In this paper, I present a complete study and hardware implementation of a charging and protection circuit dedicated to a 48 V – 20 Ah lithium iron phosphate (LiFePO₄) power battery pack. This work is motivated by the fast-growing market of electric power tools, light electric vehicles, and the increasing demand for an improved EV battery pack management system. The central objective is to develop a reliable, efficient, and safe power conversion and monitoring architecture for small and medium sized electric powertrain applications. My approach combines a two-stage switched-mode power supply with active power factor correction, a digital protection and monitoring unit based on a high-performance mixed-signal microcontroller, and a practical state-of-charge estimation method. Throughout the paper, I emphasize the importance of an EV battery pack as the primary energy storage element, and I analyze every design decision from the perspective of safety, cost, efficiency, and long cycle life.
1. Introduction
The global increase in the number of fossil-fuel vehicles has led to severe environmental pollution and energy shortages. It is widely reported that motor vehicles emit more than 200 million tons of harmful gases annually, accounting for over 60% of atmospheric pollution. In addition, the depletion of petroleum resources has forced the transportation sector to seek alternative energy carriers. Electric vehicles (EVs), including electric bicycles, electric motorcycles, and electric cars, represent a promising solution. The key component of any EV is its energy storage system, and the EV battery pack is the most challenging part to optimize. Traditional lead-acid batteries are heavy, bulky, have low specific energy, and suffer from limited cycle life. Moreover, lead is highly toxic and poses environmental threats. Therefore, the development of high-performance lithium-based batteries has become a critical research area.
Lithium-ion batteries have been successfully commercialized for portable electronics, but their application to high-power traction requires careful attention to safety, thermal management, and charge/discharge strategies. Among the various cathode materials, LiFePO₄ has emerged as the most promising candidate for an EV battery pack due to its excellent thermal stability, long cycle life, high discharge current capability, and low material cost. Table 1 compares the main lithium battery cathode materials.
| Cathode Material | Theoretical Capacity (mAh/g) | Actual Capacity (mAh/g) | Working Voltage (V) | Safety | Cost |
|---|---|---|---|---|---|
| LiCoO₂ | 274 | 140–150 | 3.7 | Fair | High |
| LiNiO₂ | 274 | 190–210 | 2.5–4.2 | Poor | Medium |
| LiMn₂O₄ | 148 | 90–120 | 3.0–4.0 | Good | Low |
| LiMnO₂ | 286 | 200 | 3–4.5 | Good | Low |
| LiFePO₄ | 170 | 110–165 | 3.2 | Excellent | Low |
For this project, I selected a LiFePO₄ EV battery pack with the following specifications:
| Parameter | Value |
|---|---|
| Nominal voltage | 48 V (15 cells in series) |
| Nominal capacity | 20 Ah |
| Maximum charge current | 20 A (1C) |
| Maximum discharge current | 20 A (1C) |
| Operating temperature | 0–50 °C |
| Charge cutoff voltage (cell) | 3.6 V |
| Discharge cutoff voltage (cell) | 2.1 V |
The design must achieve the following main functions: charging with high power factor, over-voltage/over-current protection, temperature monitoring, state-of-charge (SOC) estimation, and communication with external devices using the SMBus protocol. My research addresses both the hardware architecture and the software algorithms needed to manage an EV battery pack efficiently and safely.
2. Lithium Power Battery and Its Characteristics
An EV battery pack is not simply a scaled-up version of a consumer lithium-ion battery. The power requirements of electric vehicles demand high current capability, high energy density, low internal resistance, and excellent safety. The LiFePO₄ battery, with its olivine crystal structure, has become one of the most researched systems for automotive applications. Its nominal cell voltage is 3.2 V, the charge cutoff voltage is 3.6 V, and the discharge cutoff voltage is 2.1 V. The specific energy is about 60–75 Wh/kg, and the volumetric energy density is around 220–240 Wh/L.
Experimental studies have shown that LiFePO₄ exhibits:
- High efficiency output: standard discharge rate of 2–5C, continuous high-rate discharge up to 10C, and pulse discharge (10 s) up to 20C.
- Excellent high-temperature performance: the cell remains safe even at 160 °C internal temperature.
- Outstanding safety: no combustion or explosion even under internal or external damage.
- Long cycle life: more than 95% capacity retention after 500 cycles.
- Fast charging ability: up to 2C current.
- Smaller volume and weight compared to lead-acid batteries.
However, LiFePO₄ has a disadvantage in low-temperature performance. At 0 °C, the discharge capacity drops to 78% of the capacity at 23 °C, and at -20 °C it drops to 65%. Therefore, any EV battery pack management system must incorporate temperature compensation and thermal protection.
