1
\$\begingroup\$

For building a constant current LED driver, I'm looking at the TI LM3407 datasheet: https://www.ti.com/lit/ds/symlink/lm3407.pdf

For calculating the necessary inductor, there's several things I don't understand:

  • Under 8.1.4 Selection of Inductor The datasheet lists ripple current as \$ I_{L(ripple)} = \frac{\left(V_{IN} - V_{OUT} - 0.198 \times \left(1 + \frac 1 R_{ISNS}\right)\right) \times V_{OUT}}{L\times V_{IN} \times f_{SW}} \$

What I don't understand is where the term with the sense resistor is coming from, my understanding of a buck converter is that it depends only on duty cycle, frequency and inductance, but not output current?
Edit: I guess that's the on resistance of the built in MOSFET - which is given as 0.77 Ohm typical, 1.45 Ohm max (at 80mA).

  • On the next page, there's Table 1. Suggested Inductance Value of the Inductor, which e.g. lists 22µH for driving 5 LEDs with 1MHz from 25V, with \$I_{L(ripple)} = 0.8 \times I_{OUT}\$. But inserting those values into above formula, I arrive at >250mA ripple current, indpendent of \$I_{OUT}\$.
    E.g. with \$I_{OUT} = 100mA, R_{ISNS} = 2\Omega\$, that ripple current seems excessive?
  • The datasheet has some suggested output capacitor sizes, but no information at all for how output capacitor size relates to LED ripple current. What's a good resource to learn about this?
\$\endgroup\$
6
  • \$\begingroup\$ Welcome to EE.SE! Have you tried simulating it? \$\endgroup\$
    – winny
    Commented Mar 26, 2021 at 14:52
  • \$\begingroup\$ I'm relatively new to the field, so no idea how I would do that. \$\endgroup\$ Commented Mar 26, 2021 at 15:59
  • \$\begingroup\$ I don't see precisely that formula in the data sheet 8.1.4 \$\endgroup\$
    – Andy aka
    Commented Mar 26, 2021 at 17:25
  • \$\begingroup\$ It has number of LEDs times forward voltage in there instead of Vout. \$\endgroup\$ Commented Mar 26, 2021 at 18:06
  • 1
    \$\begingroup\$ A great time to learn! Download LTspice (it’s free) and draw your circuit. We’ll help you from there. You can of course derive a formula for it, but if your end goal is a real circuit and not a theoretical exercise, you probable reach your goals quicker with simulations. \$\endgroup\$
    – winny
    Commented Mar 26, 2021 at 18:12

0

Browse other questions tagged or ask your own question.