A Cost-Effective Passive/Active Hybrid Equalizer Circuit Design

Abstract

This paper proposes a novel hybrid equalizer circuit (HEC) for a battery management system (BMS) to implement the passive HEC (P-HEC), active HEC (A-HEC), or active/passive (AP-HEC) with the same equalizer circuit architecture. The advantages of an HEC are that it is simple, cost-effective, highly energy efficient, and fail safe. The P-HEC can further use a cooling fan or heater instead of a conventional resistor as a power dissipation element to convert the energy of the waste heat generated by the resistor to adjust the battery temperature. Even if the P-HEC uses the resistor to consume energy as in conventional methods, the P-HEC still dramatically improves the component lifetime and reliability of the BMS because the waste heat generated by the equalizer resistor is outside of the BMS board. Three significant advantages of an A-HEC are its (1) low cost, (2) small volume, and (3) higher energy efficiency than the conventional active equalizer circuits (AECs). In the HEC design, the MOSFETs of the switch array do not need high-speed switching to transfer energy as conventional AECs with DC/DC converter architecture because the A-HEC uses an isolated battery charger to charge the string cell. Therefore, the switch array is equal to a cell selector with a simple ON/OFF function. In summary, the HEC provides a small volume, cost-effective, high efficiency, and fail-safe equalizer circuit design to satisfy cell balancing demands for all kinds of electric vehicles (EVs) and energy storage systems (ESSs).

1. Introduction

Due to growing concerns about the environmental impact of fossil fuels, policymakers are increasingly turning their attention to clean energy and electric vehicles [1,2,3,4,5]. Therefore, the energy storage system (ESS) plays a critical role, and the lithium-ion battery (LIB) shares more than 90% of the ESS market. For these ESSs, they need to consist of many LIBs grouped in series and in parallel to assemble a battery pack to provide sufficient power and the desired energy. In practice, the difference in each cell’s characteristics (e.g., internal capacity, impedance, self-discharge rate, etc.) will influence the voltage difference of all the series-connected cells of the battery pack. Therefore, a battery management system (BMS) [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21] was designed to protect, monitor, and control the state of all cells of the battery pack. A good BMS can ensure safe operation, maximize the available capacity, and provide a real-time estimate of the remaining discharge capacity of the battery pack [19].

However, much effort and resources are required to test a BMS with a real ESS. Furthermore, by pushing the ESS operation under extreme operating conditions, it is almost impossible to verify that the BMS functions are safe and reliable. Therefore, some research uses a hardware-in-the-loop (HIL) simulation tool to test BMS. The HIL can be used to simulate string cells of a battery pack [20] to test BMS functions for cells, e.g., test passive equalizer function [21]. Briefly, HIL is a good auxiliary tool in the development stage of a BMS, which can significantly reduce the development time and testing risk of the BMS. In many practical experiences, most battery pack failures are caused by the cell voltage imbalance issue, which means the voltage difference is enormous, drastically reducing the battery pack’s available power, capacity, lifespan, and safety. Therefore, most BMS designs have an equalization unit to solve the imbalance issue [22,23,24,25,26,27,28,29,30,31,32,33,34,35].

In general, the BMS is composed of measurement, communication, calculation, memory, equalization units, etc. The equalization unit is also called the cell balancing unit. The equalization unit is composed of two parts: an equalization control strategy [23,24,25,26,27,28,29] and an equalization circuit (EC) [28,29,31,32,33,34]. Its purpose is to reduce the maximum and minimum cell voltage differences of the battery pack to avoid battery overcharge and overdischarge, which can improve the safety and lifespan. Besides, the low cell voltage difference also increases the available capacity of the battery pack. Therefore, the equalization unit is the most crucial part of all BMS units. Retired vehicle lithium battery packs with good cell balance can be reused in energy storage systems for secondary use to produce huge economic and environmental benefits. Various ECs have been proposed in the past, which can be divided into two types: (1) the passive equalizer circuit (PEC) [28], also called dissipative balance, discharges the energy with a power resistor to the series-connected cell with the highest voltage, and (2) the active equalizer circuit (AEC) [27,28,29,30,31,32,33,34,36,37,38], also called non-dissipative balance, charges the energy to the cell with the lowest voltage. Many review articles have provided a detailed summary or comparison for the equalization design [17,26,28,30,32,33,36]. Therefore, the following summary is more focused on commercialization.

