1

LECTURE 12
Lossy Converters (Static State Losses
but Neglecting Switching Losses)
HW #2 Erickson Chapter 3 Problems 6 & 7
A. General Effects Expected:
V(load)↓ as I(load)↑
B. Switch and Resistive Losses Change M(D)
Modified DC Voltage Transfer
Functions M(D) with loss ≠ M(D) for no loss
1. General aspects of the origin of M(D)
changes in static state loss terms:
RL,Vd, Rd, and Ron
a. Vmax / Vmin
b. V/L Slopes
2. General solution for lossy Boost Converter
a. RL only: M(D, RL), η ≠ 1
b. Static State Device DC losses:
M(D, Vd, Ron, RD)
3. Buck Converter
a. Problem 3.3 of Erickson
4. General solution for lossy Buck-Boost
Converter
a. M(D, Vd, Ron, RL), Problem 3.4 of
Erickson
5. Compare for a specific task:
a. Buck vs. Buck-Boost Circuits,
Problem 3.5 of Erickson
 2

A. GENERAL EFFECTS EXPECTED ON M(D)

Lossy Converters: Consider the static state DC loss only,
no dynamic switch loss will be considered here
Generally one expects, due to finite Rout, from any power
supply for VL to fall as IL increases. In converters Rout is
much more complex, especially with feedback.

VL Rout effect on VL

IL
For fixed Pout and Vin for a converter the input current
increases proportional to the losses so that Pin=Pout+losses.
The losses we will consider will also include device losses
but only in the static states due to finite Ron and Von values.

B. Effects on voltage DC transfer function M(D) due to

resistive and device voltage drops. Besides decreasing
converter efficiency static state losses will also effect dc
transfer functions. M(D) becomes a function of RL, Von, Ron,
etc. when losses are included as shown below.
1. Vmax/Vmin changes as seen by any converter
inductor will cause di/dt slope changes.
⇒ ∆iL = V/L (in units of A/sec) slopes change,
affecting D and D’themselves just by including
resistance and device voltage drops. Hence, the
value of M(D) changes.

2. Including inductor resistance RL, device Ron, as

well as Von changes M(D), from M(D) with these
effects ignored. Lets consider the inductor
 3

resistance effects.

a. Illustrative RL (inductor series equivalent

resistance) effects in boost topology

M will change from M(D) to M(D, RL) as shown below.

RL is finite not only because of simple copper resistive
losses. Wound inductors at high frequencies (0.1 - 1 MHz)
have additional core and proximity effect losses. To a first
approximation: Rac/ Rdc =0.5∏ (wire dia. in cm)X f1/2 x100
Larger diameter wire has more effect than a small diameter
Below we carefully distinguish between RL with the dominant
R for the load.
Boost with RL during D Boost with RL during D’
i L RL iC i L RL

+ - + + - +
vL vL iC
Vg C R v Vg C R v

- -

vL(t)
Vg-IRL
1. V-sec balance across VL DTs D'Ts
t

Vg-IRL-V

D[Vg - IRL] + D’[Vg - IRL - V] = 0

⇒ Vg - IRL - D’V = 0
iC(t)
I-V/R
2. Current-time balance
across the capacitor DTs D'Ts t

-V/R
 4

D[-V/R] + D’[I - V/R] = 0

⇒ D’I - V/R = 0

Usual  
Second term has all effects
   of including finite RL.
Vo 1 Boost  1
=   ← RL>D’2R big effect
Vg D' lossless1 + L R
  2 
Also big effects occur for
term  D ' R
D’→ 0 or for D→ 1.0
Note some extreme converter operating conditions will
amplify finite RL effects.

Note well:
(1) R is the load
resistor
(2) RL = inductor losses

For D→ 1.0
RL/R has a big effect
on M(D)
As RL/R→ 0 no effect
on M(D)

Note also as D → 1.0, M(D) has big “rollover” if RL is even

1% of the load resistance. Note also in DC gain at high D
that for D< 0.2, RL effects are minimal.

