Performance Comparison of Coreless PCB AFPM Topologies for Duct Fan

Abstract

Duct fan motors must provide high torque within limited space to maintain airflow while requiring low vibration characteristics to minimize fluid resistance caused by fan oscillation. Axial Flux Permanent Magnet Motor (AFPM) offers higher torque performance than Radial Flux Permanent Magnet Motor (RFPM) due to their large radial and short axial dimensions. In particular, the coreless AFPM structure enables superior low-vibration performance. Conventional AFPM typically employs a core-type stator, which presents manufacturing difficulties. In core-type AFPM, applying a multi-stator configuration linearly increases winding takt time in proportion to the number of stators. Conversely, a Printed Circuit Board (PCB) stator AFPM significantly reduces stator production time, making it favorable for implementing multi-stator topologies. The use of multi-stator structures enables various topological configurations depending on (1) stator placement, (2) magnetization pattern of permanent magnets, and (3) rotor arrangement—each offering specific advantages. This study evaluates and analyzes the performance of different topologies based on efficient arrangements of magnets and stators, aiming to identify the optimal structure for duct fan applications. The validity of the proposed approach and design was verified through three-dimensional finite element analysis (FEA).

1. Introduction

To effectively drive airflow within the duct, the duct fan must maintain a constant static pressure during operation. Accordingly, the motor must be capable of delivering sufficient rated torque. Moreover, due to installation space constraints, the duct fan motor requires a compact form factor in the axial direction [1]. Vibrations generated during fan rotation are transmitted through the motor and increase the fluid resistance experienced by the fan blades. Therefore, low-vibration characteristics that minimize such effects are essential. The AFPM, with its large rotor diameter and short axial length, is well-suited to deliver high torque, making it an appropriate choice for duct fan applications [2,3].

AFPM can be classified into core type and coreless type. In core-type AFPM, design strategies such as laminated structures are required to minimize eddy current paths generated in the stator core. However, the stator core typically features complex shapes such as teeth and shoes, making it difficult to manufacture using rolled electrical steel sheets or laminations, as commonly done in RFPM. To address these challenges, research is being conducted using materials such as amorphous alloy sheet and soft magnetic composite (SMC). Specifically, amorphous alloy sheet can be formed into rolls, then cut or punched and re-rolled to create an effective laminated structure for reducing eddy current losses. However, when applied to an AFPM, the use of amorphous alloy sheet becomes disadvantageous for mass production due to interlayer alignment issues that worsen with increasing outer diameter and high mechanical forming complexity. SMC is fabricated by press-molding insulated magnetic powder, offering ease of manufacturing and excellent high-frequency core loss characteristics. Nevertheless, for motors operating in relatively low-speed regions, SMC exhibits inferior magnetic properties and mechanical strength compared to electrical steel. Additionally, the insulation coating on the powder particles can be damaged during the pressing process [4]. Furthermore, core-type AFPM requires sufficient axial length to accommodate windings on the stator core, limiting axial compactness. The use of complex geometries such as teeth, shoes, and slot openings also causes magnetic reluctance variation, leading to vibration issues due to cogging torque.

The coreless type AFPM, which does not include a stator core, offers the advantage of significantly lower cogging torque, resulting in superior vibration characteristics. However, it presents limitations in winding implementation, and active research is ongoing to address this challenge [5,6]. A coreless type AFPM utilizing a PCB stator eliminates the need for separate coil-winding machinery, as the winding process is integrated into standard PCB fabrication steps. This drastically reduces the production takt time relative to core-type AFPM, making the design well suited to high-volume manufacturing; moreover, it is well established that PCB-stator fabrication is significantly faster than conventional coil-winding processes [7]. Additionally, while the thickness of a PCB substrate varies depending on its oz rating, a 6-layer board with a 2 oz specification typically achieves a thickness of approximately 1 mm, thereby minimizing the magnetic air gap. Figure 1 illustrates the stator structures and cross-sectional views of a core-type AFPM stator and a PCB stator for a coreless-type AFPM.

