Ben Gurion University of the Negev
Department of Electrical and Computer Engineering
Power Electronic Laboratory
Senior Project: P-2013-049
Digital controller for a resonant converter
Students: Bar Halivni and Nadav Cohen
Supervisor: Prof. Shmuel (Sam) Ben-Yaakov
Mr. Michael Evzelman
Project Submission: August 2013
Abstract
Traditional Switched Capacitor Converters (SCC) suffer from switching losses, to overcome this problem a DCDC resonant convertor with Zero Current Switching (ZCS) is used. To successfully control a system of this type,
advanced design and programming of a control system based on digital microcontroller is required.. Zero
Current Switching requires us to perform the switching when there is no current flowing through the circuit;
therefore there is no power loss during the switching time. By using a resonant converter the current flows in
a sinusoidal pattern created by the resonant components in the circuit, enabling ZCS.
Nowadays the control system of the convertor we are about to use in this project is performed by an open
loop method that requires an outside interference which takes time and human resources to achieve optimal
performance.
The project goal is creating an automated control system with a closed loop feedback, which changes the
duty cycle of the switches and tracks Zero Current Crossing, that doesn’t demand manpower. Implanting zero
current crossing control grants us benefits in pricing and efficiency such as: lower power-loss, programming
time and achieving more accurate switching times for changing conditions within the system.
The challenges we expected to run in to at this project were: finding a microcontroller that answers all the
requirements needed for the control system, coping with circuit’s noise and optimizing zero current crossing
track. We will be tracking the progress of the project by simulations and experimental tests.
Resonant Converter
Resonant converters belong to the family of DC-DC converters called SCC (Switching Capacitor Converters).
The SCC converters use capacitors as energy storage and delivery component. The basic topology of the SCC
has two phases: A charging phase which charge the capacitor from the source and then discharge the
capacitor to the load ) Figure 1(.
In case of resonant SCC a coil is added in serial connection to the capacitor, forming the LC pair in order to
create sinusoidal capacitor current. The coil inductance will determine the resonant frequency of the
sinusoidal current that will dictate the switching frequency of the control system. The three phase resonant
converter (Figure 2) used in this project developed at the Power Electronics Laboratory, including three
switching phases: Charging, Discharging and Auxiliary.
Cf
Figure 1: The basic opology of SCC (Cf is flying capacitor)
L
Vin
Charging
Auxiliary
Ro
Discharging
Co
Cf
Figure 2: The basic opology of Resonant Converter
Design System
In order to achieve the goal of the project we designed and build a control system which consists of an FPGA
digital controller and Current Sense Network which allow us to perform ZCS accurate as possible.
The switching cycle can be divided into two main processes, the first is the detection of the zero current
crossing which performed by the current sensing network, that convert the converter analog current signal to
a digital signal using an analog comparator. The second part of the processes is the reaction stage, at this
stage the FPGA use the signal received from the sensing network to determine if ZCS did accrued using the
software algorithm (Figure 5) and then change the state of the switches if need (Figure 3).
The current sensing network – contains of four main parts, the first part is the current transformer which
gives us a duplicate image of the capacitor current that we want to detect for zero current crossing. The
second part is a full diode bridge resulting a positive polar current signal regardless of the input polarity; this
step is required due to the capability of the digital converter to receive only positive voltage inputs. The
second stage is a conversion of the current signal to a voltage signal using a conversion resistor of a 330 Ohm.
The final part of the network is the high speed alanog comparator converting the analog voltage signal to a
digital signal which can be read by the FPGA, to protect the FPGA from high voltage level we used a open
collector comparator with a digital 3.3V output level (see Figure 4).
For a digital controller we selected the Altera's CycloneIV FPGA using his high speed operating and
programming versatility to create the best conditions for a successful ZCS. Using those attributes while
programming the controller by the algorithm presented at figure 5 using Verilog, we were able to achieve real
time ZCS as shown at the experimental results chapter.
Power Stage
IL
VIN
S1
VOUT
S2
VIN
IL
VREF
S3
VCOMP
VSW1
Vsw(1,2,3)
VSW2
FPGA
VCOMP
+
VIN
-
VSW3
Current Sense
Network
VREF
1. Charging Phase
(b)
(a)
2. Discharging Phase
3. Auxiliary Phase
Figure 3: (a) System Block Diagram (b) wavefoems
To/From
Development
Board DE2-115
Resonant
Current
To Switches
Operating
System
Full Diode Bridge
Conversion Resistor 330 Ohm
Voltage Regulator 3.3V
Analog Comparator LM319N
Figure 4: Current Sense Network & Analog Comparator
Start
Set
State
Output by
Present State
YES
NO
Zero crossing
occurred?
YES
Was it in
Mask Time?
YES
End Dead
Dead Time
time
NO
Update
State to
State
Next State
State
NO
Start Dead
Time
Dead time
Mask Time-Time when the
noises occurred
Dead time-Time when
switches are turned off.
Figure 5: FPGA algorithm
The Complete Project:
Figure 6: A Picture of Final System
1. Digital controller and development board – Altera DE2 -115 with EP4CE115F29C8 FPGA chip.
2. Switches Operating system – 15V high speed floating drivers.
3. Input Power Source Filter.
4. Current Sense Network and Analog Comparator – Current sensing components with a LM319N
comparator.
5. Input Capacitors.
6. Current transformer - CST306 -3A current transformer.
7. Resonant system - LC Tank contains of the resonant coil and capacitor.
8. Switches – Three switches build from six IRFP3077 MOSFET with a back to back connection.
9. Output Capacitance.
10. Load – Grid of a 50 ohm power resistor to create any wanted output resistance manually.
Specifications of Resonant Converter:
Switching Rate
50KHz-200KHz
Input Voltage
10V-60V
Output Voltage
10V-60V
Converter with whom we work capable
of producing variable Input Output
ratio. Attenuation or amplification up to
four.