Summary
The operating duty cycle, D value, amplifies equivalent
series resistance ,RL, effects.
D > 0.5 and RL/R even 1% you see big changes to M(D).
D < 0.5 you see no effects.
 5

I RL

+
+ Equivalent Circuit with
Vg D'V
+
D'I V R
“DC Transformer”:
-
- Use to get the
-
operating
efficiency
η = Pout / Pin

I
RL Equivalent circuit model
D':1
of the boost converter,
+ including a D’:1 dc
transformer and the
V R inductor winding
resistance RL.
-

Referring to the secondary to get one loop equation.

D'I

RL/D'2 + Clearly want RL << D’2R

Easy if D’= 1.0, D = 0
Vg/D' Vo R
Hard to achieve if D’→ 0
- D→ 1.0

−1
Ideal  RL 
Vo = Vg  1 + 
Case D'2 R
↑
1/D’
 6

−1
Vg Ideal  RL 
I(input) = 1+ 
2 
D' Case  D' R
2
−1
P V  RL 
η = out = o D' = 1 + 
Pin Vg  D '2 R 
η(D, RL/R)

1) As the on duty
cycle D→ 0 one
gets high η but
efficiency is little
effected by RL

2) D→ 1 and D’→ 0
η → 0 still we get
much lower η for
RL not negligible
and much greater
sensitivity to RL
changes.
Compare the above to the Lossless Boost:
D':1

I1 I2 +

Vg Vo

I2 = D’I1 η = D’Vo/Vg = Po/Pin

Vo = Vg/D’ = 1.0 (lossless)
Next we include static device losses to the system loss.
b. Static State Device Loss Effects in the Boost
Topology.
 7

In our static states of the switching devices we only consider

Ron from the transistor but both RD and VD for the diode when
on, because Von(TR) ≈0. We expect to obtain M(D, RL, Von,
RD, VD) and η(D, RL, Ron, RD and VD).

2
i L +
i L + RL
RL
1 Vg C R v
Vg C R v
-
-

Fig. 3.22. Boost converter

Fig. 3.6. Boost converter example.
circuit, including inductor
resistance RL.
SW position 1: Transistor is on case: Ron(I)
SW position 2: Diode on case: VD and RD

Effects of Ron, VD and RD follow from revised circuits which

change values of VL and hence circuit D and D’periods as
well.

Period DTs Circuit Period D’Ts Circuit

Connections Connections
L L RL RD

i iC + i + vL - +
RL VD iC
Vg Ron C R v Vg R v
C
- -

Next, we invoke both volt-sec and current-sec

balance on the inductor voltage ,VL, and capacitor
current ,IC, respectively.
 8

(1) Volt-sec balance on VL vs. time

vL(t)
Vg-IRL-IRon D[Vg - IRL - IRon] +
DTs D'T s D’[Vg - IRL - VD - IRD - Vo]
t

Vg-IRL-VD-IRD-V

Input Conditions = 0 yields:

Vg - IRL - IDRon - D’VD - ID’RD - D’Vo = 0

This is the input circuit loop for the above equation:

RL DRon D'VD D'RD

I + - + - + -
IRL IRon ID'RD
Vg D'Vo +
-

(2) Current-sec balance on Ic vs. time

iC(t)
I-V/R

DTs D'Ts t D[-V/R] + [I - V/R] = 0

-V/R

Output conditions yield the simple equivalent output circuit

result:
+
+ V
D’I - V/R = 0
D'I - R

-
 9

Combining Input/Output with a DC transformer:

RL DRon D'VD D'RD D':1
I +

Vg V R

Vo 1  D'V D  1
= 1 −  R + DR + D' R =M(D, RL, Von, RD, VD)
Vg D'   V g 1 + L ON D
D'2 R
↑ ↑ ↑
lossless VD Resistive Effects
ideal factor Only for R → ∞ do
we get full output

For the extreme D = 0 ⇒ DRon = 0, D’RD = RD and

resistive factor is slightly simplified
P VD' I V
η= o = = D' , but V/Vg differs with losses
Pin VgI Vg
 D' VD   RL + DRon + D' RD 
η = 1 −  1 + 2 = η(D, RL, Ron, RD, VD)

 Vg   D' R 
↑ ↑
Diode on Resistive
Voltage Effects

(3) The following is problem 3.3 of Chapter 3

Erickson
Buck Converter with:
•Input L1-C1 filter L1 has RL1 winding
resistance due to copper wire
 10

