High-power LED problem

The nice thing about the logic-level FET is that the Rds(on) is specified at logic levels in the datasheet table. To use a standard FET at logic levels means more calculation.

In general choosing a FET with lower breakdown voltage (BVDSS) will give you lower Rds(on). Unlike Rds(on) BVDSS increases with temperature so do add margin for temperature increase.

The LM317 in the Instructables circuit regulates the current through the LED using closed-loop feedback. IIRC the LM317 maintains a voltage of 1.2V between the Out and Adj pin. The current through the resistor is 1.2V / R. Since no current goes into the adjust pin all of the current has to go through the LED. The LM317 is acting like a variable resistance.

How are you proposing to shutoff the LM317? I do not see a practical way to do it

(* jcl *)

How are you proposing to shutoff the LM317? I do not see a practical way to do it

Couldn't you put a transister after the led, so the cathode of the led is connected to the collecter, emitter going to ground and the gate controlled by an arduino output?

You will gain current regulation but you now have three parts in series with the high current -- LM317, current setting resistor and switching transistor. The only benefit is regulation of the current.

You would be better off with a current sink which would use an op-amp, FET and current-setting resistor. Since the FET is used as a variable resistance the Rds(on) isn't critical.

For a schematic example scroll down to the "load cell" schematic at http://www.luciani.org/geda/util/matrix/index.html

That basic cell design is the same as the one I use on my LED driver. Pictures are at http://wiblocks.luciani.org/not-quite-ready/ZB1-ISINK-index.html The LEDs in the pictures are running at 350mA.

(* jcl *)

Thanks for the input jluciani.

Yeah it does seem a little redundant... Ive spent hours looking for a simple circuit that allows logic level control of a constant current source/sink for high power led's...

In your load circuit, Is the op amp regulating the current output of the FET? Or is the Resistor in series with the FET acting like a current limiting resistor? I dont understand op-amps at all.

The op-amp is controlling the resistance of the FET by varying the gate voltage. The amount of current is determined by voltage on the +IN of the op-amp and the resistance of the sense resistor.

With negative feedback the op-amp will vary the output to maintain 0V between +IN and -IN. The voltage between +IN and GND will equal the voltage between -IN and GND. The voltage between -IN and GND is the voltage across the current setting resistor (Vsense). The current through Vsense, Isense is equal to Vsense / Rsense. Since no current flow into the gate of the FET the current out of the source is the same as the current into the drain. The current out of the source is Isense. The cathode of the LED connects to the drain. The anode to a voltage source.

When I finish the datasheet for the current sink kit I will have a detailed description of the sink and the thermal calculations.

(* jcl *)

Wow thank you all for the help, advice, and analysis ;D! I don't have the money to spend on a switching power supply, or a constant current supply, so I am still sticking with 'quick and dirty'. I have removed the 3.3V regulators, and replaced them with LM7805 5V ones. I now have a set of 3 adjoined LEDs in parallel (so same voltage drop, right? Not stacking?). Each of the 7805's drives 3 LEDs (so 2 7805's driving 3LEDs each). The two 3-LED groups each have a 10-ohm, 1W resistor tying them to +5V. Now what transistor should I use to switch each LED on (brought to GND)? I am not bringing the 2 groups to ground, but each LED individually. The 2N2222's can't handle this setup as you all said before (I tested it and they do some pretty weird stuff). Let's take note that I CAN use the 2N2222's to turn on higher-powered transistors that I get, if necessary. Radioshack is the limiting factor here because I need to get this project done really soon. It looks like I have a choice between the: TIP42 (PNP), IRF510 (MOSFET), TIP31 (NPN), TIP3055 (NPN). I think one of those has a base-to-emitter voltage of > 5V so obviously I can't use it. Thanks a lot to you all; I really appreciate it.

wait...scratch that part about driving a transistor with a transistor (2N2222). Forgot about the voltage drops

Ive spent hours looking for a simple circuit that allows logic level control of a constant current source/sink for high power led's

If the constant current supply voltage is the same as the arduino then a simple transistor with the collector to 5V and the emitter to the constant current circuit will do. Otherwise just treat the constant current supply like you would treat a LED in a matrix and switch it by either sinking current from it or sourcing current to it.