3. EV Battery Pack Management System
The battery management system (BMS) is the core of the EV battery pack. Its basic structure is shown in Figure 1. The system is divided into two main sections: the power section (charger) and the control/protection section. The power section includes the switch-mode power supply with power factor correction and DC/DC conversion. The control section uses a microcontroller to monitor voltage, current, temperature, and remaining capacity, and to execute protection actions.
Figure 1. Conceptual block diagram of an EV battery pack management system.
The main functions of a complete BMS for an EV battery pack include:
- State monitoring: continuously measure each cell voltage, pack voltage, current, temperature, and internal resistance.
- State-of-charge estimation: calculate the remaining capacity to inform the user of the remaining driving range.
- Cell balancing: compensate for differences between series-connected cells to maximize usable capacity.
- Thermal management: activate cooling or heating devices to keep the battery within its optimal temperature range.
- Protection: disconnect the battery under over-charge, over-discharge, over-current, short-circuit, and over-temperature conditions.
- Communication: exchange data with the charger, the vehicle controller, or a PC through a serial bus.
The protection of the EV battery pack is crucial because the consequences of a fault can be severe, including fire or explosion. My design incorporates both analog hardware protection and software-based protection to provide redundant safety layers.
4. Charging Circuit Design
The charging part of my design is an AC/DC switch-mode power supply that converts the 220 V mains input into a regulated DC output suitable for the EV battery pack. Due to the power level (up to 550 W), European and international standards require a power factor correction (PFC) stage to reduce harmonic pollution. I selected a two-stage architecture: a boost PFC front-end followed by an isolated DC/DC converter. The system block diagram is shown earlier in Figure 1.
4.1 Switching Power Supply Types
There are many topologies for switch-mode power supplies. The choice depends on output power, isolation requirements, cost, and efficiency. Table 3 summarizes the main isolated and non-isolated topologies.
| Topology | Isolation | Advantages | Disadvantages | Output Power Range (W) |
|---|---|---|---|---|
| Buck | No | Simple, high efficiency | Not isolated | <50 |
| Boost | No | Simple, high efficiency | Not isolated | <30 |
| Buck-Boost | No | Simple, high efficiency | Not isolated | <30 |
| Single-ended forward | Yes | Reliable, simple drive | Transformer unidirectional, low utilization | 50–200 |
| Single-ended flyback | Yes | Simple, low component count | High switch voltage stress | 20–100 |
| Push-pull | Yes | Bidirectional core excitation, high utilization | Flux imbalance, saturation risk | 100–500 |
| Half-bridge | Yes | Low switch voltage stress | Floating drive required | 100–700 |
| Full-bridge | Yes | High power density | Complex drive and high component count | 500–2000 |
For my design, the output power is 550 W. I chose a parallel two-transistor forward converter (also called a double-ended forward converter), which can be seen as two single-ended forward converters operating interleaved. This topology provides twice the output power of a single forward converter, reduces the ripple frequency, and simplifies the drive circuit compared to a full-bridge. The control is implemented with a PWM controller that supports both PFC and DC/DC switching signals.
4.2 PFC Circuit Design
I selected the boost PFC topology operating in continuous conduction mode (CCM) with average current-mode control. This method is known for low total harmonic distortion, good noise immunity, and stable operation. The circuit diagram of the average current-mode boost PFC is shown in a previous figure (not reproduced here due to space constraints). The key equations are given below.
The maximum DC voltage after rectification is:
$$V_{\text{DC},\max} = \sqrt{2} \cdot V_{\text{AC},\max} \approx 375 \text{ V}$$
The maximum input current is estimated as:
$$I_{\text{DC},\max} = \frac{2 \cdot P_{\text{OUT},\max}}{\eta \cdot U_{\text{AC},\min}} \approx 7.04 \text{ A}$$
The maximum duty cycle at minimum AC input voltage is:
$$D_{\max} = 1 – \frac{\sqrt{2} \cdot V_{\text{AC},\min}}{V_{C}} \approx 0.52$$
The boost inductor value is calculated from the desired current ripple coefficient \(K = 0.2\):
$$L = \frac{V_{\text{AC},\min} \cdot D_{\max}}{\eta \cdot K \cdot f_{\text{PFC}} \cdot P_{\text{OUT},\max}} \approx 2.44 \text{ mH}$$
With an AL-value of 200 nH/N² and a gap of 0.8 mm, the number of turns is:
$$N = \sqrt{\frac{L}{AL}} = \sqrt{\frac{2.44 \times 10^{-3}}{200 \times 10^{-9}}} \approx 110 \text{ turns}$$
The PFC output capacitor is designed for a hold-up time of 60 ms:
$$C = \frac{2 \cdot P_{\text{OUT}} \cdot t_{HU}}{V_{C}^{2} – V_{\text{min}}^{2}} \approx 580\ \mu\text{F}$$
I selected a 600 μF/400 V electrolytic capacitor with a parallel 0.22 μF/400 V film capacitor for high-frequency decoupling. The PFC switch is a MOSFET (2SK1249, 500 V, 15 A) and the boost diode is a fast-recovery type (MUR860, 600 V, 8 A). The bridge rectifier is KBPC1010 (1000 V, 10 A).
