Vbatt > Vf OR Vbatt < Vf OR Vbatt = Vf ?

[MrNaz]

I always thought that you wanted to make Vbatt and Vf as close as possible to reduce driver inefficiency. Well I just finished make a 4LED light, hoping to power 2S2P LEDs with a 3S2P LiIon battery array. Not much difference between Vbatt and Vf. Great. Until I discivered that a voltage regulated driver would either just direct drive for the first half or blow, and a buck driver would only work for the first half.

What gives? Do you have to design such that there is a difference to provide breathing room for the driver? How about those regulated 1 LED 1 LiIon lights?

There's always the option of using boost/buck drivers, but they all seem so expensive. Are there any cheap (under $10) boost/buck drivers? Ideally, I'd like to design around AMC7135 boards, as they are under $2/board and those chips are fairly efficient, and wiring them up like this.

However, is there a solution for 2 LED 2 cell arrays? As in, a cheap solution that doesn't cost over $50 like the hipCC drivers? Surely a basic boost/buck IC shouldn't cost more than an entry level Opteron CPU?

 

[rmteo]

The AMC7135 is a linear shunt (step down) regulator so Vbatt > Vf. Also, anything not lighting the LED is going to be turned into heat. Its max operating is 6V so you are limited to 1SxP LiIon configurations. The larger difference between Vbatt and Vf, the lower the efficiency. Typically, efficiency for shunt regulators = (Vf/Vbatt*100)%.

Switching regulators such as buck (step down), boost (step up), buck/boost (step up/down, inverting) and SEPIC (step up/down, non-inverting) can be more efficient.

Boost and buck/boost are typically 80% efficient (SEPIC can be slightly better at around 85%) whilst buck is the most efficient at 90-95%.

For your battery and LED configuration, go with a buck regulator.