The commonly used PEC architecture [28] is shown in Figure 1. The advantages of PEC are that it is very simple, easy to apply and low cost, and all switches (${\mathrm{S}}_{\mathrm{i}}$) and discharge resistors (${\mathrm{R}}_{\mathrm{i}}$) are placed on the BMS circuit board (where i = 1, 2, …, n). Therefore, the temperature of BMS board will increase a great deal when performing the cell balancing. In most commercial products, the typically balancing resistor is 33 Ω [39], and the current of each channel is around 100~127 mA which maps to 3.3~4.2 V and the waste heat of the PEC is around 0.3~0.5 W per channel in the BMS board. To avoid overheating the BMS board, the BMS also limits the number of balanced channels.

Briefly, the AEC uses the switch array, capacitors, inductors, or transformers to absorb the energy from the cell with the highest voltage and release the energy to the cell with the lowest voltage. Many kinds of AECs are shown in Figure 2, but most are based on the DC/DC converter principle.

Figure 3 shows the fundamental topology of an AEC with the DC/DC converter working principle, and all switches are the bidirectional switch (BS) which consists of a back-to-back MOSFET string, called BS-MOSFET. Obviously, all switches in Figure 3 need high-speed switching to transfer energy between the string cell and the transformer.

Based on the topology of Figure 3, the chipmaker analog device had commercialized the active balancing control IC LT8584 [40]. The AFE IC (LTC680x family) can drive the active balancing IC LT8584 and the transformer NA5743-AL to generate 2.5 A balancing current from the cell to the module. Moreover, the standalone balancing IC (LT3300 family) enables a bidirectional balancing current between cell and module up to 10 A.

Figure 4 shows another implementation of the AEC with a single switched capacitor (SSC) with DC/DC converter principle, and the switches in Figure 4 also need high-speed switching to transfer energy between the auxiliary battery, capacitor, and string cells.

In fact, the working principle of most AECs is based on a DC/DC converter architecture to transfer energy between string cells or the module, and the energy storage components (e.g., capacitor or inductor) can be regarded as an energy transfer buffer. Therefore, the switches of the switch array of the AECs are not only used to select the string cell for sinking/draining energy but also needed high-speed switching for transferring energy through the energy buffer (e.g., inductor or capacitor). Due to high-speed switching requirements, most of the switching arrays composed of BS-MOSFETs in AEC circuits are mostly driven by gate drivers. Based on the above descriptions, it is easy to find the cost of AECs is related to the component’s specification and the number of power resistors, inductors, capacitors, transformers, and BSs.

Thus, the low-cost or the low-series cell string pack prefers the PEC. The AEC is more applicable to high-end or high-series cell string packs. Therefore, this paper proposes the HEC architecture, which can be both PEC and AEC or either PEC or AEC. The advantages of the proposed HEC are its (1) cost-effective, (2) high efficiency, and (3) fail-safe features.

2. Operation Principle of Hybrid Equalizer

In Figure 5, the HEC corresponds to a PEC by replacing the discharge unit with a power resistor, which is called a P-HEC. All the switches in Figure 5 are replaced with BS-MOSFET, as shown in Figure 6.