4.3 DC/DC Converter Design
The isolated DC/DC converter provides galvanic isolation and converts the 380 V DC bus to the required battery charging voltage. My chosen topology is the parallel two-transistor forward converter. The switching frequency is set to 200 kHz. The duty cycle is limited to 44% to keep the transformer flux reset safe.
For the transformer design, I calculated the minimum secondary voltage:
$$U_{s,\min} = \frac{(U_{o,\max} + U_{L} + U_{F}) \cdot T}{t_{on,\max}} = \frac{(55 + 0.2 + 0.5) \cdot 5}{2.2} \approx 126.6 \text{ V}$$
The turns ratio is:
$$N = \frac{U_{s,\min}}{U_{dc,\min}} = \frac{126.6}{200} \approx 0.633$$
Using an EI-40 core with \(B_m = 0.2 \text{ T}\) and effective area \(S = 148 \text{ mm}^2\):
$$N_s = \frac{U_{s,\min} \cdot t_{on,\max}}{B_m \cdot S} \times 10^{4} = \frac{126.6 \times 2.2 \times 10^{-6}}{0.2 \times 148 \times 10^{-6}} \times 10^{4} \approx 9.4$$
I chose \(N_s = 10\) turns. The primary turns are:
$$N_p = \frac{N_s}{N} = \frac{10}{0.633} \approx 15.79 \text{ turns}$$
I chose \(N_p = 16\) turns. The primary current is:
$$I_{DS} = \frac{N_s}{N_p} \cdot I_o = \frac{10}{16} \times 10 \approx 6.25 \text{ A}$$
The RMS current in the primary winding is:
$$I_p = I_{DS} \cdot \sqrt{D_{\max}} = 6.25 \times \sqrt{0.44} \approx 4.15 \text{ A}$$
The secondary current RMS is:
$$I_s = I_p \cdot \frac{N_p}{N_s} = 4.15 \times \frac{16}{10} \approx 6.64 \text{ A}$$
For high-frequency operation, I selected Litz wires to reduce skin effect. The output inductor is calculated with a 1 A current ripple:
$$L_o = \frac{(U_{s,\min} – U_{o,\max}) \cdot t_{on}}{\Delta I_L} = \frac{(126.6 – 55) \times 2.2 \times 10^{-6}}{1} \approx 71\ \mu\text{H}$$
The output inductor has 19 turns. The output filter capacitor is:
$$C_o = \frac{U_s \cdot T}{8 \cdot L \cdot \Delta U_o} = \frac{126.6 \times 5 \times 10^{-6}}{8 \times 71 \times 10^{-6} \times 0.01} \approx 111.4\ \mu\text{F}$$
4.4 Feedback Control Design
For the PFC stage, I designed the voltage and current loops. The voltage loop must attenuate the 100 Hz ripple. The current loop must have high bandwidth to shape the input current. The key parameters are summarized below.
| Component | Value |
|---|---|
| \(C_f\) (voltage loop) | 110 nF |
| \(R_f\) (voltage loop) | 16.7 kΩ |
| \(C_z\) (voltage loop) | 1.1 μF |
| \(R_{IN}\) | 2.2 MΩ |
| \(R_{I}\) (current loop) | 2 kΩ |
| \(R_f\) (current loop) | 27 kΩ |
| \(C_p\) (current loop) | 100 pF |
| \(C_z\) (current loop) | 470 nF |
For the DC/DC converter, I implemented a current-mode control with an outer voltage loop. To achieve constant-current (CC) and constant-voltage (CV) charging, two error amplifiers are used. Their outputs are diode-ORed to the PWM controller. The CC loop limits the charging current, while the CV loop limits the charging voltage. When the battery voltage is low, the CC loop dominates; after the battery voltage reaches the setpoint, the CV loop takes over. Both setpoints are provided by the microcontroller through a D/A converter.
4.5 EMC Design
Electromagnetic compatibility is often an afterthought in power supply design, but I incorporated it from the beginning. The main EMI sources are the switching action of the MOSFETs, the high-frequency transformer leakage inductance, and the rectifier diodes.
