How to build intelligent LED lamps

Light-emitting diodes have long been used as low-cost indicator lights in a variety of electronic products. Today, they have become powerful light sources for indoor lighting, signage, display screens, and decorative lighting applications. Compared with incandescent and fluorescent lamps, LEDs are gaining popularity because they consume much less energy to produce the same brightness. Energy is one of the hottest topics of this century and will soon become one of the most important issues that designers around the world need to consider.

For lighting manufacturers, using LEDs has many advantages. However, there are also many manufacturers trying to catch up with the LED wave as early as possible, so product differentiation is very important. In addition, when energy consumption and labor costs become the main issues that need to be considered in design, large lighting equipment almost needs to be "smart". The ability for the lamp to communicate with the "parent" controller, monitor its own condition, adjust the working mode based on the monitoring results, and enter a safe state after failure is all expected of the new generation of LED lamps. This article will explore some of the "smart" options suitable for LED lamps and the steps to achieve them.

Input Low Voltage Lockout:

The input voltage to an LED driver system is typically a DC voltage. The power is provided by an offline AC-DC converter or by the bus . In addition to powering the LED driver, the power is also used to power the controller in the system (after converting to a 5V or 3.3V voltage suitable for the controller). The controller power supply is generally designed to start operating when the input voltage is slightly higher than the required output voltage. For example, a 5V regulator will start operating at an input voltage of 6-7V. However, the steady state of this power supply can be a 24V power supply for a string of 5-6 LEDs at 1A per string. Once the controller is powered up, the controller assumes that power is available and turns on the LED driver system (assuming it is configured to do so), which then attempts to enter normal operation. If the input voltage is only 10V at this time, the current required from the power supply will be much higher than the steady state current, and the system will crash due to this instantaneous current surge. This excessive current demand will also exceed the rated capacity of the cables, connectors, and other components of the input power supply, which may cause permanent damage to the system.

To avoid this, the system should have a "low voltage lockout" function. The hardware used is a resistor divider that can gradually reduce the input voltage to a range that the controller input can withstand. The input is internally connected to a comparator. The controller (firmware) should be designed to act so that the power supply section is enabled only when the input voltage exceeds the threshold for normal operation. In addition, the voltage system does not start immediately when the comparator is turned on. The firmware should poll the output of the comparator to check whether this state is consistent (because the comparator is part of the combinational logic circuit) before starting the power supply system. Figure 2 is a hardware schematic (simplified diagram) to implement this function.

Load (LED) monitoring:

The load here has a regulated constant current through the LED. Although the current regulation system is already monitoring the load, the goal here is to ensure that the appropriate load current is flowing. LEDs are prone to damage, especially when an open circuit or short circuit occurs. Causes of these types of failures include loose cables, loose connectors, PCB assembly issues, etc. A short circuit in the channel will cause the MOSFET (which acts as a switch) to be damaged. Given the power of these systems, in the event of a fault, a large current will be drawn and a lot of heat will be released. In order to protect the system and its surroundings from the adverse effects of faults, the controller should have the ability to monitor the load condition in real time.

Let's consider an open circuit situation, where there is no path for the current to flow. If the current regulation system is left to its own devices, it will keep the switch (MOSFET) open to try to get the current to flow where it wants. But this will not solve the problem. Similarly, in a short circuit situation, there is an uncontrolled surge in current. The feedback system will try to close the switch, but if the MOSFET is damaged, it will not respond to these control signals and the problem will not be solved.

An intelligent LED luminaire should be able to detect these conditions and put the system in a condition where it can safely avoid the consequences of a fault. One way to achieve this is to force a fuse to blow, thereby cutting off power to the entire system. Another way is to send a signal to the "mother" controller, or stop sending a signal to the "mother" controller, to indicate a fault condition. To do this, the system must be able to first monitor the load current or voltage value. To measure the current, a current sensing resistor is introduced in the LED line, and the voltage across it (after amplification) is input to the ADC. The digital output of the ADC is monitored by the processor and appropriate actions are taken based on the measured current value. For example, if the current through the LED is 500mA, but the ADC only measures 10mA, it can be considered a fault. The controller then sends a signal to the "mother" controller to activate the "fuse blown" circuit and force the fuse to blow.

In a circuit such as a boost circuit, where a large capacitor is present, it is important to continuously monitor the voltage across the load. Under normal operating conditions of the boost system, the large capacitor is charged during the switch-off cycle and discharged during the switch-on cycle. If the load is open circuited, the capacitor does not discharge but charging continues. If left unchecked, the voltage across the capacitor can quickly rise to very high levels, potentially damaging components such as MOSFETs. If the circuit is suddenly closed with a loose connection, the overcharged capacitor can cause excessive current to flow through the load for a short period of time, potentially permanently damaging the LED.

A resistor divider network is connected across the bulk capacitor to reduce the output voltage to a level that the microcontroller can handle. The signal is then fed into a comparator, whose output is connected to a current-regulating system that shuts it down. When the voltage exceeds a preset limit, the comparator opens a switch, shutting down the system.