•
Transistor with Ron
•
Diode with RD and VD
Lossless Buck:
1:D

I1 I2 +

Vg Vo

Vo = DVg η = Vo/(DVg) = Po/Pin

I1 = I2/D = 1.0 (lossless)

L1 RL1 L2 RL2

i1 iC1 i2 iC2 +
+
Vg vC1 C1 D1 C2 R V
-
-

a) Q(on)
L1 RL1 Ron L2 RL2

i1 + - i2 + - +
+
Vg C1 C2 R V
-
-

VL1 = Vg - VC1 - I1RL1

VL2 = -I2Ron + VC1 - V - I2RL2
IC1 = I1 - I2
IC2 = I2 - V/R
Next we will consider the switch case when the transistor is
off and the diode is on.
 11

Q(off)
L1 RL1 L2 RL2

i1 + - + - iC2 +
+ iC1 VD
Vg C1 C2 R V
- RD
-

VL1 = Vg - VC1 - I1RL1

VL2 = -VD - I2RD - V - I2RL2
IC1 = I1
IC2 = I2 - V/R

Do Volt-second balance on L’s / Charge balance on C’s

1. L1: D(Vg - VC1 - I1RL1) + D’(Vg - VC1 - I1RL1) = 0
∴ Vg = VC1+I1RL1
2. L2: D(-I2Ron + VC1 - V - I2RL2) + D’(-VD - I2RD - V - I2RL2) = 0
∴ -DI2Ron + DVC1 - D’VD - D’I2RD - V - I2RL2 = 0
I1
3. C1: I1 - DI2 = 0 ∴ I1 = DI2
DI2

4. C2: I2 - V/R = 0 ∴ I2 = V/R

Input Circuit
V
From 1, 3, and 4 VC1=Vg - I1RL1=Vg - DI2RL1=Vg - DRL1
R
V
∴ VC1 = Vg - DRL1
R
Output Circuit:
Then:
V V V V
(2) -DRon + DVg - D2RL1 - D’VD - D’RD - V - RL2 = 0
R R R R
 12

DRon D2RL1 D'VD D'RD RL2

+ - + - + - + - +
DVg R Fig. 1
I2
-

DC Trf Model:
DRon D2RL1 D'VD D'RD RL2
1:D
+ - + - + - + - +
Vg R Fig. 2
I2
-

b) From Fig. 1
V V V V
V = -DRon + DVg - D2RL1 - D’VD - D’RD - RL2
R R R R

c) Using the Voltage Divider:

R
V= (DVg - D’VD)
R + DR on + D R L1 + D' R D + R L2
2

Filtered Buck
V V R
= (D - D’ D )
Vg Vg R + DR on + D 2 R L1 + D' R D + R L2

V D' VD R
η= = (1 - )
DVg DVg R + DR on + D 2 R L1 + D' R D + R L2
 13

Next we solve the general equations for the Buck-Boost

4. Buck-Boost solved in general

DI 1:D L D':1
+
I
Vg Vo

(a) Equivalent Series Resistance ,RL, losses only

•Assume lossless except for RL
•Reflect all to load side. For DC L → short and
C → open.
D'I
+ For DC conditions:
RL/D'2
Vg(D/D') R Vo
R D
Vo = Vg
- RL D'
+ R
2
D'
(b) Device losses only
Buck-Boost with device loss (RL ≡ 0)
Transistor: Ron
Diode: VD
D'VD DRon
DI 1:D D':1
+

Vg I Vo

-
 14

VD
D'I
+
DRon/D'2
Vg(D/D') R Vo

-
 D  R
Vo = Vg − VD 
 D' R + DRon D'2
(c) Buck-Boost Converter including all static
losses. This is problem 3.4 from Erickson

1A
+

load
Vg=1.5V 100 µH 5V

-
fs = 40 kHz

Vout required is +5 V and Vg = -1.5 V

Diode: VD = 0.5 V Schottky
RD = 0 (approx.)
Transistor: Ron = 35 mΩ

Ideal Buck-Boost:
Vo − D − D
= =
Vg 1 − D D'

Real Buck-Boost: we will find in the next few pages

D  R
Vo =  Vg − VD 
D' R + other (Ron,RL )
↑ ↑
Diode Resistive
Effect Effect
 15

Below we work out Erickson problem 3.4 for illustration of

how to do the HW for Chapter 3
HW # 2 will be Erickson Chapter 3 Pbms.6 and
7. This will be due next week along with any
questions asked in the lectures
16
 17

(5) Compare Buck vs. Buck-Boost Topology for DC

Conditions Only. This is problem 3.5 of Erickson.