Figure 7 is an example to illustrate the principle of P-HEC. To discharge the B₂ cell through the EQ-Bus, the BMS controls the ${\mathrm{S}}_{2,\mathrm{n}}$ and ${\mathrm{S}}_{2,\mathrm{p}}$ switches to be conducting. A discharge unit ${\mathrm{R}}_{\mathrm{EQ}}$ can be a fan or a heater to regulate the battery temperature or be a power resistor to dissipate the cell energy as conventional ones. The ${\mathrm{R}}_{\mathrm{EQ}}$ is installed out of the BMS board with proper cooling to prevent the BMS board temperature rising quickly when executing the balance operation. The cell balancing current (${\mathrm{I}}_{\mathrm{b}}$) is related to the resistance of ${\mathrm{R}}_{\mathrm{EQ}}$, conducting resistance of ${\mathrm{S}}_{2,\mathrm{n}}$ and ${\mathrm{S}}_{2,\mathrm{p}}$, and the resistance of all connected parts from B₂₊ to B₂₋. For example, the ${\mathrm{R}}_{\mathrm{EQ}}$ is 1 Ω and the resistance of the BSs and wires is 0.5 Ω so that the total resistance is 1.5 Ω. The balancing current is 2.27 A (${\mathrm{V}}_{\mathrm{B}2}$ = 3.4 V) to 2.8 A (${\mathrm{V}}_{\mathrm{B}2}$ = 4.2 V) within cell operation voltage range. The fuse and switch specification will be calculated after given ${\mathrm{R}}_{\mathrm{EQ}}$. When the HEC is only used as PEC, it allows the opposite electrode of EQ-Bus. Therefore, the switches ${\mathrm{S}}_{\mathrm{i},\mathrm{n}}$ and ${\mathrm{S}}_{\mathrm{i},\mathrm{p}}$ can be replaced with a single switch ${\mathrm{S}}_{\mathrm{i}}$ to save costs.

Figure 8 shows how the proposed HEC design can satisfy the fail-safe function when any switch of the switch array fails in short-circuit. Figure 8 assumes the ${\mathrm{S}}_{1,\mathrm{p}}$ incurs a short-circuit failure, and a short current (${\mathrm{I}}_{\mathrm{s}}$) blows the fuse F₂ or F₃ when discharging the B₂ cell by conducting ${\mathrm{S}}_{2,\mathrm{n}}$ and ${\mathrm{S}}_{2,\mathrm{p}}$. The current path of the short current (${\mathrm{I}}_{\mathrm{s}}$) goes through ${\mathrm{F}}_3\to {\mathrm{S}}_{2,\mathrm{p}}\to \mathrm{EQ}-\mathrm{Bus}+\to {\mathrm{S}}_{1,\mathrm{p}}\to {\mathrm{F}}_2$. Therefore, the HEC can avoid the cell being continuously discharged when the failed switch keeps conducting. In addition, it is easy to identify the failure switch path by detecting the balancing current or voltage change.

Figure 9 shows an HEC implement AEC architecture by replacing the charge unit of Figure 5 to an isolated CC-CV battery charger, which is called an A-HEC. The isolated CC-CV battery charger is composed of an isolated DC/DC converter and a CC-CV battery charger circuit. The switch ${\mathrm{S}}_{\mathrm{B}}$ controls the input power, and the ${\mathrm{S}}_{\mathrm{A}}$ controls the charger output power to the EQ-Bus.

Figure 10 shows the HEC architecture can be either PEC or AEC, but only one mode can be used at one time. It can be P-HEC by conducting the switch ${\mathrm{S}}_{\mathrm{C}}$ and be A-HEC by conducting the switch ${\mathrm{S}}_{\mathrm{A}}$ and ${\mathrm{S}}_{\mathrm{B}}$. The switches ${\mathrm{S}}_{\mathrm{A}}{-\mathrm{S}}_{\mathrm{C}}$ must be isolated switches because of the different ground levels.

In summary, the proposed HEC combines three essential parts: (1) a switch array, (2) fuses, (3) the charging unit, or the discharging unit. The proposed HEC uses the switch array architecture of conventional AECs to implement the PEC and AEC. The P-HEC can solve the thermal issue of conventional PECs by placing the discharge unit out of the BMS board. The discharge unit might not only be a power resistor for dissipating the energy but also could be a fan or heater to reuse the energy to regulate the battery temperature. Furthermore, the balancing current (>2 A) of the P-HEC is more than 20 times that of the conventional PEC (10–127 mA) with a power resistor on the BMS board. The proposed HEC also adopts the auxiliary lead–acid battery of the EV to provide energy to the isolated DC/DC charger to charge the cell with the lowest voltage in the battery pack. The isolated charger is composed of an isolated DC/DC converter and a CC-CV charger circuit. The significant advantages are shown below: (1) simple and easy to implement with low-cost components, (2) its volume is much smaller. Whether the proposed HEC is a P-HEC or an A-HEC, the proposed HEC also satisfies the fail-safe requirement of functional safety because the balancing current will be cut off by blowing a fuse when any switch of the switch array fails in short-circuit state. The BMS also can detect the cell voltage or balancing current to identify the failed switch.

3. Cost-Effective Bidirectional Switch Circuit Design for HEC

The switch array of the HEC is very similar to the traditional AEC, but the basic control concept is entirely different even though the BSs of the switch array use the same BS-MOSFET structure. In conventional AECs based on the DC/DC converter principle, the BS-MOSFET in many AEC articles needs a gate driver [29] or an isolation optocoupler (TLP250) [34] to high-speed switch the BS-MOSFET to transfer the energy between the cell and the energy storage components. The disadvantage is the number of gate drivers dramatically increases the cost of the switch array. However, the BS-MOSFETs of the HEC only use an ON/OFF switch, which is equal to a string cell selector for performing cell balancing. Therefore, Section 3 proposes a low-cost control circuit to replace the gate driver to control the BS-MOSFETs.

3.1. Low-Cost BS-MOSFET Drive Circuit

Based on the working principle of Figure 7, we further explain how to design the low-cost BS-MOSFET switch circuit. Therefore, the BS-MOSFET control circuit and the working principle are shown in Figure 11.

In Figure 11, we use two back-to-back P-MOSFETs to be the BS-P-MOSFET drive circuit if the minimum voltage of the node is $\ge$8 V. Otherwise, we use two back-to-back N-MOSFETs to be the BS-N-MOSFET drive circuit for the maximum voltage of the node is <8 V and the battery pack voltage B₊ is $\ge$8 V.

3.2. BS-P-MOSFET Drive Circuit

Figure 11 can be simplified as Figure 12, where only two cells are series connected. The B₁ in Figure 12 corresponds to the battery consisting of B₁ to B_n−1 of Figure 11. The B₂ in Figure 12 corresponds to the B_n of Figure 11. We use Figure 12 to demonstrate the principle of a BS-P-MOSFET driving circuit with a B₂₊ path to the EQ-Bus+ because the circuit of the B₂₋ path to the EQ-Bus- is the same structure. Therefore, we only use B₂₊ to demonstrate the operation principle.

3.2.1. Schematic and Calculation for BS-P-MOSFET Circuit

In Figure 12, the principle of step 1 to step 3 to switch on the BS-P-MOSFET path is shown below.

Step 1: Switch on the transistor Q₁ by giving control voltage ${\mathrm{V}}_{\mathrm{B}2\mathrm{ctrl}}$.

$${\mathrm{V}}_{\mathrm{R}3}={\mathrm{V}}_{\mathrm{B}2\mathrm{ctrl}}\times \left(\mathrm{R}3/\left(\mathrm{R}3+\mathrm{R}4\right)\right),$$

(1)

$${\mathrm{I}}_1=\left({\mathrm{V}}_{\mathrm{B}2\mathrm{ctrl}}-{\mathrm{V}}_{\mathrm{BE}\left(\mathrm{Q}1\right)}\right)/\mathrm{R}4,$$

(2)

where ${\mathrm{V}}_{\mathrm{R}3}>1.2\times {\mathrm{V}}_{\mathrm{BE}\left(\mathrm{ON}\right)}$ is recommended before switch on Q₁ and the current of R₃ can be ignored after conducting Q₁; $50\mathsf{\mu }\mathrm{A}<{\mathrm{I}}_1<100\mathsf{\mu }\mathrm{A}$ is recommended and C₁ is parallel with R₃ to avoid malfunction.