Active research is underway on PCB stator motors, focusing on various aspects inherent to the coreless motor structure. These include studies on AC loss caused by the time-varying flux linkage directly interacting with the windings [8,9], designs utilizing Halbach magnets to reduce leakage flux and enhance air-gap flux density [10,11], coil pattern optimization with slit structures for AC loss reduction [12], thermal design investigations considering PCB vias and pads [13], and comparisons between PCB-stator and laminated-core-stator axial flux permanent magnet motors [14]. When the rotor of a coreless motor rotates, eddy currents are induced in the conductive windings due to the time-varying magnetic flux generated by the permanent magnets. This is one of the major sources of loss in a coreless motor. In a conventional AFPM with a stator core, most of the magnetic flux follows the high-permeability core path, resulting in minimal flux passing directly through the windings. In contrast, a coreless motor has a stator with permeability similar to that of air, allowing more flux from the magnets to flow directly through the windings. This leads to additional AC losses caused by the skin effect and proximity effect. The aforementioned AC loss refers to the loss occurring in the stator winding. Although AC losses in the permanent magnets should ideally be considered, they are neglected in this study due to the use of ferrite magnets, which have relatively low electrical conductivity and therefore exhibit negligible eddy current losses. In addition, since the study focuses on a small-scale AFPM, rotor core iron losses are also considered negligible.

Therefore, it is essential to identify methods for reducing AC loss in PCB stator AFPM. Since AC loss primarily arises from the skin and proximity effects, it can be mitigated through modeling based on finite element analysis and trend analysis of these effects induced by external magnetic fields. Figure 2 illustrates the magnetic flux linkage between the magnets and the stator in both core-type and coreless-type AFPM.

In addition, various studies have been conducted on AFPM configurations that utilize multi-stator structures. For example, research has been conducted on high-speed axial flux induction motors employing dual-stator structures [15], comparative analyses between core-type and coreless type configurations in dual-stator single-rotor AFPM [16], and comparative study on novel dual stator radial flux and axial flux permanent magnet motors [17]. However, when a multi-stator structure is applied to a conventional core type AFPM, additional winding processes are required, resulting in increased takt time. Moreover, sufficient axial length for teeth is necessary to accommodate the windings. In contrast, a PCB stator can be reused with identical boards, which significantly reduces production time when implementing various topologies such as multi-stator configurations compared to the core type AFPM. In addition, the PCB stator offers high design flexibility, allowing various topologies to be implemented by simply changing its placement depending on the desired configuration. A PCB stator motor can be categorized into different topologies based on the magnetization pattern of the rotor and the arrangement of the stator and rotor. As examples, the stator can be arranged as a single module composed of four 6-layer PCB stator units connected in series, or as two separate modules, each consisting of two 6-layer PCB stator units connected in series. The most representative examples are the single stator–single rotor type and the single stator–double rotor type [7,18,19]. In the single stator–single rotor type, the PCB stator is attached to the stator yoke, and the rotor is placed on one side, forming a single-sided air gap. This structure minimizes the air gap length, which most significantly affects electromagnetic performance, and enables heat dissipation through contact between the stator back yoke and the housing. However, since only one magnetic source is present, phase back-EMF imbalance may occur. The single stator–double rotor structure is the most commonly utilized configuration. It does not require a stator back yoke, and the presence of magnetic sources on both sides results in minimal phase back-EMF imbalance. In PCB stator AFPM, relatively simple methods such as increasing the magnet thickness or connecting multiple PCB stators in series to increase the total number of series turns can be used to improve performance within a limited outer diameter. While these approaches are effective for initial performance enhancement, stacking multiple PCB layers increases the magnetic air gap, which may degrade electromagnetic performance. Additionally, when the magnet thickness exceeds a certain threshold, performance improvement no longer scales linearly due to magnetic flux saturation. Accordingly, this study investigates a topology design approach for enhancing performance by adjusting the arrangement of the rotor and PCB substrate. The most widely used single stator–double rotor structure was selected as the reference model, and the electromagnetic performance of different topologies was analyzed based on the same number of PCB stators and the same amount of magnets. The structure of this paper is as follows: (1) analysis of performance variation according to pole-slot combinations in the single stator–double rotor type PCB stator AFPM based on conventional motor design specifications; (2) selection of pattern end-turn parameters and design of a linear conductor structure for AC loss reduction based on the selected pole-slot combination; and (3) design of several PCB stator AFPM topologies using the same PCB stator as the reference model, with no-load back-EMF and load torque, and loss analysis to determine the optimal model. The proposed models and design process were validated using finite element analysis (FEA).

2. Selection of the Baseline Model

Table 1. Motor specifications for 168 W Target Duct fan.

Dimensions	Conventional Motor	Unit
Motor Outer Diameter/stack length	107/43	mm/mm
Permanent Magnet	Ferrite Magnet	-
Parameters	Conventional Motor	Unit
Rotating Speed	3350	rpm
Rated Torque	0.48	Nm
Efficiency	82	%

presents the design specifications for a conventional duct fan motor system. The target rotational speed is 3350 rpm, and a rated torque of 0.48 Nm is required. Under constant speed conditions, torque is the primary factor determining motor output. Equation (1) expresses the torque of an AFPM in terms of motor size and magnetic flux density. Equation (2) expresses the torque of an AFPM in terms of motor size and magnetic flux density. Commonly in both equations, $k_{\omega 1}$ denotes the winding factor, $B_{avg}$ represents the average air-gap flux density, and $ac$ refers to the specific electric loading, defined as the total electric loading divided by the air-gap circumference. Here, the total electric loading is expressed as the product of the equivalent series turns and the armature current. In Equation (1), $D_g$ represents the air-gap diameter, while in Equation (2), $D_{out}$ indicates the rotor outer diameter. Furthermore, $K_d$ is a constant that denotes the ratio of the rotor inner diameter to the rotor outer diameter. Since the torque of an AFPM is proportional to the cube of the rotor outer diameter, the motor was designed using the maximum allowable outer diameter while minimizing the lamination stack height. This approach was used to select an appropriate motor size.

$$T_{RFPM}=\left({\displaystyle \frac{\pi }{4}}ac{k}_{\omega 1}{B}_{avg}\right)D_g^2{L}_{stk}$$

(1)

$$T_{AFPM}=({\displaystyle \frac{1}{8}}ac\pi k_{\omega 1}{B}_{avg})(1-K_d^2)D_{out}^3$$

(2)

In the PCB stator proposed in this study, the winding is distributed, and electrical connection between the active conductor and end-turn sections is essential. Since the end-turn does not contribute to torque generation, the permanent magnets on the rotor are applied only up to the outer region, excluding the end-turn area. Accordingly, the PCB stator AFPM is designed such that the stator outer diameter is larger than the rotor outer diameter, and the previously selected maximum allowable outer diameter is used as the stator outer diameter. The maximum lamination height was determined based on the lamination size of a conventional motor, as summarized in Table 1 along with the overall motor specifications and dimensional constraints. The thickness of the PCB copper pattern was selected as 2 oz, which provides optimal electrical resistance and dimensional characteristics (approximately 1 mm in axial direction for 6 layers). For the base model analysis, a 12-layer PCB stator was configured by serially connecting two sets of 6-layer boards with 6 parallel windings. Figure 3 shows the series connection schematic of two 6-layer PCB, along with copper thickness and the structure of a 12-layer PCB.

