LM3409 schematic check?

That looks like a valid way of driving it and I don't see any issues. Keep in mind however that a resistor to ground for current setting is not the most accurate method. The internal current source could vary possibly up to +/-20% over the full temperature range. So if accuracy is critical in the lower settings (the full ON will be very accurate) you may be better off driving IADJ directly with an accurate voltage source.

Thank you Clinton for the input.

Regarding the varying - why should it vary +/- 20%? According to the datasheet page 7 figure 4 (http://www.ti.com/lit/ds/symlink/lm3409-q1.pdf) the I_adj current will only fluctuate 2%. In the electrical specs page 6 the baseline current can fluctuate from 3.8 to 6.4 uA. That means setting the current for each board will be different. But it shouldn't fluctuate as it heats up... right? I'm all ears if I'm missing something.

I'm going to look into using a 1.24V voltage source as well, since that will be more consistent from board to board anyway. Do you have a recommendation for a component and how to incorporate this?

Yes, I meant from part to part and lot to lot as well as over temperature. In reality they will generally be much tighter than the spec even over all of those variables. But the methods used to generate datasheet limits incorporate wide guardbands to be safe since they are guaranteed specs. It is just something I wanted to make sure you are aware of, that there could be variation. It's the same with the internal 1.243V clamp on IADJ. It's pretty safe to use that since in reality it will vary very little. The way you are doing it should work just fine. But some people require very tight accuracy (better than the 4.5% or so spec for the IADJ clamp voltage) in which case a precision reference or analog voltage supply should be used.

So, I got the circuit board built. It's a beautiful thing! (My first PCB ever)

And miraculously, it worked on the first try. There is just one small issue. I configured the circuit targeting a 2.70A LED average current. However, I'm actually seeing 2.87A average current. This number varies slightly as my power supply approaches Vin=Vout and it is possible to exceed the rated current of the LEDs, which is 3A.

The resistor WEBENCH chose for this current was 80mOhm +/- 1% 3W. However, plugging it back into the calculator after the fact, this yields 2.96A which is closer to what I'm seeing practically.

My question is - this resistor is entirely responsible for determining the current, correct? So, if I choose an 85mOhm resistor or equivalent parallel array, the rest of my circuit can stay the same? The other issue here is resistor selection is pretty slim in this range. It goes from .07 to .08 to .10. Is there a standard way of dealing with this?

I'm not sure exactly what you are driving IADJ with typically, but for full IADJ voltage you would need a 9.11mohm resistor for 2.7A. Webench picks a closest standard value and I'm not sure they have all of the values in there. But the typical way to tune your current would be to use a parallel combination and/or adjust your maximum IADJ voltage to get the maximum current you need.

For full power mode, I'm not using anything to drive the I_Adj pin. (For low power mode, I use a 50k trimmer to ground using the internal 5uA current.)

To clarify, the 80mohm resistance I am using is for R_sense. How did you determine 9.11mohm?

My understanding is that if I_Adj is open, the circuit provides full power, and a variable voltage or resistance is only needed for dimming.

Thank you for the continued support!

Edit: Along the same lines ... how do I determine what power rating is required for R_Sense? WEBENCH selected a 3W resistor which really limits my choices when it comes to R values. Is it simply [I^2 * R] where I is the average LED current and R is the sense resistor. In this case, [(2.7)^2 * .080] = 582mW? If so, I wonder why a 3W was chosen. Perhaps that's just a limitation of the webench tool.

I'm sorry, typo. I meant 91.1mohm. That is assuming the full 246mV sense voltage. But I'm not sure how much the opto is loading the IADJ pin. There must be some leakage and it could affect the internal clamp voltage a little. As for the power rating the way you are calculating is correct. You only need a 1W resistor, not a 3W. I always tell people that webench is a good starting point, but it isn't perfect. You will need to use some common sense just as you have and realize that everything it picks out is not necessarily ideal in the end.

Awesome :) Really appreciate the quick replies.

It's a great tool. Considering my limited but growing EE knowledge, it's a real life saver to get a good starting point.