Step 2: Switch on the P-MOSFETs Q₂ and Q₃ after Q₁ is switched on and the current I₁ drives ${\mathrm{I}}_2$ and${\mathrm{V}}_{\mathrm{R}1}$.

$${\mathrm{I}}_2=\left({\mathrm{V}}_{\mathrm{B}2+}-{\mathrm{V}}_{\mathrm{CE}\left(\mathrm{Q}1\right)}-{\mathrm{V}}_{\mathrm{DS}\left(\mathrm{Q}2\right)}\right)/\left(\mathrm{R}1+\mathrm{R}2\right),$$

(3)

$${\mathrm{V}}_{\mathrm{R}1}={\mathrm{I}}_2\times \mathrm{R}1,$$

(4)

where ${\mathrm{V}}_{\mathrm{R}1}>{\mathrm{V}}_{\mathrm{gs}\left(\mathrm{th}\right)}$, $8\mathrm{V}<{\mathrm{V}}_{\mathrm{R}1}<12\mathrm{V}$ is recommended for switch on, and ${\mathrm{V}}_{\mathrm{R}1}<0.2\mathrm{V}$ for switch off.

Step 3: Finish to conduct BS-P-MOSFET path to EQ-BUS.

3.2.2. Implementation and Test Results for BS-P-MOSFET Circuit

The transistor Q₁ uses the BC846, and the Q₂ and Q₃ use a dual P-MOSFET (MTB60B06Q8) in SOP8 package for size reduction, whose basic specification is shown in Figure 13 and Figure 14 shows the implementation of the BS-P-MOSFET drive circuit.

Based on the operation principle of Figure 12, the settings are defined as B₁ = 42 V (corresponding to the B₁~B_n-1 string of the battery pack) and B₂ = 4 V (corresponding to the B_n). Other settings are R₁ = 20 kΩ, C₁ = 10 nF, R₂ = 75 kΩ, R₃ = 50 kΩ, and R₄ = 50 kΩ. We use the constant resistor mode of electrical load to simulate ${\mathrm{R}}_{\mathrm{EQ}}$ for obtaining ${\mathrm{I}}_{\mathrm{EQ}}\cong 1\mathrm{A}$. Therefore, the calculated result is ${\mathrm{V}}_{\mathrm{R}3}=1.65\mathrm{V}({\mathrm{V}}_{\mathrm{B}2\mathrm{ctrl}}=3.3 \mathrm{V})$, ${\mathrm{I}}_1\cong 52\mathsf{\mu }\mathrm{A}$, and the voltage range of ${\mathrm{V}}_{\mathrm{R}1}$ and ${\mathrm{V}}_{\mathrm{R}1\text{’}}$ is 9.2 V (B+ = 46 V) and 8.4 V (B+ = 42 V), respectively. Figure 15 and Figure 16 show the signal of ${\mathrm{V}}_{\mathrm{B}2\mathrm{ctrl}}$, ${\mathrm{V}}_{\mathrm{R}1}$, and ${\mathrm{V}}_{\mathrm{EQ}}$ to exhibit the transient state of the switch on and off.

3.3. BS-N-MOSFET Drive Circuit

Figure 11 can be simplified as Figure 17, where only two cells are series connected. The B₁ in Figure 12 corresponds to the B₁ cell of Figure 11. The B₂ in Figure 17 corresponds of B₂ to B_n of Figure 11. We use Figure 17 to demonstrate the principle of BS-N-MOSFET driving circuit with B₁₊ path to EQ-Bus+ because the circuit of B₁₋ path to EQ-Bus- is the same structure. Therefore, we only use B₁₊ to demonstrate the operation principle.

