How to solve the LED heat dissipation problem?

Achieving high brightness lighting with LEDs requires using the maximum current allowed by the manufacturer, but the average life of an LED is highly dependent on the operating temperature. An increase in operating temperature of only 10°C can reduce its life by half. This situation forces designers to reduce the regulated current and sacrifice brightness to extend life. If the LED is required to operate at a higher ambient temperature, the current must be further reduced to minimize the ambient to chip temperature rise to ensure life. However, due to the upper temperature limit, this will reduce the lighting brightness in the medium and low ambient temperature range. Essentially, we achieve high temperature operation by reducing the brightness. Figure 1 shows an LED driver circuit using a thermistor-controlled operational amplifier (op amp) that reduces the drive current when the LED board temperature rises.

Figure 1: An op amp reduces LED current when it senses rising temperature.

The LED array current is regulated by sensing the voltage across the current sense resistor R7 and used as feedback control for a controller such as the TPS40211. The op amp circuit (including R9) injects a current into the feedback node (FB) to reduce the regulated current, or injects a current into it to increase the regulated current. The FB node voltage remains constant at 0.26V. Increasing the voltage at the op amp output (TP1) must be compensated by reducing the voltage across R7, thereby reducing the LED current. When the op amp output is exactly 0.26V, the injected current is zero and LED regulation is unaffected.

Thermistor RT1 is a negative temperature coefficient (NTC) device. Its nominal resistance is 10K ohms at 25°C, but increases to over 300K ohms at –40°C and decreases to under 1K ohm at 100°C in a nonlinear fashion. Resistors R8 and R10 adjust the 5V bias voltage down to close to the FB voltage, while the value of R9 controls how quickly the current decreases with high temperature. Using a well-regulated bias voltage is important because the accuracy of the circuit is affected by the bias tolerance. Resistor R9 must be placed as close as possible to the current-mode boost controller to minimize noise sensitivity. Using thermal epoxy, place thermistor RT1 as close as possible to the center LED connection on the PWB.

Figure 2 shows the data obtained at various temperatures. Only the LED and thermistor were operated in this temperature range. The temperature sensed by the thermistor is plotted against the ambient temperature. The calculated LED die temperature is also plotted, which is equal to the board temperature plus the power per LED times the junction to chassis thermal impedance (8°C/W). We can see that at high ambient temperatures, the op amp circuit reduces the LED current while the LED die temperature approaches the LED board temperature. In this case, the LED board temperature approaches the ambient temperature because the LED current is almost zero. This results in a stable LED die temperature with no change. The RT1 nonlinearity is the cause of the sharp change in LED current at the highest temperature. The temperature “control voltage” at TP1 is also plotted and matches the expected value very well.

Figure 2 As the ambient temperature rises, the LED current decreases, resulting in a lower chip temperature rise rate.

Summarize

In high temperature environments, driving high-power LEDs can degrade the LED brightness and shorten its lifespan, which is where thermal feedback circuits are useful. They can reduce the LED current, which in turn reduces the LED power consumption and ultimately reduces the LED temperature rise. Since LED brightness decreases with temperature rise, this approach may not be practical in some applications that require constant brightness. However, this circuit can extend the effective lifespan of LEDs in extreme environments.