10A
500V DC-DC
400V
DC Bus Converter

Transistor: Ron = 0.5 Ω RD → 0

Diode: VD → 0 Inductor: RL → 0
 18

Find D and η for placing different topologies inside the DC-

DC converter. Then compare why one topology might be
favored over the other for the specified conditions.
Prob. 3.5
Only loss is Mosfet with Ron = 0.5Ω
Buck Buck-Boost
L D1
ILoad =10A ILoad=10A
i + vL - iC + i iL iC +
-
Vg V Vg V
D1 C R vL L C R
500v 400v 500v 400v
+
- -

Mosfet on and Diode off. Mosfet on and Diode off.

Ron L ILoad=10A ILoad=10A
i i iL iC +
+ vL - iC + Ron -
Vg vL L C R V
Vg C R V +
-
-

vL(t)= Vg - i Ron - V ≈Vg - I Ron -V vL(t)= Vg - i Ron ≈Vg - I Ron

ic(t) = i -10 ≈I - 10 ic(t) = -10
Mosfet off and Diode on. Mosfet off and Diode on.
L 10A
10A
i iC + iL iC +
+ vL -
-
Vg vL L C R V
C R V +
- -

vL(t)= -V vL(t)= -V
ic(t) = i - 10 ≈I - 10 ic(t) = i - 10 ≈I - 10
Thus averaging we get: Thus averaging we get:
<vL> = (Vg - I Ron - V)D + (-V)D’= 0 <vL> = (Vg - I Ron)D + (-V)D’= 0
<ic> = D(I - 10) + D’(I - 10) = 0 <ic> = D(-10) + D’(I - 10) = 0
So ⇒ DVg - DIRon - V = 0 So ⇒ DVg - I Ron D - V + VD = 0
I - 10 = 0 ⇒ I = 10 ID’- 10 = 0 ⇒ I = 10/D’
 19

DRon DRon

+ + + +
DVg V DVg I VD'
- I - - -

DRon Ig DRon
1:D 1:D D':1
. .+ -
+
. . . . +
Vg I R V Vg I R V
- -

First we need D:
DVg - 10RonD - V = 0 DVg - 10RonD/D’- VD’= 0
DVg - VgD2 - 10RonD - V + 2VD -VD2 = 0
D = V/(Vg - IRon) D2(-Vg - V) +D(2V + Vg - 10Ron) -V = 0
Using the quadratic equation for D:
D = 0.81; D’ = 0.19 D = 0.45; D’ = 0.55
Solving for efficiency: Ig = DI Solving for efficiency: Ig = ID
η = Pout/Pin = V Iload/VgDI η = Pout/Pin = V Iload/VgID
= (4000)/[DVg(10)] =(4000)/[VgD(10/D’)]
η = 0.99 = 99% η = 0.978 = 97.8%
Ploss = Pin - Pout Ploss = Pin - Pout
Ploss = 4050 - 4000 = 50 Watts Ploss = 4090 -4000 = 90 Watts
The BUCK is better for this application!
Another way to calculate the Power loss is to use the voltage divider and find Vg with
Respect to V as shown below:
RDVg RDVg
V= V=
R + DRon DRon
D'(R + )
D'2
Knowing that:
V = DVg for the lossless Buck: V = DVg/D’for the Buck-Boost:
R
we can see that =1
R + DRon
for the lossless Buck case thus:
R R
η= η=
R + DRon DRon
(R + )
D'2
η = 0.99 = 99% η = 0.98 = 98%
since η = Pout/Pin:
Pin = Pout/η = 4000/.99 = 4040.4 W Pin = Pout/η = 4000/.98 = 4081.63 W
Ploss = 4040.4 - 4000 = 40.4 W Ploss = 4081.63 - 4000 = 81.63 W
The BUCK is better for this application!