Importance of lifetime and reliability in LED driver design

With the increasingly serious energy and environmental problems, traditional incandescent lamps are being replaced on a large scale around the world. Although a large part of them are compact fluorescent lamps, the concern about the mercury content of energy-saving lamps (which may cause environmental pollution) and the demand for greater energy-saving potential have led more and more users to turn to LED lamps. The latest LED lamps consume less than 80% of the energy of incandescent lamps and do not contain toxic substances. According to market research company iSuppli, global sales related to LEDs will continue to grow strongly, and despite the slow recovery of the global economy, the market is estimated to reach approximately US$14.6 billion in 2013.

As energy-saving lamps rich in electronics are widely used, consumers are beginning to see frequent failures of these products. LED lighting has a much longer potential life and offers better reliability.

The question is whether the electronic drivers used in LED lamps have achieved the same lifespan and high reliability? Poorly designed products can ruin the LED lighting industry. Well-designed LED lighting systems can last up to 50,0004 hours. However, unless the electronic devices can also become correspondingly long-lived and highly reliable, the advantages of LED lighting will not be realized.

In design and production, it is important to understand that the service life of the product and the reliability of the product are two very different concepts. Although they are not unrelated, they are often confused. Lifespan refers to the length of time after which users can expect a single product to work properly until the product is no longer suitable for use. Reliability is used to reflect the failure rate of batch products, which can be expressed as MTBF (mean time between failures) or the inverse of the failure rate. A service life of 50,000 days means that the product can serve for 50,000 hours. And a MTBF of 50,000 hours means that for 1,000 products, every 50 hours, from a probability perspective, theoretically people will see a random failure. These two concepts are very important for the successful implementation of LED lighting.

2 Lifespan

Estimating the life of any product is primarily about determining the wear mechanisms of all electronic components and then finding the shortest component life. For most power supplies, including LED drivers, the shortest life component will be the electrolytic capacitor. The electrolyte in the capacitor will evaporate to varying degrees with operating temperature and operating time, and the capacitor ripple current will also affect the life. Although electrolytic capacitors may vary from manufacturer to manufacturer or component model to model, the life of a typical electrolytic capacitor can generally be expressed using the following formula:

Where:

Lx——result of life calculation;

k is a coefficient determined by the RMS ripple current and operating voltage of the capacitor, which can be a value or a function;

L0——Life data under specified standard conditions provided by the capacitor manufacturer;

Ts - the surface temperature of the capacitor under specified standard conditions provided by the capacitor manufacturer;

Ta——Capacitor surface temperature under target working conditions.

With this formula, optimizing lifetime design becomes fairly simple.

The first thing is to choose a high-quality, long-life capacitor. Second, engineers should strive to reduce the RMS ripple current flowing through the capacitor, as well as the operating voltage. So the selection of capacitors needs to reduce the ripple current and voltage to obtain sufficient design margin. However, excessive selection of high-specification capacitors will lead to the use of larger and more expensive products, resulting in increased product costs. However, insufficient design margin may greatly damage the service life of the product. The industry believes that the most effective way is to reduce the surface temperature of the capacitor. The surface temperature of the capacitor is determined by the surrounding operating environment temperature, the heat dissipation capacity of the driver, and the heat generated.

For a given existing design and application, the main determinant of temperature will be the efficiency and heat dissipation of the driver. In other words, high efficiency and low thermal resistance design can significantly increase its service life. Efficiency has a greater impact on temperature than many people think. For example, going from 95% to 85% efficiency does not mean that the loss amount will only be 10% different, but it corresponds to a 3.3-fold increase in losses, which are converted into heat in the driver. Inventronics has invested a lot of R&D efforts to improve the efficiency of LED drivers. For the popular EUC-150S (150W constant current output) series, the efficiency at full load of 220V AC reaches 92%, and the loss is only 13W. As the output power increases, only a 1% efficiency difference can see completely different power losses. Figure 1 shows the relationship between efficiency and losses.

Figure 1 150W LED driver loss efficiency curve

Since products of different designs can have significantly different efficiencies, the temperature inside the driver housing can be very different. As shown in formula (1), a 10°C temperature difference can double the lifespan. Even if the assumed thermal design is the same, which means that the thermal resistance from the capacitor to the air is the same, drivers of different efficiencies will inevitably lead to different capacitor temperatures and therefore very different lifespans. Still using a 150W product as an example, Figure 2 shows the relationship between efficiency and lifespan.

Figure 2 150W LED driver life efficiency curve

However, even if the driver has high efficiency, limited power loss will lead to high temperature of internal electronic components if there is no good heat conduction or convection design. The use of good thermal conductive glue materials and a solid aluminum alloy shell can greatly reduce the thermal resistance from electronic components to the environment. In this way, the driver can achieve a life of 87,000 days at an ambient temperature of 45°C. This is better than most LED drivers on the market today, which will greatly improve and promote the application and development of LED lighting projects.