The input filter consists of both differential-mode and common-mode chokes. The power filter uses a common-mode choke \(L_2\) and two Y capacitors \(C_2\) of 0.22–0.47 μF. The EMI filter also includes X capacitors and Y capacitors to attenuate conducted emissions both from and to the AC mains.
For the high-frequency transformer, I reduced leakage inductance by careful winding arrangement, using multi-filar wires, and selecting a core shape with a low height-to-width ratio. The transformer was also shielded to prevent magnetic field coupling to nearby circuits.
For the power switches, I added RC snubbers to absorb the voltage overshoot caused by the transformer leakage inductance. The snubber capacitor is charged through a fast diode during the switch off-time, and the resistor dissipates the stored energy during the on-time.
The PCB layout was designed with the following guidelines:
- Minimize loop areas of high dv/dt and high di/dt paths.
- Separate power ground and signal ground.
- Use ground planes to reduce impedance.
- Place decoupling capacitors close to the IC pins.
- Keep sensitive analog circuits away from switching nodes.
4.6 Simulation Results
I simulated the complete charging circuit using PSpice. The PFC switch drive waveform, inductor current, inductor voltage, and output voltage were observed. The PFC output voltage settled at 380 V with less than 2% ripple. The DC/DC output provided a stable 55 V supply. The simulations confirmed that the design meets the requirements for charging the EV battery pack.
5. Protection and Control Circuit Design
5.1 Hardware Design
The control unit is based on the C8051F020 microcontroller, a fully integrated mixed-signal System-on-Chip from Silicon Laboratories. It integrates:
- 12-bit ADC with multiple input channels and internal temperature sensor
- 8-bit ADC
- 12-bit DAC
- Analog comparators
- Voltage reference
- 16-bit timer array for PWM generation
- Multiple serial buses: SMBus/I²C, SPI, UART
- 64 kB flash memory and 4 kB RAM
- On-chip debug via JTAG
The C8051F020 is ideal for an EV battery pack management system because it can handle all the monitoring and control tasks without external analog circuitry.
5.1.1 Voltage Detection
The battery pack voltage is divided by a resistor network and filtered before being fed to the ADC. The protection thresholds are:
| Parameter | Value |
|---|---|
| Over-charge voltage | 54.5 V ± 0.1 V |
| Over-charge release voltage | 53.5 V ± 0.1 V |
| Over-charge delay | 2 s |
| Over-discharge voltage | 31.5 V ± 0.3 V |
| Over-discharge release voltage | 37.5 V |
| Over-discharge delay | 500 ms |
| Short-circuit voltage | 5 V |
| Short-circuit delay | 10 μs |
| CC/CV transition voltage | 52.5 V |
5.1.2 Current Detection
I used a Hall-effect current sensor based on the magnetic balance principle. This sensor provides galvanic isolation, high accuracy, and fast response. The sensor output voltage is proportional to the battery current. The charging current is sensed separately from the discharge current due to direction. The protection current thresholds are 20 A for both charging and discharging. The over-current delay is 50 ms. To detect full charge, the charger monitors the current during CV mode; when the current drops below 200 mA for 3 seconds, the charge is terminated.
5.1.3 Internal Resistance Detection
Battery internal resistance is correlated with capacity and health. I employed an AC injection method with a 1 kHz, 50 mA signal to estimate the internal resistance. The measurement is performed periodically when the battery is idle. Although less accurate than the DC method, it does not damage the battery and is acceptable for the EV battery pack health assessment.
5.1.4 Temperature Detection
Temperature is measured with an integrated sensor (AD590) placed on the battery pack. The ADC reads the temperature and the microcontroller implements thermal protection. If the temperature exceeds a safe limit, the charging or discharging MOSFETs are turned off. The system also compensates the SOC estimation based on the temperature data.
5.1.5 Cell Voltage Detection
For the 15 series cells, I used a voltage-controlled current source circuit to measure each individual cell voltage. The circuit converts the differential cell voltage into a current, which then flows through a resistor to produce a voltage proportional to the cell voltage. This technique rejects common-mode voltages and simplifies the interface to the microcontroller’s ADC.