The winding layout of the PCB stator proposed in this study adopts a distributed winding configuration with a pole-to-slot ratio of 1:3, which enables the winding factor to be set as 1. Within the selected design dimensions, performance analysis was conducted based on combinations of optimal pole and slot numbers and the number of conductors per slot. Using the 8-pole configuration of the conventional model as a reference, analysis was performed on a PCB stator AFPM with 8 poles and 24 slots, with 6 conductors per slot, as well as other combinations that provide similar or higher total series turns. In a PCB stator AFPM, via holes are required to establish electrical connections between conductors in the axial direction. However, implementation of via holes requires a minimum spacing between them, which imposes a constraint on the minimum inner diameter. Therefore, the analysis was carried out using only combinations of pole number, slot number, and conductors per slot that satisfy the via hole spacing requirements. Figure 4 shows the variation in the number of conductors per slot under the same pole-slot combination.

Table 2. Performance table according to pole-slot combinations and conductor per slot.

Parameters	Single Stator-Doble Rotor									Unit
Conductor per Slot	3	4			5			6		-
Pole/Slot	16/48	16/48	14/42	12/36	14/42	12/36	10/30	10/30	8/24	-/-
Output power	55.6	77.6	60.8	42.4	60.2	44.2	27.7	39.1	32.5	W
No Load Back EMF	7.6	10.2	9.0	7.6	11.0	9.4	7.3	6.7	5.1	Vrms
Current	3.8	3.0	2.9	2.8	2.1	2.0	1.9	2.3	2.7	Arms
Power factor	0.99									-
Current Density	21									Arms/mm2
Phase Resistance	0.36	0.61	0.58	0.23	0.96	0.88	0.85	0.78	0.60	Ohm
Copper Loss	15.8	9.8	9.2	8.6	4.8	4.4	3.9	5.8	8.0	W
AC Loss	31.6	14.3	17.8	21.9	9.5	12.0	13.7	7.3	9.2	W
Efficiency	54.0	76.3	69.2	58.2	80.8	73.0	61.1	75.0	65.4	%

presents the performance of comparison models based on the analysis of pole-slot combinations and conductors per slot in the PCB stator AFPM. Under the 1:3 pole-to-slot configuration, an increase in the number of poles leads to an increase in the number of slots, which in turn increases the total number of series turns on the board and results in a higher no-load back-EMF. However, increasing the number of conductors per slot extends the current path on the PCB, leading to higher resistance. Conversely, when the number of poles decreases, the total number of series turns is reduced, which lowers the resistance. However, the effective conductor area on the PCB increases, resulting in greater AC loss in the winding. Additionally, the reduction in series turns also leads to a decrease in no-load back-EMF. Based on this analysis, the model with 16 poles, 48 slots, and 4 conductors per slot was selected as optimal, as it delivers the highest output within the allowable current density limit. Although the selected current density exceeds the values commonly applied in conventional motors considering standard cooling approaches, it was validated through prototype fabrication and performance evaluation, confirming that the design operates without thermal or functional issues [16].

3. Design of the Geometry for a Single Stator–Double Rotor

3.1. Output Improvement Through Analysis of PCB End-Turn Parameters

Following the selection of the pole-slot and conductors-per-slot combination, the basic PCB stator design was carried out through the end-turn pattern layout. The end-turn regions were analyzed separately for the outer and inner diameter sides, which are referred to as the outer end-turn and inner end-turn, respectively. Initially, an analysis of the outer end-turn thickness was conducted. As the outer end-turn thickness increases, the conductor area in the current path also increases, resulting in reduced resistance of the PCB stator. However, since the outer diameter of the PCB stator was constrained by a predefined limit, any increase in the outer end-turn thickness was implemented in the inward direction, toward the inner diameter. This design concept is illustrated in Figure 5. As the outer end-turn thickness increases, the resistance decreases, allowing a higher maximum current to be applied within the current density limit, thereby increasing the output. Conversely, the increase in end-turn thickness reduces the effective conductor length, which results in a smaller rotor outer diameter and a slight decrease in the no-load back-EMF. However, the effect of the increased current due to reduced resistance is more significant than the decrease in back-EMF. Figure 6 illustrates the output according to the variation in the outer end-turn thickness. Based on the analysis of the trade-off between these two parameters, the optimal point for maximizing output was identified, and the outer end-turn thickness was set to 2.75 mm.