3 Reliability

Reliability is a concept related to the failure rate of a product operating under its rated conditions during its rated service life. The method commonly used to express reliability is MTBF, which is the mean time between failures of a product. Although reliability and life are often calculated in hours, they are still quite different concepts. Formula (2) shows how to calculate the mean time between failures, which is the total operating time divided by the number of product failures.

For example, a sample population of 1,000 products generates a total of 24,000 hours of time per day. If this population has 4 failures in 1 month of operation, the MTBF of this product is (1,000 units × 24 hours/day × 30 days)/4, or 180,000 hours. To provide another example, if the MTBF of a product is 300,000 hours, it is likely that one of these products will fail every 300 hours on average when 1,000 of them are working simultaneously. If the number of these products is 10,000, then it can be expected that one failure will occur every 30 hours on average. It must be clear that a product with an MTBF of 300,000 hours does not mean that any particular product will be expected to have a lifespan of 300,000 hours. There is another way to understand reliability, and that is to look at its failure rate. Formula (3) shows that the failure rate is simply the inverse of the MTBF.

When determining the life of a product, one simply finds its shortest component and calculates its life. However, when determining the reliability of a product, one needs to understand the failure rate of each component that could cause the product to fail and look at the combined failure rate.

Predecessors have spent a lot of time on the evaluation and research of the reliability of electronic equipment. The most common of these methods is MIL-HDBK-217, which is widely used in military equipment. It is considered to be the standard reliability prediction method and is also the reliability prediction method used by Infit Power. Another common method is the Telecordia reliability prediction model. Usually, the evaluation results of military equipment are more conservative than the second Telecordia commercial method, that is, the life value is reduced. These two methods are called "calculated" reliability, rather than being obtained by running a large number of individual experiments. In fact, it is difficult to obtain the MTBF value through the latter. The reason why MTBF is used to measure the reliability of a product is actually because these standards make it possible to make effective comparisons under the same conditions using the same method.

In fact, the challenge facing drive manufacturers is mainly how to produce the most reliable products under certain scale and cost constraints.

There are several key factors in reliability design. The first is the choice of the main power stage topology of the design. The reliability of a semiconductor component is usually determined by the operating junction temperature. Topologies such as zero-current soft switching flyback and LLC half-bridge can be used to minimize the switching losses of the power semiconductor, thereby improving the efficiency of the semiconductor and the entire drive and reducing the temperature. The second is to consider the selection of high-quality components and ensure appropriate stress margins for the components. For example, design a 20% operating voltage margin for high-voltage electrolytic capacitors and a 10% spike voltage margin for semiconductor components to ensure a reliable design. Third, the protection circuit can enable the product to survive extreme conditions, including various abnormal over-current, short circuit, over/under voltage or overheating. In addition, surge suppression circuits are needed to prevent lightning damage. Fourth, as mentioned earlier, we are back to the issue of efficiency and thermal design. Since temperature has a direct and significant impact on the reliability of semiconductors such as MOSFETs, integrated circuits and optocouplers, we have to put the efficiency of the driver again in an important position. Figure 3 shows how efficiency affects the mean time between failures of a 150W product.

Figure 3 150W LED driver MTBF efficiency curve

The last important issue about reliability is to eliminate the high failure rate period at the beginning of product completion. The concept of product reliability is only valid during the service life of the product. In fact, the life of the product itself also includes the period from production completion to delivery. But for users, the service life starts after delivery. Figure 4 shows the well-known "bathtub" curve. In this figure, the y-axis is the failure rate of the product and the x-axis is time. Most foreign manufacturers call the high failure rate period just after production completion infant mortality, and then enter the service life of the product, which is the flat bottom of the curve. Finally, as the product reaches its rated life, the failure rate begins to rise. The challenge for manufacturers is to ensure that these products have passed the high failure period before leaving the factory. To achieve this, Infit uses rigorous double aging for all products.

Each product is subjected to a 1 to 2 hour aging test before glue filling.

Then, after the final potting and assembly, all products are run at high load and temperature for a period ranging from 4 to 12 hours, which is determined by the failure rate of the product during the aging process.

Figure 4 “Bathtub” curve

4 Conclusion

Understanding lifespan and reliability is critical to LED driver design. Lifespan and reliability are even more urgent for projects such as LED lighting that require long-term uninterrupted work to achieve returns. The long life of the power supply makes the lamp cheaper to use, and high reliability makes the maintenance cost of LED lamps lower, which forms a virtuous circle of return on investment. There are many factors that must be considered to achieve this, but in this article, we can see that efficiency as a key indicator further explains the importance of LED driver power supply to some extent. How LED driver power supply can meet the requirements of today's lighting projects still requires a lot of time for in-depth research and analysis to ensure that the goals of long-term use and energy conservation can be achieved.