Optimization of LED based lighting applications
LED lighting is becoming the technology of choice, especially where severe environmental conditions pertain. LEDs are a robust solution and never completely fail, however, their light intensity does fade over time. It is beneficial to implement automatic dimming where the light output of the LED is measured and the power supply regulated to maintain constant light output over the lifetime of the lamp. This enables a much longer life span and constant brightness across manufacturing variations. In this article, circuit examples are shown that work with various sensor technologies for use in street, office and factory lighting.
A new approach for ecodesign at system level
In ecodesign terms, it is always desirable to improve the lifespan of a given application because it directly influences the ecological impact. This is not as easy as it sounds because innovation never stops. New standards and application possibilities need to be considered, and new features need to be implemented. It is important to not just implement the minimum standard, but to also build “better” products that pre-empt many of the new features and help to achieve a certain flexibility. An excellent example is the “Mars Rover,” where a fair amount of support helped to keep it going, and it exceeded the originally planned mission time by a significant amount.
It is important that not only flexibility is implemented but that the lifespan of the LED has the potential to live that long. Here, three categories need to be considered: fast changing impacts (e.g., turn-on currents), slow changes (e.g., drift) and external influences.
Turn-on currents are a significant stress for all applications. In power supplies, there are large bus capacitors that can cause this stress. Additionally, undefined states at power-on can cause stress on the application. In order to avoid these situations, the behavior needs to be carefully analyzed throughout the design cycle with a FMEA (Failure Modes and Effects Analysis). Then, with the finished application, it is important to verify the currents and temperatures during turn-on and during turn-off. It is possible that they might go unrecognized during design and can cause field failures later.
This picture shows the temperature change of the input bridge rectifier of a power supply caused by the turn-on current. While the other components are cool, the rectifier is heating up a lot, which can potentially impact its lifespan.
Many of the applications developed twenty or thirty years ago had circuit techniques that could avoid failure due to drifts. Since many of the components used back then had larger tolerance levels and stronger drifts, the circuits were developed to be less dependent on the component characteristics. Many of these techniques are out of ecodesign consideration today in order to increase the life span of an application. This increases the complexity of the circuit and leads to a higher component count. The ecological impact of the electronic circuits is generally lower than the impact caused by the operational energy consumption, and the increased lifespan compensates for the additional complexity.
Power electronics are a significant part of all applications and products, especially under ecodesign considerations, because the subsystem affects its ecological impact through characteristics such as efficiency. For example, lower efficiency means an application consumes more energy during production and application through larger heatsinks; and more energy is needed for transportation and recycling. Therefore, it is recommended that designers develop applications for the greatest efficiency and a wide input voltage range to achieve maximum flexibility with low component usage, as this cannot be changed later (e.g. through a software change).
In lighting applications, the trend is to use LED-based solutions. In addition to greater efficiency, LEDs intrinsically have a very long lifespan, and the electronics driving the LEDs should be designed to make maximum usage of that lifespan. A block schematic of such a lighting solution is shown below:
Street lighting is a special implementation of these lighting solutions. No other lighting application combines so many requirements that are not always converging. Changing environmental factors such as temperature, humidity and vibrations combined with the need for high efficiency, reliability, and low energy consumption all need to be factored into the lighting solution, while being produced at high volumes and meeting corresponding pricing pressures. The temperature can change from -40°C to 85°C with the power dissipation of the lamp, which leads to a greater increase in temperature. It is important to note that every (daily) turn-on cycle leads to a temperature cycle. This means the lamp sees almost 2000 temperature cycles at system level in five years, which puts a big stress on the system. And since the cost for street lighting can be a significant portion of a town’s budget, high efficiency is of particular interest to help reduce the cost of operating the lamps. The goal is to use as little electrical energy as possible to achieve the required luminosity levels on the streets or walkways. This is apart from maintenance cost which, for lower lifespan lamps, can be significant as well.
A rough estimate on how to calculate the number of lamps in a given demographic is to use a factor of one lamp per six residents. For example, using this calculation, there are approximately 13.3 million streetlamps in Germany – and the potential for savings is quite significant. It is not surprising that many cities and communities are switching to these new LED-based lighting solutions, where either the complete lamp is replaced or the lamp head is being changed.
Preemptive maintenance on a streetlamp
LED control circuit. For full resolution, click here.
