Practical Task 2: On-board
Charger for Electric Vehicles
Name: Lucía García-Tristan García
Email: go35top@[Link]
Task:
The main task is to design an LLC resonant converter to charge an electric vehicle battery from
its minimum voltage to the maximum at constant power
Component: C3M0025065K
Vehicle: Tesla Model X Dual Motor
Parameters of the vehicle:
Cell strings in Cell Voltage [V]
Cells in series Power [kW]
parallel min/nom/max
110 72 17 3.0/3.6/4.2
Specifications:
The specification of the charger are the following:
Input voltage range: 360 to 400 volts
Output voltage ripple: 1%
Maximum voltage across resonant capacitor: 1500 volts
Possible switching frequency range: 50 kHz to 300 kHz
Thermal impedance between transistor case to heat-sink: 0.22 K/W
Heat-sink temperature: 60ºC
Maximum semiconductor junction temperature: 20◦C below its maximum allowed junction
temperature limit. (Transistors can be parallelized if the thermal condition is not reachable
with only one component.)
The charger should be able to operate in any input and output voltage condition, but the
power can be derated to 50% when the battery cell voltage is between its minimum and
nominal values.
Battery voltage calculation:
Explanations: Calculate the battery voltages considering that we have 110 cells in series
Minimum: V bat , min=110 ×3.0=330 V
Nominal: V bat , nom =110 ×3.6=396 V
Maximum: V bat , max=110 ×4.2=462 V
The charger must operate at constant power from minimum to maximum voltage.
Constant power: Pout =17 kW
Choice of resonant topology:
Now we analyse the characteristics of each of the possible topologies to choose from:
Full-Bridge: Higher power capacity, better transformer utilization, more components (4
primary MOSFETs).
Half-Bridge: Fewer components, but greater stress on resonant capacitors and more
demanding to achieve the desired gain range.
Decision: Given the required power of 17 kW, we will use a full-bridge configuration to achieve
better losses and a wider operating range. Furthermore, the MOSFET used is SiC, which is
suitable for high frequencies and this topology reduces currents in the primary and improves
thermal management.
Diode selection:
Now we select the device to use on the secondary, to choose one correctly, let's look at the
current and voltage ranges it will have to withstand.
V out , max=V bat ,max =4 62V
Pout 17 kW I out ,max
I out ,max = = =51.5 A → I D= =25.75 A
V out ,min 330 V 2
We chose the component IDW30G120C5B, we look now at its features
1. Nominal current (30A), that provides a margin over the required 25.75A
2. The voltage rating (1200V) is oversized but safe for transients
3. SiC Schottky, this is ideal for high frequency applications
4. It belongs to the package TO-247-3 (the same as TO-274)
LLC design:
We choose the resonant frequency:
The permitted range is 50-300 kHz. Considering that higher frequencies result in smaller
component sizes but higher losses, and that the SiC MOSFET is optimal between 100-200 kHz,
the selected component is f r=100 kHz → wr =2 π f r=628.32 krad /s
Turns ratio n:
In a full-bridge LLC converter, at resonant frequency (f r=f sw ) the gain is approximately 1. For
the design point selection, we chose to design for the nominal point where we want to operate
close to f r for maximum efficiency.
To obtain the turns ratio, we obtain the expression for the gain; for this, we refer the voltage
from the secondary to the primary in the form V o ut ´ =n V out . This way:
V out ´ n ·V out
M= =
V¿ V¿
To obtain n we analyse for nominal conditions where the gain M =1, then:
n· V out , nom V ¿, nom 400 V
M =1= →n= = =1.01
V ¿, nom V out ,nom 396V
Limits of gain M:
n·V out ,min 1.01 ·330 V
M min = = =0.8 33
V ¿, max 400 V
V out , m ax
M m ax =
1.01 · 462 V
V ¿, m∈¿ = =1.296 ¿
360 V
Then we choose: N p : N s =1:1
Then we have the approximation:
V¿ V ¿,nom 380
Vo≈ → n= = =0.96
n V out , nom 396
Now we calculate the required limits:
V out ,min 330 V
M min = = =0.825
V ¿, max 400V
The circuit gain will be:
M min
M circuit , min = =0.859
n
Problem: To achieve a gain less than one, we will have to work at frequencies above f. This is
not possible, so to define the number of turns we use the design point, which will be the most
critical.
