© 2006 Zetex Semiconductors plc Feature article
Reference: ZE0500
Date: 5 October 2006
Thermal considerations for high brightness LEDs Colin Davies, Applications Manager, Zetex Semiconductors
Much attention is being focused on the dimming strategies used for controlling high brightness LEDs in a rapidly increasing number of applications. The nature of these devices demands however that even more attention be paid to thermal control.
While LED manufacturers are pushing back the technological boundaries by dramatically increasing the Lumens per Watt, there is still more electrical power converted to conducted heat than to light output. An overall strategy of thermal management is required to ensure that the power into the LED is controlled as a function of temperature.
The curve in figure 1 shows typical derating characteristics for a 1W LED. As expected, this clearly shows that the LED is driven with constant current up to a point at which the current then needs to be reduced linearly to 0 at 150°C. The point at which this derating occurs and the slope of the decrease is dependent on the mechanical/thermal heat sinking arrangement.
Fig 1
2
The control electronics must therefore be able to cope with a trip point setting and a gain setting. It’s also important to remember that the LED needs to be able to cope in fact with three potential contributors to the heating effect: self heating, ambient temperature and the LED electronic control. If the electronic control is remote from the LED illumination, then this will not be an issue – EMC however may be a concern.
Linear control of the LED
Take a trip back to the textbook and the first and most obvious way of controlling an LED is with a resistor. Although a low cost method, it inevitably results in power loss which negates or diminishes the LED’s key efficiency attribute. Using a variable resistor as the dimming element is also impractical for high brightness LEDs since the power in the resistor will be too large and a dedicated wire wound device would be required. As an illustration, for a single 1W LED taking 350mA from a 12V supply, around 2.5W would be wasted in the dimming resistor at full brightness. And if the resistor is in close proximity to the LED then the extra heating it produces will only add to the problem.
The pass element could of course be a transistor, which would mean the power dissipation resides in the transistor and not the variable resistor. This approach provides more flexibility by enabling logarithmic response, negative (NTC) or positive (PTC) temperature coefficient thermistors to be used for thermal control and brightness definition. A bit of imagination then easily leads to the use of optical feedback for automatic brightness control.
The transistor could be of any type: MOSFET, NPN bipolar or PNP bipolar. Amazingly some engineers of a more digital persuasion still believe that MOSFETs are better for this application because of their low on resistance! It simply has to be understood that linear power dissipation is the same regardless of silicon choice. It’s still wasted power in the form of heat which needs to be managed.
© 2006 Zetex Semiconductors plc
百度搜索“77cn”或“免费范文网”即可找到本站免费阅读全部范文。收藏本站方便下次阅读,免费范文网,提供经典小说教育文库thermal_considerations_for_high_brightness_LEDs_5在线全文阅读。
相关推荐: