Every motor development program begins with a familiar question:
Which motor topology should we use?
It sounds like the logical place to start.
It isn't.
The best motor architecture is not selected by comparing radial flux, axial flux, or transverse flux machines in isolation. It is selected by defining the operating conditions, manufacturing realities, package constraints, reliability requirements, and economic objectives of the application.
Motor topology is an answer — not the starting point.
This Engineering Design Study uses commercial HVAC as the case study because HVAC forces engineers to balance variables that rarely align neatly: variable-speed performance, long operating hours, thermal management, acoustic behavior, package size, production cost, supply-chain maturity, and total cost of ownership.
The purpose is not to name a universal winner. It is to identify the conditions under which each architecture becomes the stronger engineering decision.
Commercial HVAC is a systems problem disguised as a motor problem. A motor may run for thousands of hours each year, spend much of that time below rated output, and still be expected to meet demanding requirements for efficiency, noise, reliability, cost, and service life.
Unlike applications dominated by one metric, HVAC requires simultaneous optimization. Higher torque density may reduce package size but complicate assembly. Higher efficiency may justify a higher acquisition cost only when operating hours and utility rates allow the savings to accumulate. A mature topology may be less compact but easier to source, validate, and scale.
These tradeoffs make HVAC an effective teaching vehicle. It exposes the difference between optimizing a motor and optimizing the system that uses it.
| Driver | Engineering question |
|---|---|
| Performance | What torque, speed range, efficiency map, and dynamic response are required? |
| Operating profile | How many hours will the motor run, and where will it spend time on the load curve? |
| Packaging | Which dimensions, interfaces, cooling paths, and integration constraints are fixed? |
| Manufacturability | Can the design be tooled, assembled, inspected, and scaled repeatably? |
| Economics | How do manufacturing cost, utility rates, and capital constraints shape the decision? |
| Lifecycle | What creates value over the expected service life of the equipment? |
A disciplined topology study progresses from intent to constraints, application, operating characteristics, architecture, and implementation strategy. Architecture appears late in the sequence because the design space must be defined before a topology can be judged fairly.
This framework prevents a familiar architecture from becoming the default before the application has earned that decision.
The study compares conventional radial flux, yokeless axial flux, and transverse flux architectures for commercial HVAC blower or fan applications. The values are representative concept-study estimates intended to expose trade-offs - they are not production quotations or validated product claims.
The baseline comparison holds output rating, operating environment, material-pricing logic, and manufacturing assumptions constant so that the effect of architecture can be examined. A second operating case then changes utilization and part-load behavior to test whether the preferred decision shifts.
| Input | Baseline case | High-utilization case |
|---|---|---|
| Motor ratings | 1, 5, and 10 HP | 1, 5, and 10 HP |
| Speed range | 500–3,000 RPM | Reduced-speed operation within the same application class |
| Annual utilization | 50% | 90% |
| Average output | Rated-case comparison | Approximately 50% of rated mechanical output |
| Control | Comparable drive assumptions | VFD or inverter control for all topologies |
| Electricity cases | $0.12/kWh and $0.18/kWh | $0.12/kWh and $0.18/kWh |
| Price escalation | None in static comparison | 2.2% annually in lifecycle sensitivity |
Radial flux motors remain the industrial benchmark because decades of refinement have produced standardized frames, mature lamination and winding processes, proven serviceability, and a global supply base. Their primary advantage is not a single electromagnetic metric; it is manufacturing readiness. Where cost, scale, reliability, and sourcing dominate, radial flux remains difficult to displace.
Yokeless axial flux motors shorten the magnetic path and use a disc-shaped package that can provide higher torque density and substantially reduced axial length. In HVAC, that can create value when the motor is integrated into a blower, duct, or fan module. The tradeoff is a more demanding production system: permanent magnets, tighter tolerances, specialized stator construction, and less standardized assembly.
Transverse flux motors use a three-dimensional magnetic circuit to produce high torque at low rotational speed. They are most relevant where direct drive, high pole count, or exceptional low-speed torque outweigh manufacturing complexity. In the operating cases considered here, transverse flux remains technically capable but economically specialized.
A motor does not become commercially successful when the finite-element model converges. It becomes successful when the predicted performance can be manufactured consistently, economically, and repeatedly.
Radial flux benefits from mature stamping, stacking, winding, and assembly infrastructure. Axial and transverse flux designs can reduce package size or improve torque density, but magnets, complex geometry, and specialized assembly increase production risk. This is why manufacturing must begin the day the architecture is selected — not after electromagnetic optimization is complete.