To ensure that all required gains are ≥ 1 (operation on the resonant frequency):
V ¿, min 360 V
n= = =0.78
V out ,max 462 V
Verification:
330 V
M min = =1.06
400 V·0.78
462 V
M max = =1.65
360 V· 0.78
First Harmonic Simplification:
In an LLC converter, the waveforms are complex because the bridge voltage is a square wave
with multiple harmonics, and the resonant current is sinusoidal but not perfectly so. This makes
the analysis extremely complex. To address this, only the first harmonic is considered, as it is
primarily responsible for transmitting most of the power.
In this way we obtain the following approximations of the voltages in the primary and
secondary:
4V¿
v p ,1 ( t ) ≈ sin (ws t+ ϕ)
π
4 V out
v s ,1 ( t ) ≈ sin (ws t)
π
For calculations, the rectifier is seen as a resistor R. This is because the voltage fundamental
harmonic component over at the input of the rectifier is in phase with its current.
Analysing the secondary side, we reach the expressions:
π V out
I s ,1=
2
⟨ io ⟩ → ⟨ io ⟩ =
Rout
4 V out
V s , 1=
π
2
2
V s ,1 8 V out 8 396
Req = = 2· = 2· =7.47 Ω
I s ,1 π Po π 17 kW
After these calculations, the converter circuit diagram considering the first harmonic is
obtained:
Figure 1: Circuit with transformer and rectifier equivalent impedance
Figure 2: Circuit with transformer equivalent circuit and load referenced to primary side
2 2
Req ´ =n Req =1.01 · 7.47 Ω=7 .54 Ω
The voltage gain, M, is now calculated:
π V s ,1
V out 4 V s ,1
M= = =
V ¿ π V p ,1 V p , 1
4
Calculation of resonance components:
We have 3 parameters to select (usually provided by the professor or chosen by us):
Quality Factor (Q): Controls bandwidth and currents
With a low Q (< 0.3) we obtain a wide gain range, high currents. On the other hand, with high Q
(> 0.5) we obtain a narrow range, lower currents. We choose Q=0.3
Inductance Ratio (r L ) : Controls the gain range
With a low ratio (3-4) we obtain higher gain range, higher magnetizing current. On the other
hand, with a high ratio (6-8) we obtain lower range, lower current, more difficult to achieve
high gains → We choose r L=4.5
Resonant Frequency (f r ): Determines component size. The selection of f r=100 kH has
already been explained.
Parameter selection:
For this selection, we must ensure that the desired gain is achieved under any operating
conditions. We will define the critical point as the operating point at maximum load and
maximum required gain. At this point, the converter must provide the maximum gain (Mg,max)
without entering the capacitive region of the resonant tank.
We now represent the gain as a function of Q and assign different values to rL. We select a
point where the gain is greater than our Mmax. Now we need:
{
D 6=Q2 r L 2−2 Q2 r L +Q2
1
M ( f n , Q ,r L , n ) = ·
fn
2
√ r L −2 r L+1
2
where D =−2Q 2 r +2 Q2 +r 2
n √ f 6 D +f 4 D +f 2 D +1 4 L L
n 6 n 4 n 2
D 2=Q2 r L2 −2Q2 r L−2 r L
Design of Lr :
Starting from the expressions
1
w r , HF =
√ Lr C r
Zr
Q=
Req ´
Q R eq ´ 0.3 ·7 .54 Ω
→ Lr = = =3 .54 μH
w r , HF 628.32 krad /s
Design of Lm:
Starting from the expression
Lr + Lm
r L=
Lr
→ Lm=3.5 · Lr =12.4 μH
Design of C r:
Starting from the expression
1
w r , HF =
√ Lr C r
1 1
→ Cr = 2
= 2
=715.5 n F
wr , HF · Lr (628.32 krad /s) ·2.17 μH
Z r=
√ Lr
Cr
Z m=
√ Lm
Cr
We now verify that w r , HF > wr , LF:
1
w r , LF = =296.08 krad /s <628.32 krad / s
√(Lr + Lm )C r
Thermal analysis:
The specifications indicate Rθ , CS=0.22ºC /W and T s=60 ºC
Looking at the datasheet for component C3M0025065K: Rθ , JC=0.46 ºC/W
T j ,max −T s 150 ºC−60ºC
Pcond , max = = =132.35 W
Rθ , JC + Rθ ,CS Rθ , JC=0.46ºC /W + 0.22ºC /W
Output voltage ripple analysis: