6.622 Power Electronics Prof.

David Perreault

Lecture 7 — DC/DC Lecture 3

1 Converters
Consider device and passive component ratings of converters:
Indirect:

For now, assume L, C big

iL ≈ IL , vc ≈ Vc

IL = |I1 | + |I2 |
Vc = |V1 | + |V2 |

vsw,max = vd,max = vc = |V1 | + |V2 |

So we have
isw,max = id,max = IL = |I1 | + |I2 |

Let’s look at direct converters (L, C: big: iL ≈ IL , vc ≈ Vc )

vsw,max = vd,max = Vc = V1 vsw,max = vd,max = Vc = V2

isw,max = id,max = iL = I2 isw,max = id,max = iL = I1
vsw,max = vd,max = vc = max|V1 | + |V2 |
For the direct converter types
isw,max = id,max = IL = max|I1 | + |I2 |

1
 So based on device and passive component stresses, we would choose a direct converter over an indirect
converter whenever possible!

In practice, component election does depend on ripple in many classes. Let’s see how to approximately
calculate ripple efects.
To calculate capacitor voltage ripple, we:

1. Neglect ripple in inductor (assume L ≈ inf so ∆i2,pp ≈ 0)

2. assume all current voltage ripple goes into capacitor

3. calculate voltage ripple

4. verify assumption afterward

Ex: Boost converter ripple

So we model the system assuming all ripple current component (ĩd ) goes into the capacitor, and the old
1
dc component < id > goes into the resistor. For this to be true, 2πfsw >> RC

2
 Under this assumption, a “ripple only” model is:

• ripple in vc is triangular
vc,pp
• the average value of a ∆ wave is half-way between its peaks, so we get ripple between ±δ 2

Z
dvc 1
i=c → ∆vc = ic dt
dt c

D(1 − D)IL T
∆vc,pp = where IL = I1
C
D(1 − D)I1 T
So to limit ripple to be low a specifed value C≥
∆vc,pp
Similar calculations can be made for current ripple in iL .

The fat black line says: neglecting capacitor voltage ripple (a good approximation)
Blue line says: real case (is ∞ with ripple calculated above)

Modeling vx with equivalent source, throw out dc components:

(Text on left:

3
neglecting zx , vL = −v̂x ) (On right arrow: Neglecting any drop on zi )

∆i
V =L∗
∆t

D(1 − D)V2 T
iL,pp =
L
D(1 − D)V2 T
Therefore, L≥
∆iL,pp
From these waveforms we can defne Ripple Ratios: ripple magnitude as a fraction of dc value

∆ ∆vc,pp/2
Rc =
Vc
∆ ∆iL,pp/2
RL =
IL

∴ vc,pk = Vc (1 + Rc ) and i2,pk = IL (1 + RL )

So from our previous results
D(1 − D)T v2 D(1 − D)I1 T
L≥ C≥
2I1 RL 2Vc Rc
Energy storage is one metric for the minimum size of an energy storage component. What is required
energy storage?
Capacitor:
1
Ec = CV 2
2 c,pk
1 D(1 − D)I1 T
= [Vc (1 + Rc )]2
2 2Vc Rc
1 D(I1 V1 )T (1 + Rc )2
2 2 Rc

DT P0 (1 + Rc )2
Ec =
4 Rc
So energy storage increases with
1. switching period
2. output power
3. conversion ratio
4. smaller ripple spec (? cannot read space or spec)
2
DP0 (1+Rc )
Similar arguments for inductor EL = 4f sw Rc
It can be shown that direct converters always require less energy storage (+ hence smaller components)
than indirect converters.

4
 We can also factor in ripple on our peak device stresses:
DIRECT:
vc,pk = vsw,pk = vd,pk = max(|v1 |, |v2 |)(1 + Rc )
iL,pk = isw,pk = id,pk = max(|I1 |, |I2 |)(1 + R2 )
INDIRECT:
vc,pk = vsw,pk = vd,pk = (|v1 | + |v2 |)(1 + Rc )
iL,pk = isw,pk = id,pk = (|I1 | + |I2 |)(1 + R2 )

If time, defne metric for switch size: switch stress parameter

∆
S.S.P. = Vsw,pk ∗ isw,pk

5
MIT OpenCourseWare
https://ocw.mit.edu

6.622 Power Electronics

Spring 2023

For information about citing these materials or our Terms of Use, visit: https://ocw.mit.edu/terms