Datasheet

Table Of Contents
DocID024275 Rev 2 19/29
AN4260 Feedback loop calculation guidelines
7.2 Compensation procedure for a DCM buck
The first step is to choose the pole and zero of the compensator and the crossing frequency.
In this case C(f) has only a zero (fzc) and a pole at the origin, thus a possible setting is:
fzc=k*fp
fcross = fcross_sel_fsw /10
where k is chosen arbitrarily. A starting point could be k=5
After selecting fcross_sel, G1(fcross_sel) can be calculated from Equation 2 and, since by
definition it is
C(fcross_sel)*G1(fcross_sel)
= 1, C
0
can be calculated as follows:
Equation 11
At this point the Bode diagram of G1(f)*C(f) can be plotted, in order to check the phase
margin for the stability.
If the margin is not high enough, another choice should be made for k and fcross_sel, and
the procedure is repeated.
When the stability is ensured, the next step is to find the values of the schematic
components, which can be calculated as follows:
From Equation 9
Equation 12
and from Equation 10
Equation 13
The quantities found in Equation 12 and Equation 13 are suggested values. Commercial
values are chosen, let us call them C7_act, R7_act, resulting into fzc_act.
Equation 14
C
0
value is also recalculated from Equation 9
Equation 15
C
0
j2π fcross_sel⋅⋅⋅
1
j fcross_sel
fzc
----------------------------------+
-----------------------------------------------------
H
COMP
G1 fcross_sel()
---------------------------------------------=
C7
Lfsw
V
IN
V
OUT
-------------------------
Gm
C
0
----------------


R4
R4 R5+
----------------------⋅⋅=
R3
1
2 π fzc C7⋅⋅
------------------------------------=
fzc_act
1
2 π R3_act C7_act⋅⋅
-----------------------------------------------------------=
C
0
_act
Lfsw
V
IN
V
OUT
-------------------------
Gm
C7_act
-------------------


R4_act
R4_act R5 4()_act+
------------------------------------------------------⋅⋅=