3.3.1. Schematic and Calculation for BS-N-MOSFET Circuit

In Figure 17, the principle of step 1 to step 4 for conducting the BS-N-MOSFET path is shown below:

Step 1: Conduct the transistor Q_d by giving control voltage ${\mathrm{V}}_{\mathrm{B}1\mathrm{ctrl}}$.

$${\mathrm{V}}_{\mathrm{Re}}={\mathrm{V}}_{\mathrm{B}1\mathrm{ctrl}}\times \left(\mathrm{Re}/\left(\mathrm{Re}+\mathrm{Rf}\right)\right),$$

(5)

$${\mathrm{I}}_{\mathrm{a}}=\left({\mathrm{V}}_{\mathrm{B}1\mathrm{ctrl}}-{\mathrm{V}}_{\mathrm{BE}\left(\mathrm{Qd}\right)}\right)/\mathrm{Rf},$$

(6)

where ${\mathrm{V}}_{\mathrm{Re}}>1.2\times {\mathrm{V}}_{\mathrm{BE}\left(\mathrm{ON}\right)}$ is recommended before conducting Q_d and the current of R_e can be ignored after conducting Q_d; $50\mathsf{\mu }\mathrm{A}<{\mathrm{I}}_{\mathrm{a}}<100\mathsf{\mu }\mathrm{A}$ is recommended and C_a is parallel with R_e to avoid malfunction.

Step 2: Conduct the transistor Q_c after Q_d is conducted and the current I_a drives;

$${\mathrm{V}}_{\mathrm{Rc}}=\left({\mathrm{V}}_{\mathrm{B}+}-{\mathrm{V}}_{\mathrm{EB}\left(\mathrm{Qc}\right)}\right)\times \left(\mathrm{Rc}/\left(\mathrm{Rc}+\mathrm{Rd}\right)\right),$$

(7)

$${\mathrm{I}}_{\mathrm{b}}=\left({\mathrm{V}}_{\mathrm{B}+}-{\mathrm{V}}_{\mathrm{EB}\left(\mathrm{Qc}\right)}-{\mathrm{V}}_{\mathrm{CE}\left(\mathrm{Qd}\right)}\right)/\mathrm{Rd},$$

(8)

where ${\mathrm{V}}_{\mathrm{Rc}}>1.2\times {\mathrm{V}}_{\mathrm{BE}\left(\mathrm{ON}\right)}$ is recommended before conducting Q_c and the current of R_c can be ignored after conducting Q_c; $50\mathsf{\mu }\mathrm{A}<{\mathrm{I}}_{\mathrm{b}}<100\mathsf{\mu }\mathrm{A}$ is recommended to avoid malfunction.

Step 3: Conduct the N-MOSFETs Q_a and Q_b after Q_c is conducted and the current I_b drives;

$${\mathrm{I}}_{\mathrm{c}}=\left({\mathrm{V}}_{\mathrm{B}+}-{\mathrm{V}}_{\mathrm{B}1}-{\mathrm{V}}_{\mathrm{EC}\left(\mathrm{Qc}\right)}-{\mathrm{V}}_{\mathrm{SD}\left(\mathrm{Qa}\right)}\right)/\left(\mathrm{Ra}+\mathrm{Rb}\right),$$

(9)

$${\mathrm{V}}_{\mathrm{Ra}}={\mathrm{I}}_{\mathrm{c}}\times {\mathrm{R}}_{\mathrm{a}},$$

(10)

where ${\mathrm{V}}_{\mathrm{Ra}}>{\mathrm{V}}_{\mathrm{gs}\left(\mathrm{th}\right)}$, and $8\mathrm{V}<{\mathrm{V}}_{\mathrm{Ra}}<12\mathrm{V}$ is recommended.

Step 4: Finish to conduct BS-N-MOSFET path to EQ-BUS.

3.3.2. Implementation and Test Results for BS-N-MOSFET Circuit

The transistor Q_d uses the BC846 and Qc uses the MMBT5401, and the Q_a and Q_b use a dual N-MOSFET (MTB20A06KQ8) in SOP8 package for size reduction, whose basic specification is shown in Figure 18 and Figure 19 shows the implementation of the BS-N-MOSFET drive circuit.

