LM3424 Datasheet, PDF (26/50 Pages) National Semiconductor (TI) – Constant Current N-Channel Controller with Thermal Foldback for Driving LEDs

And the uncompensated DC loop gain (T_U0) is approximated:
Buck
Boost
Buck-boost
For all topologies, the primary method of compensation is to
place a low frequency dominant pole (ω_P2) which will ensure
that there is ample phase margin at the crossover frequency.
This is accomplished by placing a capacitor (C_CMP) from the
COMP pin to GND, which is calculated according to the lower
value of the pole and the RHP zero of the system (shown as
a minimizing function):
10. INPUT CAPACITANCE
Set the nominal input voltage ripple (Δv_IN-PP) by solving for
the required capacitance (C_IN):
Buck
Boost
Buck-boost
Use D_MAX to set the worst case input voltage ripple, when
solving for C_IN in a buck-boost regulator and D_MID = 0.5 when
solving for C_IN in a buck regulator.
The minimum allowable RMS input current rating (I_CIN-RMS)
can be approximated:
Buck
If analog dimming is used, C_CMP should be approximately 4x
larger to maintain stability as the LEDs are dimmed to zero.
A high frequency compensation pole (ω_P3) can be used to
attenuate switching noise and provide better gain margin. As-
suming R_FS = 10Ω, C_FS is calculated according to the higher
value of the pole and the RHP zero of the system (shown as
a maximizing function):
The total system loop gain (T) can then be written as:
Buck
Boost
Buck-boost
11. NFET
The NFET voltage rating should be at least 15% higher than
the maximum NFET drain-to-source voltage (V_T-MAX):
Buck
Boost
Buck-boost
Boost and Buck-boost
www.national.com
26