BLDC Motor Efficiency: What the Datasheet Doesn’t Tell You

Every BLDC gear motor catalogue carries an efficiency figure, and most sourcing decisions treat it as a property of the machine. It is not. It is a property of one operating point, measured under conditions the catalogue rarely states, in a part of the torque-speed map your application may never visit. If you size a drive from that single number, the gap between the catalogue value and the duty-weighted value ends up as heat in the housing and as fewer amp-hours in the battery.
This article takes the published data for a 42 mm BLDC planetary gear motor and rebuilds the efficiency curve from three values the datasheet does give you: no-load current, the stated maximum-efficiency point, and stall data. The arithmetic is simple enough to run in a spreadsheet, and it shows why two motors with identical peak efficiency can differ by 15 percentage points once a gearbox and a real duty cycle are involved. If you are comparing motors on our BLDC planetary gear motor line, this is the calculation we would run before quoting.
One Number, Three Different Efficiencies
The first thing to pin down is the measurement boundary, because “efficiency” changes meaning depending on where you put the probes:
- Motor only: shaft power out of the rotor, electrical power in at the motor terminals. The controller is outside.
- Gearbox only: torque at the output shaft divided by torque at the gearbox input, at a stated input speed. This is where most catalogue confusion lives.
- Drive train: electrical power in at the DC bus, mechanical power out at the gearbox output shaft. This is the number your battery actually feels, and it is always the lowest of the three.
A supplier who quotes 88% and a supplier who quotes 82% may be describing the same machine with different boundaries. A 42 mm BLDC that reaches roughly 70% motor efficiency at its best point, pushed through a four-stage 1:246 gearbox, delivers under half of that at the output shaft — and the controller takes another share before the battery even sees the motor. We will put real numbers on that stack later in this article.
Ask any supplier one question before comparing figures: is the quoted efficiency measured at the DC bus, at the motor terminals, or calculated? If the answer is “calculated from the torque-speed curve,” the number is a model, not a measurement, and the controller losses may not be in it at all.
The Datasheet Gives You Three Points, Not a Curve
A small BLDC datasheet rarely publishes an efficiency map. What it does publish is a torque-speed line described by three anchor points: no-load, the maximum-efficiency point, and stall. Our guide to reading a motor datasheet covers how to locate those anchors; here we use them to reconstruct efficiency.
Take the 12 V winding of the DM-BL4260-012-6000, the bare motor used in the 42 mm planetary platform:
| Datasheet anchor | Value | What it tells you |
|---|---|---|
| No-load | 6,000 rpm at ≤320 mA | Fixed losses: iron, friction, windage |
| Max efficiency point | 3,900 rpm at 600 gf·cm | Where the curve tops out |
| Stall | ≥1,710 gf·cm at ≥5.0 A | Torque constant and winding resistance |
A permanent-magnet DC motor’s torque-speed line is close to linear between no-load and stall, so the first check is whether the rated point actually sits on that line:
T(3,900 rpm) = 1,710 × (1 − 3,900/6,000) = 598.5 gf·cm ≈ the published 600 gf·cm
It does, which tells you two useful things. The published maximum-efficiency point is internally consistent, and you can interpolate torque at any intermediate speed without inventing data. From the same two anchors you can also derive the two constants the datasheet omits:
- Torque constant: kT = 1,710 gf·cm ÷ 5.0 A ≈ 342 gf·cm per amp
- Winding resistance (idealized): R ≈ 12 V ÷ 5.0 A = 2.4 Ω
Both are approximations — stall current includes some inductance and saturation effects, so treat them as estimates within a few percent. But they are estimates built from the supplier’s own published numbers, which makes them a fair basis for comparing quotations.

Rebuilding the Curve from No-Load Current and Stall Data
With kT, R and the no-load current in hand, the efficiency at any speed can be estimated in three steps. The no-load current matters because at 12 V it represents the fixed loss floor:
P_fixed(6,000 rpm) = 12 V × 0.32 A ≈ 3.8 W
Iron loss falls with speed (roughly with frequency to the power 1.3–2) and friction with speed squared or more, so we scale that 3.8 W down as speed drops, using an exponent of 1.8 as a reasonable middle. Copper loss is I²R, and I is torque divided by kT. Running this model at three points on the line:
| Operating point | Torque (gf·cm) | Shaft power | Copper loss | Fixed loss | Estimated efficiency |
|---|---|---|---|---|---|
| 2,000 rpm | 1,140 | 23.4 W | 26.6 W | 0.5 W | ≈ 46% |
| 3,900 rpm | 600 | 24.0 W | 7.4 W | 1.8 W | ≈ 72% |
| 5,700 rpm | 86 | 5.0 W | 0.15 W | 3.5 W | ≈ 58% |
Three conclusions fall out of that table, and none of them appear in a catalogue.