5.2 Software Design
5.2.1 State-of-Charge Estimation
Accurate SOC estimation is essential for an EV battery pack. The SOC indicates the remaining capacity as a percentage of full charge. I adopted an enhanced ampere-hour counting method with several compensation factors for temperature, discharge rate, aging, and self-discharge. The fundamental equation for accumulated capacity is:
$$Q_{use} = \int_{0}^{t} i(t) \, dt$$
In discrete form:
$$Q_{use} = \sum_{k=1}^{n} i_k \Delta t$$
The uncompensated SOC is:
$$SOC = \frac{Q_E – Q_{use}}{Q_E} \times 100\%$$
To compensate for charging efficiency and self-discharge, I introduce a loss current \(i_g\):
$$i_g = \eta \cdot U + \mu \cdot (T – T_0)$$
where \(\eta\) and \(\mu\) are empirical constants, \(U\) is the battery voltage, \(T\) is the temperature, and \(T_0\) is the reference temperature (20 °C). The effective current is \(i_f = i_t – i_g\) during charging and \(i_f = i_t + i_g\) during discharging.
Temperature compensation is applied directly to the SOC value:
$$SOC_{T} = SOC \times \left[1 – \nu \cdot (T – T_0)\right]$$
where \(\nu\) is the temperature coefficient. Additionally, an aging factor \(A_F\) is defined as:
$$A_F = \frac{Ah_{ref} – Ah_{pre}}{Ah_{ref}}$$
where \(Ah_{ref}\) is the maximum capacity over the battery’s lifetime, and \(Ah_{pre}\) is the current capacity. The Peukert equation accounts for the discharge rate effect:
$$I^n \cdot t = K$$
where \(n\) and \(K\) are constants determined experimentally. For variable discharge currents, I discretize the current profile and compute the effective discharged capacity:
$$SOC = 1 – \sum_{j} \frac{I_j \Delta t_j}{Q_{E,j}}$$
where \(Q_{E,j}\) is the capacity available at discharge current \(I_j\) at 20 °C. Combining all compensation factors yields:
$$SOC = \left[1 – \sum_{j} \frac{(I_j + \eta U + \mu(T-T_0)) \Delta t_j}{K I_j^{1-n}}\right] \times \left[1 – A_F\right] \times \left[1 – \nu(T-T_0)\right]$$
To prevent error accumulation, I periodically correct the SOC using the open-circuit voltage (OCV) method when the battery has been at rest for a sufficiently long time. The OCV-SOC relationship is stored in a lookup table, and the system replaces the calculated SOC with the OCV-based value if the deviation exceeds a threshold.
5.2.2 SMBus Communication
I implemented the System Management Bus (SMBus) protocol for communication between the EV battery pack protection circuit and external devices such as the charger, the vehicle controller, or a PC. SMBus is a two-wire interface based on I²C, with a maximum data rate of 100 kbps. It uses SDA and SCL lines, both open-drain, with pull-up resistors. The bus protocol includes start/stop conditions, 7-bit addressing, read/write bits, acknowledgments, and arbitration.
A typical SMBus transmission consists of:
- START condition: SDA falls while SCL is high.
- Address byte: 7-bit slave address followed by a read/write bit.
- ACK: The receiver pulls SDA low on the 9th clock.
- Data bytes: up to 32 bytes, each followed by an ACK/NACK.
- STOP condition: SDA rises while SCL is high.
Arbitration is performed on the SDA line: if two masters start simultaneously, the one that sends a high bit while the other sends a low bit loses the bus and becomes a slave. The protocol defines timeouts: any SCL low period exceeding 25 ms triggers a communication reset in all devices.
SMBus provides a simple, low-pin-count solution for the EV battery pack to report voltage, current, SOC, temperature, and fault status. It can be easily bridged to USB or RS-232 via commercial converters.
6. Conclusion and Outlook
In this paper, I have presented a complete design of a charging and protection circuit for a 48 V, 20 Ah LiFePO₄ EV battery pack. The charging part uses a boost PFC stage and a parallel two-transistor forward DC/DC converter to achieve high power factor, high efficiency, and stable output. The protection and control part is based on a C8051F020 microcontroller, which performs voltage, current, temperature, and internal resistance monitoring, as well as over-charge, over-discharge, over-current, short-circuit, and thermal protection. I also proposed a practical SOC estimation method combining compensated ampere-hour counting with OCV-based correction. The SMBus interface enables communication with external devices.
Some aspects can be further improved in future work:
- Fine-tuning the compensation parameters through more extensive battery testing.
- Adding a cooling fan and thermal management system to the charger to improve reliability.
- Implementing soft-switching techniques in the DC/DC converter to reduce switching losses and EMI.
- Developing intelligent cell balancing algorithms to extend the EV battery pack cycle life.
- Integrating a wireless communication module for remote monitoring.
With the rapid development of battery technology, the EV battery pack will become more energy-dense, lighter, smaller, and safer. Correspondingly, the charging and protection electronics must evolve to meet the demanding requirements of future electric vehicles. My work provides a solid foundation for such an evolution.