Material utilization also changes the economics. Laminated cores can generate substantial stamping scrap, while pressed SMC components can approach near-net-shape production. The engineering comparison must therefore account for purchased material, usable material, secondary operations, tooling, inspection, and assembly rather than raw material price alone.
| Rating | Radial flux | Yokeless axial flux | Transverse flux |
|---|---|---|---|
| 1 HP | $140 | $175 | $260 |
| 5 HP | $370 | $465 | $665 |
| 10 HP | $610 | $710 | $1,065 |
The original concept model used $0.12/kWh. That assumption remains useful as a lower-cost reference, but it is not representative of every commercial market. The revised study adds a $0.18/kWh case to test how a higher energy price changes the economic weighting of efficiency.
The escalation factor is based on U.S. Energy Information Administration commercial-sector data. The national average commercial electricity price increased from 8.67 cents/kWh in 2005 to 13.41 cents/kWh in 2025. The compound annual growth rate over that period is approximately 2.2%. The lifecycle model applies that historical rate annually over ten years; it is a planning assumption, not a forecast.
Regional utility rates vary widely, and large facilities may face demand charges, time-of-use rates, negotiated tariffs, or transmission-level service. Detailed programs should therefore replace the national planning rate with the facility tariff.
| Electricity assumption | Purpose |
|---|---|
| $0.12/kWh, no escalation | Retains the original engineering baseline for comparison. |
| $0.18/kWh, no escalation | Tests a higher-cost commercial market without compounding. |
| $0.18/kWh + 2.2% annual escalation | Evaluates ten year exposure using the 2005–2025 historical commercial price CAGR. |
Under the original 50% utilization case, operating cost dominates total ownership cost, but the differences among architectures remain modest. Axial flux slightly reduces ten year cost through improved efficiency, while radial flux preserves the lowest upfront cost. The choice therefore remains strongly influenced by packaging and manufacturing priorities.
| Rating | Radial | Axial | Transverse |
|---|---|---|---|
| 1 HP | $4,600 | $4,535 | $4,670 |
| 5 HP | $21,660 | $21,305 | $21,725 |
| 10 HP | $42,340 | $41,930 | $42,795 |
The second case assumes 90% annual utilization, approximately 50% average mechanical output, and VFD or inverter control. Permanent-magnet architectures retain efficiency more effectively at reduced speed and torque, so the operating-cost term begins to outweigh the higher acquisition cost.
| Rating | Radial | Axial | Transverse |
|---|---|---|---|
| 1 HP | $4,240 | $4,005 | $4,140 |
| 5 HP | $19,980 | $18,995 | $19,425 |
| 10 HP | $38,950 | $37,450 | $38,195 |
At $0.18/kWh, the first-year energy cost is 50% higher than in the original model. Applying the 2.2% historical commercial price escalation rate increases the ten-year energy-cost multiplier from 10.0 to approximately 11.05. The resulting totals amplify the value of part-load efficiency without changing the underlying engineering tradeoff: radial flux remains the production benchmark, axial flux produces the lowest lifecycle cost in this operating case, and transverse flux requires a specialized torque or integration need to justify its complexity.
| Rating | Radial | Axial | Transverse | Axial savings vs. radial |
|---|---|---|---|---|
| 1 HP | $6,937 | $6,525 | $6,693 | $413 |
| 5 HP | $32,882 | $31,219 | $31,767 | $1,662 |
| 10 HP | $64,174 | $61,621 | $62,623 | $2,553 |
The preceding analysis evaluates one motor. That is the correct unit of engineering analysis, but it is rarely the unit of commercial deployment.
A supermarket may operate multiple rooftop units, condensers, evaporators, air handlers, and refrigeration fans. A data center may deploy large fleets of fan and blower motors across air-handling, computer-room air-handler, chiller, cooling-tower, and support systems. At that scale, a small per-motor difference becomes a facility-level operating decision.
The following fleet cases are intentionally illustrative. They do not claim a universal unit count or facility architecture. Their purpose is to show how the same per-motor assumptions scale.
This case applies the 90% utilization, reduced-speed profile to a fleet of sixteen 5 HP motors, beginning at $0.18/kWh and escalating electricity cost by 2.2% annually. It could represent a combination of air-handling, condenser, or refrigeration-support motors; the exact equipment mix must be defined for an actual facility study.
Reference scale note: The illustrated supermarket case uses sixteen 5 HP motors as a transparent deployment-scale reference. Actual motor counts vary by store size, zoning, refrigeration architecture, and equipment configuration; replace this count with the project-specific installed fleet during detailed analysis.
| Architecture | Ten-year fleet cost | Difference vs. radial |
|---|---|---|
| Radial flux | $526,105 | — |
| Yokeless axial flux | $499,506 | $26,599 lower |
| Transverse flux | $508,277 | $17,828 lower |
At the fleet level, the axial-flux lifecycle advantage grows from approximately $1,662 per 5 HP motor to approximately $26,600 across sixteen motors. The value may still be insufficient to justify redesign if packaging and production benefits are absent, but it is no longer negligible.