Figure 7 illustrates the method of modifying the inner end-turn structure. The selection of the inner end-turn thickness was conducted using two approaches. Similar to the analysis of the outer end-turn, performance was compared for cases in which the inner end-turn thickness was increased either toward the outer diameter or the inner diameter, as shown in Figure 7. In the former case, a trend similar to that observed in the outer end-turn analysis was identified, where the output increases initially and then begins to decrease beyond a certain point. However, as long as the inner diameter of the PCB stator is larger than the shaft outer diameter, there is no design constraint. Therefore, the inner end-turn thickness can be increased in the inward direction, constrained only by the shaft diameter. This allows the inner end-turn thickness to be increased while maintaining the effective conductor length, which helps preserve the no-load back-EMF and reduce the resistance of the PCB stator. As the end-turn area increases, the allowable current within the limited current density also increases. A trade-off occurs where copper loss increases due to higher current, but the increase in copper loss becomes saturated due to the decreasing resistance. Figure 8 illustrates the output according to the variation in the inner end-turn configuration. Consequently, a maximum output point was observed, and the inner end-turn thickness was finally set to 2.75 mm to achieve the highest output.

Although the optimal PCB stator specifications were determined through conductor pattern design, a higher voltage utilization was required to meet the target output within the allowable current density. The current configuration consists of two 6-layer PCB boards connected in series, each composed of six parallel windings through vias, resulting in a total of 64 series turns. As shown in Table 1, the stator outer diameter is limited to 107 mm, which restricts further output enhancement through diameter expansion. Therefore, to increase the voltage utilization, additional axial length was employed by connecting four 6-layer PCB boards in series. As a result, when a total of 24 PCB layers were used, the base model with a single stator–double rotor type satisfied the target output of 168 W under the current density constraint. Figure 9 shows the schematic diagrams of the 12-layer and 24-layer PCB AFPM, and the performance comparison of these two models is summarized in

Table 3. Performance comparison between 12-layer PCB AFPM and 24-layer PCB AFPM.

Dimensions	12 Layer PCB AFPM	24 Layer PCB AFPM
Output	104	168	W
Rated speed	3350	3350	rpm
No Load Back-emf	9.45	15.13	Vrms
Pole/slot	16/48	16/48	-/-
Input Current (@21 Arms/mm2)	4.1	4.1	Arms
Current Density	21	21	Arms/mm2

.

3.2. Enhancing Efficiency by Reducing AC Losses

AC loss is one of the major losses in the PCB stator AFPM, which is a type of coreless structure. One effective method to reduce this loss is the application of a linear conductor pattern, which removes portions of the coil area where AC loss is concentrated [16]. Figure 10 shows the effective conductor shape when applying the linear conductor pattern, the eddy current paths induced in the PCB stator by the time-varying magnetic flux, and a comparison of the eddy current paths with and without the linear pattern. The linear conductor method involves removing portions of the traditional circular sector-shaped coil pattern that are significantly affected by the external magnetic field and redesigning the winding into a linear form. This approach can be easily implemented using standard PCB manufacturing processes.

While the application of the linear conductor pattern reduces the cross-sectional area of the conductor and leads to an increase in winding resistance, thereby increasing copper loss, it also results in a reduction in AC loss. As such, a trade-off point exists between the two types of losses. When the 24-layer PCB board, as previously selected, is used, the increase in AC loss becomes more significant than the increase in copper loss compared to the 12-layer configuration. This occurs because the increase in no-load back-EMF due to the higher number of series turns reduces the input current, thereby limiting the increase in copper loss. To identify the trade-off point, the linear conductor pattern was applied to the original PCB stator conductor shape, and performance trends were analyzed based on changes in conductor width, focusing on the increase in copper loss and the decrease in AC loss. Figure 11a shows the trends of copper loss and AC loss as a function of linear thickness, while Figure 11b presents the efficiency variation. When the conductor width is 0.7 mm, the optimal balance between copper loss and AC loss is achieved, enabling maximum motor efficiency. Therefore, the conductor thickness was set to 0.7 mm.