Given all these requirements, the design of such a power supply is complex but not impossible. The above circuit shows an example of the input filter and bridge rectifier on the left side, followed by a slightly unusual circuit, where a BCM-PFC-controller is used in a flyback configuration. Normally, the switching frequency would change during the mains cycle, to keep the output voltage constant. Here, an oscillator (circuit with Q102) is used to maintain a constant switching frequency and hence a constant duty cycle throughout the mains cycle. The duty cycle can be changed, albeit at a much lower speed, in order to regulate the output current. This is not a problem since LEDs are not a very dynamic load. With this change, the input current will correspond to the input voltage, and the converter will constantly operate in DCM mode, yielding a very good power factor.
This picture shows the current throughout a mains cycle on the primary side (blue) and secondary side (brown). It is obvious that the peaks of the current follow a sinusoidal envelope, allowing it to achieve a good power factor and low conducted emissions, even with a small input capacitor.
The circuit can achieve a good power factor without additional PFC circuitry. A disadvantage is a high ripple current on the secondary side at double the line frequency; but for LED lighting solutions, this is not a disadvantage since the eye is not sensitive to a 100Hz flicker.
On the secondary side, a circuit is implemented that converts the output current – measured with a shunt resistor – into a feedback signal for the primary side regulator in order to implement a current output for driving the LEDs. Here, the average output current is regulated because the feedback loop must be slow in order to achieve a good power factor on the input side. An additional advantage is the use of a foil cap as input capacitor, where the small size reduces the space requirements and its lifespan is increased significantly.
The lifespan of an LED lamp is furthermore defined by the drift of the brightness of the LED. For example, OSRAM is giving its LED lamp ”Golden Dragon Plus” an average lifespan of just over 45,000 hours (continuous operation at 0.7A current and a temperature of 85°C at the solder junction). Four phases can be distinguished, with the first relatively fast drift that can go up or down, followed by an aging phase of the reflector, followed by a long phase where the LED only drifts a little bit. The final phase is the real end of the life drift of the lamp, where brightness is continuously decreasing.
The LED does not fail catastrophically, but the brightness is down to 70 percent at the rated lifespan. The lamp will become continuously darker over time. In some applications, this is not an issue, but in other applications – where constant luminosity is required through standards and regulations – this cannot be tolerated.
Compared to regular streetlamps that use high pressure sodium lamps, the LED lamp is much more robust and has a longer lifespan. However, further improvement is possible. In order to implement the measures to help increase the lifespan and achieve a reduced ecodesign impact, a brightness sensor can be used to regulate the brightness and achieve a constant luminosity over its lifespan.
One example of a secondary side constant current control that can be used for dimming high power LEDs is shown below:
In order to change the brightness of an LED without changing its color, pulse-width modulated schemes are used. To implement this, a square wave signal is added to the input of the circuit. When the voltage is higher than 1V, the regulator is disabled and no current flows through the LED string. When the voltage is low, the input node of the IC can assume the voltage implied by the shunt resistor, effectively turning on current control and operating the LED string at this constant current. This will yield pulses of constant current, followed by pauses where no current flows. This will dim the brightness of the LEDs without changing the colors. In combination with the above PFC flyback circuit, it is important to use a modulation frequency for this PWM signal that does not cause visible beat frequencies with the line frequency.
There are many ways to demonstrate how to implement brightness sensors. For example, photo transistors or diodes can be used but they will need some analog circuitry to make their signal suitable for regulation. There is also a standardized analog interface that many commercial brightness sensors use, with a control voltage range of 0…10V. In noisy environments, the signal can also be transferred with current loops, e.g. operating at 0…20mA. Also, digital interfaces are possible where sensors or sensor arrays, combined with a central controller, can perform the brightness regulation, even enabling a redundant operation, where the failure of some lamps can be compensated. This is particularly interesting in areas where the replacement of a lamp is not easily doable, e.g. in semiconductor fabs.
With such a central controller and a lamp, where regulation and turn on events do not negatively influence the lifetime, further functions can be implemented, such as regulation based on the presence of users, ambient light regulation, or even a turn-on or –off through remote control, e.g. based on a phone call or text message.
Conclusion
Initially, the low light output and comparatively low reliability of LED lighting solutions was hindering its acceptance. However, these problems have since been overcome and LED based lighting solutions are finding more and more uses. The true ecodesign impact advantage is only now becoming visible, as control circuits and applications are being developed to make its long lifespan really usable.
About the author: Alfred Hesener from Fairchild Semiconductor is responsible for applications and technical marketing in Europe.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