Based on the operation principle of Figure 17, the settings are defined as B₁ = 4 V (assumed the first string of the battery pack) and B₂ = 42 V (assumed the B₂~B_n string of the battery pack). Other settings are R_a = 110 kΩ, R_b = 390 kΩ, R_c = 50 kΩ, R_d = 680 kΩ, R_e = 50 kΩ, R_f = 50 kΩ, and C_a = 10 nF. We use the constant resistor mode of electrical load to simulate ${\mathrm{R}}_{\mathrm{EQ}}$ for obtaining ${\mathrm{I}}_{\mathrm{EQ}}\cong 1\mathrm{A}$. Therefore, the calculated result is ${\mathrm{V}}_{\mathrm{Re}}=1.65\mathrm{V}({\mathrm{V}}_{\mathrm{B}1\mathrm{ctrl}}=3.3 \mathrm{V})$, ${\mathrm{I}}_{\mathrm{a}}\cong 52\mathsf{\mu }\mathrm{A}$, and the voltage range of ${\mathrm{V}}_{\mathrm{Ra}}$ and ${\mathrm{V}}_{\mathrm{Ra}’}$ are 9.38 V and 9.24 V, respectively. Figure 20 and Figure 21 show the signal of ${\mathrm{V}}_{\mathrm{B}1\mathrm{ctrl}}$, ${\mathrm{V}}_{\mathrm{Ra}}$, and ${\mathrm{V}}_{\mathrm{EQ}}$ to exhibit the transient state of the switch on and off.

The above experimental results have verified the transient characteristics of the BS-MOSFET circuit design of the highest string cell and the lowest string cell in the battery pack, respectively. Briefly, the time of the ON/OFF transient state is around 80–100 us. Therefore, the following section will demonstrate the A-HEC design to the BMS board and test the balancing current for the actual battery pack.

Table 1. Comparison of the proposed P-HEC with conventional PECs (series cells = N).

Item	PEC (Conventional)	P-HEC (Proposed)
Topology
Limitation	---	Pack voltage > 8 V
Architecture	Switch + resistor	Cell selector + Discharge unit
MOSFET numbers/cost	N (Number of series cell) ~0.15 USD/MOSFET	(N + 1) × 2(Dual package) ~0.25 USD/Dual MOSFET
Channel cost	~0.2 USD/channel	~0.35 USD/channel
Max. ON channel	≤2 (Limited by heat)	1
Fail-safe	---	Yes (fuse)
Discharge unit	Power resistor × N (Typical: 33 Ω)	Power resistor × 1 (Typical:2 Ω), Fan or heater
Max. current	~0.127 [email protected] V	~2.1 [email protected] V
Waste heat	0.3–0.5 W/channel	4.5~8 W
Power cost	~0.4 USD/W (0.2 USD/0.5 W)	~0.04 USD/W (0.35 USD/8 W)
Location	On BMS board	Out of BMS board
Occupied area	1× (Including N resistors)	1.5–2× (Only fuse and switch arrays are placed on BMS)

shows the comparison results between the P-HEC and conventional PECs. P-HEC is suitable for a battery pack with many cells connected in series (>8S configuration), such as electric bicycles, electric scooters, or electric vehicles. The conventional PEC is suitable for the battery pack (2S or 3S configuration) of the notebook, power tool, or tablet computer.

Table 2. Comparison of the proposed A-HEC with the conventional AEC (series cells = N).