Environmental Condition Monitoring:

LED lamps are placed in various environments. In an office environment, it is beneficial to save energy if the lamps can detect when people enter and leave the room and turn off or reduce the brightness after the person leaves the room. This can save electricity, thereby saving the company's electricity bills. In addition, it can effectively utilize LEDs and extend their service life.

This can be accomplished by incorporating a simple ambient light sensor (photodiode or transistor) into the system. The output of such a sensor is usually a current (depending on the amount of light received), which can be converted to a voltage signal using a resistor. The voltage signal is then fed to the controller through its pins and converted to a digital signal using an ADC. The controller determines the appropriate action (lower the brightness, turn off, or turn on) based on this value.

LEDs also generate heat, but unlike incandescent lamps, the heat from LEDs is transferred along their terminals, while the heat transfer direction of incandescent lamps is consistent with the direction of light. In addition, LED lamps are generally installed in compact spaces with poor heat dissipation conditions. If an abnormal temperature rise occurs, it may lead to various consequences, such as reduced lifespan or even permanent damage of components and LEDs, and in extreme cases, it may even cause a fire.

In general, there are two ways to monitor system temperature. The more expensive one is to use an I2C-based temperature sensor to send a digital signal related to the temperature to the controller. When the controller has a built-in I2C interface, such as the PowerPSoC LED driver controller family, this method involves little processor overhead because the temperature value is reported directly. Another lower-cost method is to use a thermistor and a regular resistor to form a voltage divider network. The voltage divider signal is input to the controller, which uses an ADC to convert it to a digital value and takes appropriate action based on the temperature. The processor needs to perform additional work to convert the digital value to the corresponding value of the temperature. The voltage of the thermistor can then be directly input to a comparator (similar to the jumper function of load voltage monitoring). The comparator can turn the LED driver system on or off based on the preset threshold.

Real-time load current control:

Previously, we discussed how to monitor load current, voltage, and environmental conditions. Smart luminaires should also have the ability to change their behavior based on the conditions they monitor. For example, a smart luminaire should be able to adjust the brightness by changing the drive current or the digital density (PWM) of the output. This can be done based on environmental conditions, load conditions, or external inputs such as buttons or communication interfaces. Actions can be based on a variety of factors, and each factor can be assigned a priority.

For example, the controller should be able to respond to a communication interface such as I2C or UART and adjust the brightness based on the data received. However, when the input voltage drops below the latching threshold, or the temperature rises above a safe value, the controller can gradually reduce the brightness (regardless of the communication interface) and turn off the lamp at the set value. There can be many such set actions.

Traditionally, LED drivers must have a drive current set by hardware (resistance value), so real-time changes are generally impossible. In a software-configurable LED drive system, such as PowerPSoC, changing the drive current of the LED only requires rewriting some DAC registers. In addition, using a software-configurable system allows the use of a single hardware platform to design products with different functional items.

Fault logging and diagnosis:

After the lamps are installed in commercial places such as large buildings, parking lots, streets, etc., repair and replacement will affect the cost and maintenance efficiency. If the controller is equipped with an interface to communicate with the "mother" controller to plan repair and replacement work, the controller can also be used to record and report fault conditions.

For example, a lamp may fail (or a fuse may blow) due to any of the following reasons:

· Load open circuit

Load short circuit

The system temperature exceeds the threshold more than the preset number of times

Overvoltage across the load

LED life ends or premature aging occurs

The input voltage fails to reach the lockout value for longer than the preset time.

These conditions all require monitoring the operating time of the lamp. This function only requires a clock that is accurate enough to record the operating time (accurate to the second). Most internally generated clocks within the controller have a certain degree of redundancy that determines the ultimate accuracy of the timing system. More accurate clocks generally require an external oscillator or clock source.

To record the cause of the fault, the system status just before the system shuts down or the fuse blows must be stored in non-volatile media. Controllers with flash memory can store these conditions in flash memory. Another way is to use a serial EEPROM device connected to the controller through an I2C interface. The system's operating conditions and peripheral conditions can be stored in the EEPROM device continuously or before the fuse actively blows/the system shuts down.

When a faulty light fixture is removed from the equipment, the non-volatile memory device can be read out using a PC or other controller to determine the conditions before the fault occurred. The readout information can be used to determine component failure rates, unknown or unpredictable environmental conditions, and generally can be used to help engineers find the root cause of the fault.

New design examples:

Traditionally, power management circuits are implemented entirely in hardware, with the only configurability provided through hardware. As a result, creating multiple variations of a product or new products requires a unique combination of hardware components. If you want to shorten the design cycle or shorten the time to market, you need to re-examine the design paradigm. For example, if the design process replaces the included configurable software modules, the design cycle can be greatly shortened. Secondly, to make a product stand out from competitors' products or break away from the rut of past products, you only need to change the software configuration and use the same hardware platform with multiple options. Software configurability also eliminates the redundancy problem that is usually faced with hardware component configuration. If the entire product line can be put into production at approximately the same time or in sequence, it can enhance the competitive advantage of differentiation.