsh on electronics.
7. Very harsh humidity tests may be used for extreme acceleration
The commonly used 85°C /85% RH humidity test can sometimes already be extremely harsh and cause failures very quickly. However, it is possible to accelerate humidity testing even further by using HAST (highly accelerated stress testing) testing. These tests are conducted in a pressure cooker which enables the combination of a very hight temperature – above 100°C – with a high humidity level.
With such tests extreme acceleration may be achieved. However, it is important to notice that for many materials and structures these tests are too harsh and cause failures which would never occur in the field, such as cracking due to hydrolysis.
8. Popularity means tons of data
As humidity testing has been and still is a popular test both in industry and academic world, there are lots of data available about the effects of various humidity tests on different structures and materials. Numerous papers have been published about humidity tests in electronics and many of these papers include in-depth failure analysis giving a good idea of how and why humidity effected the studied structure. Therefore, it is often quite easy to find a reference to the different humidity tests and to their effects on various components.
9. Reliability prediction is possible – although with some limitations
Performing reliability testing brings always forth a question: How do the test result relate to the actual use conditions? Basically we want to predict the actual use life as accurately as possible. Answering this question is always tricky.
Since humidity testing has been widely conducted in electronics, there are also several acceleration models which have been composed on basis of these results. Many of these models are quite simple and prediction with them can be quite easily done. Some commonly used ones include Peck’s model and Lawson’s model. However, it is important to notice that acceleration models are typically built for certain structures and materials and this limits their use and accuracy considerable especially for complicated samples!
10. Excellent for validation testing
In addition to using humidity testing as reliability test, it is an excellent validation testing. As discussed earlier, validation testing is performed in the end of product development to verify functionality. Humidity testing can be very efficient test and therefore, only relatively short testing time is needed. Consequently, humidity test works very well for confirming that application performs like it should in humid conditions.
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.