SDG #184 Build and Test of a 2.4A Constant Current Buck Regulator based on LM3409 LED Driver

This build addresses a genuinely awkward design problem: driving a high current LED at a regulated current rather than a regulated voltage. The chosen controller is the LM3409, a P channel MOSFET buck LED driver that uses hysteretic, constant off time control and senses current on the high side. That topology choice matters in practice, because it avoids the low side sense resistor and level shifting complications of many generic buck converter chips, and it allows analogue dimming through the IADJ pin plus true PWM dimming through the enable input.

The project runs end to end from schematic capture and inductor and sense resistor selection, through a four layer board layout, stencil paste application, hand placement and reflow, to bench validation. The instructive part is the debugging phase. First power-up did not behave as the datasheet calculations predicted, and the fault finding works through the usual suspects on a hysteretic converter: sense resistor value, off time setting components, layout parasitics around the sense node, and whether the switching waveform is doing what the control loop expects. Anyone who has designed a current mode buck will recognise the process, and it is more instructive than a board that simply works first time.

Load testing with a real high power LED shows how well the regulator holds its setpoint as forward voltage and temperature drift, which is the whole point of a constant current supply. For anyone considering the LM3409 for a torch, a machine vision illuminator or a battery charging front end, this is a useful reality check on the component count, the achievable current accuracy, and the layout care the part demands.