4. Analysis of the APFM Topology in PCB Stators

4.1. Selection of Topology and Comparison of Model-Specific Characteristics

Based on the previously selected baseline model, the Single Stator–Double Rotor type, a comparative analysis of PCB stator AFPM topologies was conducted while maintaining the same number of PCB layers and the same amount of permanent magnets. The analysis focused on improving the efficiency of magnet and PCB stator placement and implementing additional structural modifications for new topology configurations. Model A is structured by separating the 24-layer PCB board used in the conventional single stator–double rotor type into two 12-layer boards, with a stator back-yoke inserted between them. When the 24-layer PCB boards are stacked continuously, the total magnetic air gap of the PCB stator becomes 8.6 mm, which constitutes a relatively large air gap from the motor’s perspective. By inserting a stator back-yoke with high magnetic reluctance between the boards, this configuration compensates for the excessive magnetic air gap, offering advantages in terms of magnetic flux path control and electromagnetic performance.

Figure 12 shows the structure and magnetic flux diagram of Model A. A key distinction between the two types of Model A lies in the magnetization direction of the magnets, allowing classification into NN-type and NS-type configurations. However, this topology requires the presence of a stator core. To reduce core loss in the stator back-yoke, SMC or amorphous steel sheets are typically considered. In this study, SMC was selected as the stator core material due to its suitability for the simple geometry of the stator back-yoke. Unlike the complex shape of conventional core-type stators, the back-yoke in Model A features a form that is much easier to fabricate.

Figure 13 illustrates the structure and magnetic flux flow of Model B. Model B is a modified configuration based on the placement of the 24-layer PCB board and the rearrangement of the rotor and stator from the conventional single stator–double rotor structure. In this configuration, the magnetic air gap from the magnetic source to the stator back-yoke is reduced by half compared to the original structure. One distinguishing feature of Model B is its central rotor configuration, which allows structural variations. Specifically, it can be categorized into two types: one in which the magnets are attached to both sides of the rotor core, and another in which the rotor core is removed to reduce motor weight, with additional structures used to support the magnets. For the type using a rotor core, the volume of the core was determined to maintain the same magnetic saturation level as in the original single stator–double rotor structure.

Figure 14 illustrates the structure and magnetic flux flow of Model C. Model C is configured by dividing the PCB board into two sets of 12-layer boards and placing an additional rotor at the center of the motor by splitting the magnets while maintaining the same total amount of magnet usage. This configuration is generally referred to as a multi-stator multi-rotor structure. Similar to Model B, if the central rotor core in the motor is removed and a support structure is employed to secure the magnets, the axial length can be reduced and the motor can be made lighter. Based on the selected models, performance analysis was conducted using finite element analysis (FEA).