Input Output Voltage Ratio
~100W
Power
Number of Switches
3 switches two - directions
include 6 type MOSFET
transistors IRFP3077.
Input Capacitance
28mF
1500µF
Output Capacitance
Drivers float with a 15V operating
voltage.
Switches Operating System
Resonant Capacitance
235nF
Resonant Coil
6.8µH
Figure 7: Specifications of Resonant Converter
Experimental Results
After the completion of the control system design and build, we applied the finished system on to the
resonant converter with a view to perform tests and measurements to check the performance of our control
system compared to the old manual system. The tests were performed on four different ratios of inputoutput voltage: 1:2.75 increase, 1:1.4 increase, 1:2 decrease, 1:3 decrease.
The results were gathered from two voltmeters to measure input and output voltages, active current probe
and voltage probes using a scope for the circuit waveforms. The tests measurements are presented and
compared over different test conditions using the following graphs:
Input&Output voltages [V]
1. Voltage Measurements
Input & Output voltages by Conversion
Ratio
40
27.9
27.7
30
30
25
20
20
12.5
10
10
9.08
0
1:2.75
1:1.4
1:2
1:3
Conversion Ratio
Figure 8: Input(Blue) Output(Red) voltages by Conversion Ratio
As demonstrated in figure 8, we tested the system over a wide range of conversation ratios, showing that the
control system capable of operating the converter in a wide range of input and output voltages.
2. Power Measurements
Input&Output power [W]
Input & Output power by Conversion Ratio
20
15.7 15.6
15
10
8
7.7
9.5 9.4
8.7 8.2
1:2
1:3
5
0
1:2.75
1:1.4
Conversion Ratio
Figure 9: Input (Blue) Output (Red) power by conversion ratio
From the results as showed at the graph above (Figure 9) we can see that over all the operating conversion
ratings the input and output power are very close, showing high efficiency of the converter using our control
system.
In previous tests we managed to achieve output power of 30W, but because of its proximity to the maximum
voltage of the switching transistors we preferred to do the measurements using lower power rates in order to
avoid any circuit damage.
3. Switching Error
In order to evaluate the most important goal of the project, which is an accurate ZCS, we defined the
switching error criteria. Using the switching error we can compare and evaluate the accuracy of ZCS.
The switching error is examined by a new criterion that we define as:
𝜀=
|δ|
∗ 100[%]
∆
The error parameter ε is calculated by the ratio between the error current measured in a transistor at the end
of the phase (δ) and the amplitude of phase (Δ). Since ε represents the absolute error for switching, we use
the absolute value of the error current which can be positive or negative current. As the goal of the project is
to perform an accurate ZCS we want the switching error to be as small as possible, which means we achieved
better accuracy.
Switching Error by Conversion Ratio
14.00
Switching Error [%]
12.00
10.00
8.00
ε Charge Phase %
6.00
ε Discharge Phase %
4.00
ε Auxilary Phase %
2.00
0.00
1:2.75
1:1.4
1:2
1:3
Conversion Ratio
Figure 10: Switching Error by Conversion Ratio
From figure 10, we can see that our highest error is less than 12%. The error measured at the charge phase
for each conversion ratios as shown above is almost constant in value of 11%. The graph shows that for a
voltage increase conversion ratio, a smaller error is obtained for the discharge and auxiliary phases due to
their small amplitudes, which indicate the lower current slope is we receive a lower the switching error.
Average Switching Error by Phase
Average Switching Error[%]
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
Charge Phase
Discharge Phase
Auxillery Phase
Figure 11: Average Switching Error by Phase
The graph above shows that for the charge phase which has the largest amplitude and thus the largest slope,
an 11.6% average switching error is shown for all the conversion ratios. For the discharge phase we receive a
minimum average error of 6% as shown in Figure 11.
)a(
) b(
)c(
) d(
Figure 12: Waveforms – (a) 1:2.75 increase, (b) 1:1.4 increase, (c) 1:2 decrease, (d) 1:3 decrease
The control system results on the live circuit are shown by the waveforms above, for different ratios of inputoutput voltage: 1:2.75 increase, 1:1.4 increase, 1:2 decrease, 1:3 decrease (Figure 12). The pink waveform
describes the shape of the real capacitor current achieved by a active current probe, the yellow waveform
describes the shape of the copied capacitor current after alignment using a full diode bridge using the current
transformer, the green waveform describes the comparator reference voltage and the blue waveform
describes a control signal to one of the switches.
References
[1]
S. Ben-Yaakov, and M. Evzelman, "Generic Average Modeling and Simulation of the Static and Dynamic
Behavior of Switched Capacitor Converters", IEEE Applied Power Electronics Conference, APEC-2001, 535-541,
Anaheim, California, 2001.
[2]
[3]
[4]
S. Ben-Yaakov, and M. Evzelman, "Generic and unified model of switched capacitor converter", IEEE Energy
Conversion Congress and Exposition, ECCE 20-24 Sep. 2009.
Y. P. B. Yeung, K. W. E. Cheng, S. L. Ho, K. K. Law, and D. Sutanto, "Unified analysis of switched-capacitor
resonant converters", IEEE Transactions on Industrial Electronics, vol. 51, no. 4, pp. 864 – 873, 2004.
S. Ben-Yaakov, and M. Evzelman, A. Cervera, "A Gyrator Behaved Bi-Directional Buck-Boost Resonant
Switched Capacitor Converter", Unpublished.