The plateau is narrow on both sides. Push toward stall and copper loss grows with torque squared, so efficiency collapses even though shaft power looks healthy. Run light and the fixed losses dominate, because iron and friction do not shrink when your load does. The 3,900 rpm point is not a marketing choice; it is where the two loss families cross.
Light load is not free. At 5,700 rpm the motor produces a fifth of the shaft power of the rated point, yet efficiency only drops from roughly 72% to 58%. That sounds mild until you notice that nearly 3.5 W is being burned just spinning. In a battery-powered device that idles most of the day, the fixed loss floor — not the peak number — sets the runtime.
The model has limits, and saying so is part of the method. A three-anchor reconstruction will not replace a dynamometer run. It ignores saturation above the rated point, it assumes the no-load current figure is a bare-motor value, and it puts the flat top of the curve somewhere between 3,900 and 5,000 rpm rather than at one exact point. What it does reliably is tell you whether your duty point sits on the plateau or down the slope — and that distinction is worth more than a decimal place of catalogue accuracy.
One boundary question to settle with any supplier: does the published no-load current include the driver? For a brushed motor the answer is irrelevant; for a BLDC the three-phase bridge adds a fixed overhead that may or may not be inside the 320 mA. If the driver is excluded, add its idle draw to the fixed-loss floor before running the arithmetic above.
Where the Losses Go, and What Each One Scales With
The reason efficiency moves so much across the map is that the loss terms respond to different variables. Getting the scaling right is what lets you predict behavior at operating points nobody measured:
| Loss path | Scales with | Typical share at rated point | What moves it |
|---|---|---|---|
| Copper (I²R in the winding) | Torque squared | Largest at high torque | More copper cross-section, lower kT margin |
| Iron (hysteresis + eddy) | Speed to the ~1.3–2 | Significant at high speed | Better lamination material, thinner laminations |
| Friction and windage | Speed squared and up | Small, but constant | Bearing type and preload |
| Controller switching | Switching frequency × current | 2–5 points of drive-train efficiency | Lower fsw, better MOSFETs |
| Controller conduction | Current squared | Grows with load | Lower RDS(on), parallel FETs |
| Gear mesh + grease churn | Load (mesh), viscosity (churn) | 10–35% through the gearbox | Grease choice, stage count |
The controller lines deserve a note. Commutation strategy changes the partial-load picture more than most datasheets admit: six-step control wastes torque-producing current at light load, while field-oriented control holds efficiency flatter across the map. We compared the two strategies, and their ripple behavior, in a separate article on BLDC controller selection. If your duty cycle lives below 50% of rated torque, that choice can be worth more efficiency points than any winding change.

The Gearbox Multiplies Your Error
Motor efficiency is only the first factor. Every gearbox stage takes its cut, and the cuts stack. Published gearbox data for the 42 mm platform makes the stacking visible:
| Stages | Ratios available | Gearbox efficiency | Breaking torque | Gearbox length |
|---|---|---|---|---|
| 1 | 3.7, 5.2 | 90% | 15 N·m | 33.2 mm |
| 2 | 13, 18, 25 | 81% | 25 N·m | 43.6 mm |
| 3 | 49, 67, 92, 125 | 73% | 50 N·m | 54.0 mm |
| 4 | 181, 246, 336, 458, 625 | 65% | 50 N·m | 64.4 mm |
Two things in that table are easy to misread.
First, the efficiency column is a nominal chain — roughly 90% per stage, flat. It lumps mesh loss, bearing drag and grease churning into one number. At rated torque the mesh loss of a planetary stage is typically 1–2%, so most of that 10% per stage is bearing and churning loss — which is nearly constant, not proportional to load. At half torque your real gearbox efficiency is worse than the table, because the same churning loss now divides by half the power. At high torque the table is conservative. This is why gearbox efficiency figures measured at “rated” conditions do not transfer to your duty point without adjustment.
Second, each stage is a mechanical ceiling, not just a loss. The breaking torque column means a 1:625 ratio built on four stages and a 50 N·m ceiling cannot be scaled further by adding stages — past the ceiling, more ratio only buys more length and more loss.
Now stack the drive train. Take the 72% motor estimate at 3,900 rpm and put gearboxes behind it:
- 1:5.2 single stage: 0.72 × 0.90 ≈ 65% at the output shaft; with a ~94% controller, roughly 61% from DC bus to shaft.
