MR16 lamps, also known as multi-mirror lamps, are halogen-type lamps designed in an integral multifaceted reflector and are renowned for their exceptional beam control. To encourage adoption of MR16-compatible LED-based lamps, it is essential that they match or exceed the performance of conventional halogen or incandescent lamps. This requires the LED driver circuit to be compatible with common low-voltage transformers and provide smooth, flicker-free operation with forward and reverse phase dimmers. The small form-factor requirements and high ambient temperature operation pose further limitations that make the driver design complex and challenging. This article summarizes various tradeoffs involved in LED driver design and suggests some possible alternatives.

MR16 system overview
The MR16 system is based on a low voltage alternating current (AC) power supply with 12 VAC as a standard operating voltage. Conversion from high voltage wall input AC (120 VAC or 230 VAC) to low voltage is achieved by using conventional magnetic low-voltage transformers (MLVT) or high frequency electronic low-voltage transformers (ELVT). The magnetic transformers can support a wide range of power levels, but tend to be more bulky and are prone to generate audible noise. In contrast, electronic transformers overcome the size and weight limitations and provide increased efficiency without generating audible noise. The focus here is to understand ELVT behavior and gain the perspective needed to design dimmable MR16 LED lamp drivers.

The ELVT circuit shown in Figure 1 is based on a resonant inverter that uses transistors and a high-frequency step-down transformer to perform voltage conversion. The inverter operation is governed by the self-driven resonance principle where the transistors switching sequence is controlled to generate a high frequency (10 kHz – 80 kHz) AC signal across the primary winding of the transformer. The low voltage output from the secondary winding is used to power the MR16 lamps. The inverter circuit is designed to work with the resistive filament load of high-power MR16 lamps and, thus, requires special conditions to operate correctly. The load must provide: 1) low-load impedance to initiate self-oscillation after the zero-crossing of each AC half-cycle; 2) minimum current required for sustaining self-resonance; and 3) fast slew-rate that matches the switching transitions. Failure to meet these three conditions causes erratic behavior of the ELVT and results in poor performance.

Dimming MR16 lamps requires specialized dimmers compatible with ELVT or MLVT. These dimmers operate based on phase control techniques and vary the power delivered to the load by modulating the conduction angle of the 120 VAC input main power supply. For smooth and flicker-free dimming operation, additional loading current that meets the dimmer circuit thresholds for phase control action is necessary. This adds to the complexity that a LED-based MR16 lamp must meet in order to provide satisfactory performance. With this background, let’s take a closer look at LED driver topologies that can be used for MR16 applications.

LED driver topology selection
Traditionally, choosing LEDs and their configurations dictated the LED driver configuration. A full-wave capacitor peak charging circuit followed by a DC/DC buck circuit is used to implement a step-down conversion required to drive commonly available low-voltage LEDs. Although easy to implement, this circuit tends to be incompatible with phase-cut dimmers and generally is used in lamps meant for non-dimming applications. For dimmable solutions, focus on techniques that eliminate the peak charging circuit. Realizing the limitations of simple buck-based LED drivers, many LED manufacturers now offer new low-current and high-forward voltage drop LED arrays specifically designed for MR16 applications. This opens up the circuit choices now available for designing LED drivers.

For compatibility with an ELVT, it is clear that the current drawn by the LED driver has to meet specific conditions. In this context, the set of requirements outlined previously are similar to those enforced on power factor correction (PFC) circuits. It is possible to leverage the know-how of PFC circuits for designing MR16 LED drivers. The boost circuit becomes the preferred choice as it provides the ability to shape the input current based on reference signal. With a fast control loop and proper inductor selection, a boost LED driver can operate from an ELVT output. Good dimming performance is also achievable by connecting a few lamps in parallel. However, a boost LED driver may constrain the lamp optical and thermal design aspects as it limits the LED configuration. Alternately, in cases where step-down conversion is required, a buck-boost configuration can be used to provide similar performance to a boost converter. Again, knowledge of buck-boost PFCs can be applied to design a LED driver circuit that emulates the behavior of a resistive load. With this technique you can shape the input current such that a LED driver meets ELVT operation requirements.

Driver designs for MR16-based LED lamps pose significant challenges. Simple techniques based on capacitor peak charging circuit and DC/DC buck converters tend to be suitable for non-dimming applications and provide poor dimming performance. It is possible to design boost and buck-boost circuits to provide flicker-free operation and smooth dimming.

For more information about LEDs, visit:

About the authors

Montu Doshi is a systems engineer with TI’s LED power group where he is responsible for LED lighting product development and customer support. Montu received his Ph.D. degree from the University of Colorado, Boulder.

Stephen Solanyk is a senior systems engineer in TI’s LED power group where his focus is applications design and customer support for AC/DC and DC/DC LED drivers. Stephen received his BSEE degree from Cornell University, Ithaca, New York.

David Zhang is a systems applications engineer for TI’s LED group where he is responsible for LED Lighting IC systems applications. David received his MSEE from the University of Texas at Dallas, Richardson, Texas and his MS Physcis from Marquette University, Milwaukee, Wisconsin.

Send your questions about this article to: