[trobbins]

I set up a 317T for 100mA on a small protoboard, using a 5.2Vdc regulated smps wallwart (USB power supply). The 317T tab is thermally conductive with the PCB vias underneath the tab and to one side, but no additional heatsinking to ambient has been applied as yet. The 317T package increases in temp by about 7-8degC using an IR gun, with the 100mA load being the current measurement paths of 2 meters (ie. a low voltage <<1V). The constant current changes with 317T temp by about 0.24% after 5-10 mins (and for my setup it reaches 100.00mA).

For many, that kind of current variation is a non issue, so the simplicity of using a 317 is very apparent.

At 100mA, the 317 temperature change can be alleviated in a few ways - including using a heatsink; an input dropping resistor; and a thermistor as Trev has tried.

Restricting the range of resistances to be measured to say below 10 ohm, means the 317 only has to regulate the output up to about 1V, so for a 1.7V min differential across the 317 and a 1.25V sense, the input voltage to the 317 could be as low as about 1+1.25+1.7 = 4V, and so a dropping resistor could reduce 317 power dissipation by about 2/3rds in my situation by using the 5.2Vdc supply with a 1 ohm dropping resistor.

I also confirmed that adding a 1N4004 flyback diode across the CCS output terminals does not affect the current level, and could be worthwhile for protection when measuring transformer windings.