I'm contemplating the ADP2118 for a high-power white LED driver. The usual issue with buck regulators in this application is the [relatively] high reference voltage into the error amplifier. Getting 0.6 volts from a current-sense resistor means lots of wasted power in I^2*R loss.
From the datasheet, it looks like I could drive the TRK input to something like 50 mV, making that the effective reference voltage for the error amplifier. Then I could use a small sense resistor and everything is wonderful. Are there any problems with this scenario, or should it just work? (I do realize the FB/TRK offset voltage means the fractional uncertainty will be larger.)