The data-center case applies the same high-utilization, reduced-speed profile to one hundred 10 HP cooling motors. This is a deliberately simplified fleet model - actual data centers use different cooling architectures, motor ratings, redundancy strategies, utility contracts, and demand-charge structures.
| Architecture | Ten-year fleet cost | Difference vs. radial |
|---|---|---|
| Radial flux | $6,417,413 | — |
| Yokeless axial flux | $6,162,148 | $255,265 lower |
| Transverse flux | $6,262,306 | $155,107 lower |
At one hundred motors, the axial-flux advantage exceeds a quarter of a million dollars under the study assumptions. The result does not prove that axial flux is the correct architecture for every data center. It proves that topology decisions must be evaluated at deployment scale. A small efficiency difference, repeated across a large continuously operating fleet, can justify a deeper system-level design study.
Soft Magnetic Composites (SMC) should not be treated as a universal substitute for laminated electrical steel. Laminations remain highly effective where flux is primarily two-dimensional, and the radial-flux manufacturing ecosystem creates the best system value.
SMCs matter when the application benefits from an architecture that conventional laminations cannot easily realize. Their magnetically isotropic behavior supports three-dimensional flux paths, while compaction can create near-net-shape components with integrated features and high material utilization.
That capability is particularly relevant to yokeless axial flux, transverse flux, switched reluctance, and other integrated magnetic circuits. The design question is therefore not “Should laminations be replaced?” It is “Does the application benefit from the geometry, flux path, integration, or production method that SMC enables?”
This framing keeps material selection in its proper place. The material follows the system requirement and architecture; it does not lead them.
Duty cycle, speed, torque, and utilization determine which efficiency and torque-density advantages matter.
Motor efficiency is only one contributor to packaging, thermal behavior, assembly, reliability, and lifecycle value.
Tooling, material utilization, inspection, assembly, and supply-chain readiness belong in the first architecture review.
A fixed utility rate can hide the effect of long operating life and regional energy exposure.
Per-motor economics can become strategic when repeated across a fleet.
SMCs create value when they enable a geometry or magnetic circuit that solves the application more effectively.
Commercial HVAC demonstrates why motor topology selection is an engineering decision rather than a ranking exercise.
Radial flux remains the benchmark when manufacturing maturity, supply-chain readiness, serviceability, and initial cost dominate. Yokeless axial flux becomes more compelling when compact packaging, part-load efficiency, high utilization, and direct-drive integration carry greater weight. Transverse flux retains a specialized role where exceptional low-speed torque justifies a more complex production system.
The revised electricity analysis strengthens the central conclusion. At $0.18/kWh and with a 2.2% annual escalation factor, small efficiency differences accumulate. At supermarket or data-center fleet scale, they become material operating-cost differences.
The best motor topology is not the architecture with the highest efficiency, greatest torque density, or lowest manufacturing cost in isolation. It is the architecture that best solves the application across its operating life—and at the scale at which it will actually be deployed.
No. Efficiency depends on electromagnetic design, winding strategy, inverter, cooling, speed, torque, and duty cycle. Axial flux can provide an advantage in some high-utilization or packaging-constrained cases, but topology alone does not determine efficiency.
$0.12/kWh preserves the original engineering baseline. $0.18/kWh tests a higher-cost commercial market and shows how lifecycle economics change when energy has greater weight.
U.S. commercial electricity prices increased from 8.67 cents/kWh in 2005 to 13.41 cents/kWh in 2025, equivalent to an approximately 2.2% compound annual growth rate. It is used as a historical planning factor, not a forecast.
No. The one-hundred-motor case is an illustrative scaling scenario. Actual data centers require facility-specific motor counts, ratings, utilization, redundancy, tariffs, and demand charges.
Not universally. SMCs are most valuable when they enable three-dimensional magnetic circuits, integrated geometry, or manufacturing approaches that are difficult to achieve with stacked laminations.
When low-speed torque, direct drive, or high pole count creates enough system value to justify specialized geometry, winding, tooling, and assembly.
As an early-stage screening framework. Final decisions require application-specific electromagnetic analysis, thermal modeling, controls evaluation, prototyping, cost validation, and supplier input.
Explore Horizon Technology's engineering approach to electric motor design, soft magnetic composites, and scalable production.