Item	AEC (Commercial LT8584)	A-HEC (Proposed)
Topology
Architecture	DC/DC converter	Cell selector + isolated charger
Channel cost	N × 6 USD LT8584 (~4) +Transformer (~2) Transformer: COILCRAFT NA5743-AL	(N × 0.25 × 2) + 15 USD Switch array(N × 0.25 × 2 paths) +Charger (~3) +Isolated DC/DC converter (~12) Iso. DC/DC: MEANWELL SCW12B-05
Max. ON channel	Multi-channel	1 channel
Fail-safe	---	Yes (fuse)
Max. current	N × 2.5 A	~2.1 [email protected] V

shows the comparison results between the A-HEC and the commercial active equalizer control IC (LT8584) evolved from Figure 3. The A-HEC is a more cost-effective solution than the LT8584 for the high voltage battery pack (>3S configuration) when the total balancing current of the battery pack/module is less than 3 A. Not only is the channel cost of the HEC low, but also the occupied circuit board area of the equalizer circuit is only 2~3 times of the passive equalizer.

4. Implementation of the A-HEC to the Battery Monitoring Unit of a Battery Pack

Figure 22 shows the implementation of the proposed A-HEC circuit of Figure 9 to the battery monitoring unit (BMU) board, and the maximum balancing charging current can reach 2.2 A. The battery pack has a 22S1P configuration with 25 Ah LTO cells, the cell operating voltage is 1.8~2.7 V and the pack operating voltage is 39.6~59.4 V. The battery packs can be series connected to form a high-voltage battery system in which the system voltage can be up to 600 V.

The circuit placement of the A-HEC circuit of the BMU board of Figure 22 is shown in Figure 23. The BS-MOSFET could be BS-N-MOSFET or BS-P-MOSFET, the BS-N-MOSFET and BS-P-MOSFET circuit are used to B₁–B₃ and B₄–B₂₂, respectively. We used a commercial product SCW12B-05 (12 W, η = 83%) as the isolated DC/DC converter of Figure 9 to transfer auxiliary input 18–36 V to isolated output 5 V. In Figure 9, the CC/CV charger circuit is implemented by the LIB charger IC (EUP3271) of which the maximum charge current is 4 A.

An enlarged view of the BS-MOSFET switch array of Figure 23 is shown in Figure 24, which includes BS-N-MOSFET drive circuit for the B₁~B₂ cells and BS-P-MOSFET for the B₃~B₂₂ cells, respectively.

The balance current experiment results for the proposed A-PEC circuit design of Figure 23 are summarized in

Table 3. The balancing current result for each node of the battery string.

Voltage Node	Balancing Charge Current (mA)	Voltage Node	Balancing Charge Current (mA)
B1+	2265	B12+	2265
B2+	2270	B13+	2270
B3+	2255	B14+	2265
B4+	2260	B15+	2220
B5+	2265	B16+	2270
B6+	2265	B17+	2245
B7+	2270	B18+	2270
B8+	2270	B19+	1960
B9+	2265	B20+	1955
B10+	2260	B21+	1975
B11+	2260	B22+	2260

. The balancing current for each node of the battery string is around 1.95–2.25 A.

5. Conclusions

This paper proposed the cost-effective and simple passive/active HEC design for a BMS. The key idea of the HEC is to let the switch array as a string cell selector for selecting the string cell to charge (A-HEC) or discharge (P-HEC) for cell balancing purposes. Therefore, the BS-MOSFETs of the switch array are equal to an ON/OFF switch and do not need high-speed switching to move energy. Furthermore, we proposed a low-cost and simple control circuit to drive these BS-MOSFET paths of HEC, which had been verified with a 48 V LTO battery pack. The comparison of the proposed HEC with conventional PEC and AEC are shown in Table 1 and Table 2, respectively. The power cost per channel, low-temperature rise on the BMS by placing the discharge unit outside the BMS circuit board, and fail-safe function are the competitive advantages of P-HEC. Similarly, the low cost of the BS-MOSFET driving circuit, fail-safe and small size on BMS are the significant features of the A-HEC. The limitation of the HEC is the minimum pack voltage should be higher than 8 V (>3S~4S configuration) for driving the MOSFET and control circuit. Based on the above features, the equalizer circuit design with the HEC topology is a cost-effective solution for various EVs and ESSs with high voltage battery packs.