4.2. Comparative Performance Analysis Based on Topology

Figure 15 presents the graph of no-load back-EMF according to each topology. All topology comparisons were conducted based on the same number of PCB stator layers and the same amount of magnet usage. For the rotor core, performance analysis was carried out using the rotor back-yoke thickness that yields a magnetic saturation level equivalent to that of the rotor back-yoke in the conventional single stator–double rotor structure. Model A is designed to compensate for the magnetic flux path between the PCB layers using a stator back-yoke with high magnetic reluctance. However, despite the intended improvement in the magnetic flux path, the presence of magnetic leakage within the high-reluctance stator back-yoke resulted in a decrease in performance. Specifically, the no-load back EMF RMS value was approximately 3% lower than that of the original type. In the case of Model B, both types exhibited a decrease in no-load back EMF compared to the original model. This is attributed to the less efficient arrangement of the rotor and stator compared to the single stator–double rotor configuration. When comparing the two types—one with a rotor back-yoke and one without—the type with the rotor back-yoke showed a slightly higher no-load back EMF RMS value by approximately 0.2%. Although the target output was not achieved, the use of a rotor back-yoke with high magnetic reluctance was found to be advantageous in terms of performance under the given amount of magnet usage. In the case of Model C, the no-load back EMF RMS value improved by up to 5.3% compared to the original type, due to the more efficient arrangement of magnets. Both types of Model C exhibited higher no-load back EMF than the Single Stator–Double Rotor configuration. Specifically, the structure without a rotor back-yoke showed an additional 1.8% increase in no-load back EMF compared to the one with a back-yoke. However, beyond a certain magnet thickness, the use of a rotor back-yoke for the central rotor magnets was found to be more effective in terms of efficiency.

Since all configurations share the same PCB stator geometry, the maximum allowable current based on the current density limit of 21 Arms/mm² remains constant. As a result, both topologies of Model C, which showed the greatest increase in no-load back emf, enable higher output compared to the original type. Among them, Model C-1 exhibits the highest output, with a 7.5% increase compared to the conventional model. Additionally,

Table 4. Performance comparison of PCB stator AFPM with different topologies.

Dimensions	SSDR	Model A-1	Model A-2	Model B-1	Model B-2	Model C-1	Model C-2	Unit
Output	168	163.3	165	164.8	164.6	180.6	177.7	W
Rated speed	3350							rpm
Torque	0.48	0.46	0.47	0.47	0.47	0.51	0.5	Nm
No Load Back EMF	15.15	14.63	14.78	14.79	14.76	16.24	15.95	Vrms
Current	3.81							Arms
Power factor	0.99							-
Current Density	21							Arms/mm2
Copper Loss	41.2							W
AC Loss	4.47	3.93	3.89	4.28	4.21	4.99	4.54	W
Core Loss (SMC 700 3P)	-	3.41	2.41	4.31	4.54	-	-	W
Efficiency	79	77	77	76	76	80	80	%

summarizes the specifications of the final analyzed models, indicating that the stacking length can be reduced by approximately 50% relative to the stacking limit of the conventional model. Accordingly, this study selects Model C-1 as the optimal topology for the PCB stator AFPM. Figure 16 presents the THD analysis of Model C-1, further highlighting its performance advantages.

5. Conclusions

This study presents the design of a PCB stator-based AFPM optimized for duct fan applications under size constraints and analyzes the performance of various topologies based on different rotor and stator arrangements. To address the manufacturing limitations and structural constraints of conventional core-type AFPM, a coreless structure was adopted. The stator was designed using PCB fabrication processes, enabling automated production and facilitating the implementation of diverse topologies.

In the early design stage, combinations of pole-slot numbers and the number of conductors per slot were evaluated to identify the configuration that delivers the highest output. Subsequently, outer and inner end-turn patterns were optimized to increase the allowable current, thereby enhancing output. Additionally, a linear conductor pattern was applied to effectively reduce AC loss. Based on this, a baseline single stator–double rotor topology was applied, and comparisons among topologies with different rotor and stator arrangements were conducted while maintaining equal magnet usage and an identical number of PCB stators. As a result, Model C-1 exhibited approximately 5.3% higher no-load back EMF and 7.5% higher output compared to the baseline, demonstrating its superior efficiency. While previous studies have primarily focused on topology comparisons in core-type AFPM, this study contributes by proposing an efficient rotor–stator arrangement specifically designed for PCB stator-based AFPM. This result demonstrates that performance improvement is achievable under identical resource constraints and provides valuable insights for future motor designs utilizing PCB stators in similarly constrained environments.