- 1:246 four-stage: 0.72 × 0.65 ≈ 47% at the shaft; with the controller, roughly 43%.
Same motor, same controller, same catalogue peak efficiency — an 18-point spread at the output shaft, caused entirely by ratio selection. That is the gap the catalogue number cannot warn you about. When an application asks for a very deep ratio, it is often worth revisiting whether a larger motor at a shallower ratio wins on total efficiency, noise and cost at once. For most battery-powered platforms we would rather specify a bigger rotor and two stages than chase ratio with stage count — the efficiency table above is the reason.
Temperature and Voltage Move the Number
Even a correctly measured efficiency figure is a snapshot at one temperature and one supply voltage. Both move.
Winding resistance rises about 0.393% per kelvin for copper. A motor whose datasheet efficiency was measured with the winding at 25 °C, then run continuously until the winding stabilizes at 115 °C, sees roughly 35% more winding resistance — and copper loss rises with it. In a sealed housing at 40 °C ambient, that is several efficiency points gone, and it is also the mechanism behind most overheating failures we see in enclosed applications (causes and prevention of DC motor overheating).
Supply voltage moves the operating point rather than the losses directly. A 12 V nominal battery at 12.6 V full and 10.5 V under load shifts the whole torque-speed line: for the same torque demand the current rises, and copper loss rises with its square. Efficiency optimization done at bench supply voltage quietly misses the end-of-discharge condition — which is exactly when runtime matters.
Temperature also changes the gearbox, not just the copper. Grease viscosity climbs steeply below room temperature, so churning loss in a gearbox starting at −20 °C can double or triple until the grease warms. Devices that cycle briefly and never reach thermal equilibrium never see the datasheet condition at all. How this interacts with intermittent running is covered in our explanation of motor duty cycles S1, S2 and S3.
Specifying Efficiency So a Supplier Can Answer It
A datasheet number you cannot reproduce is not a specification. When efficiency genuinely matters to your product — battery runtime, thermal budget, energy labeling — put these five items in the RFQ instead of a percentage:
- Boundary: state whether the value is measured at the DC bus, at the motor terminals, or at the gearbox output shaft, and whether the production driver is included.
- Three load points minimum: request measured efficiency at 25%, 50% and 100% of the duty torque, plus the single point where your device spends most of its time. One peak number is not a data set.
- Method: input-output measurement per IEC 60034-2-1, with the stray-loss treatment stated. Loss-summation and input-output methods can differ by over a point on small machines.
- Thermal condition: winding temperature at the time of measurement, and whether the motor reached steady state. A cold number is not a hot number.
- Driver settings: switching frequency, current control mode, and PWM scheme. The same motor on two boards can differ by several points.
None of this asks a supplier for anything unusual — it asks them to state conditions that a proper test report already contains. If a quotation resists all five requests, treat the efficiency figure as marketing until proven otherwise.
A related trap: international efficiency classes do not help here. The IE1–IE4 classes under IEC 60034-30-1 apply to line-fed AC induction motors from 0.12 kW to 1,000 kW; a converter-fed small BLDC sits outside that scheme entirely. Asking for “IE4 equivalent” on a 40 W gear motor is a category error — ask for the measured points instead.
FAQ
Is the efficiency on a BLDC datasheet measured with or without the driver?
It varies by supplier, and the datasheet usually does not say. Ask directly whether the figure is DC-bus-to-shaft or terminal-to-shaft, and whether the production driver was used. Controller losses are worth 2–5 points on small machines, so an unspecified boundary can make two quotations incomparable.
Why does the gearbox efficiency in my application not match the catalogue figure?
The catalogue value is measured at rated torque and speed, with losses that are largely load-independent (bearing drag, grease churning) folded into one number. Run the gearbox at half torque and those constant losses divide by half the power, so real efficiency drops below the table. Expect worse than catalogue at light load, and roughly catalogue or better at rated load.
What efficiency should I use in a battery-life calculation?
Use the efficiency at your duty-weighted operating point, not the peak. List the operating points the device actually visits, weight them by time, and compute the weighted average. For an idling device, remember that the fixed-loss floor consumes power even at zero output torque — runtime predictions built on peak efficiency routinely overshoot by 20% or more.
Do IEC efficiency classes like IE3 or IE4 apply to small BLDC gear motors?
No. IEC 60034-30-1 defines IE1–IE4 for line-fed AC motors between 0.12 kW and 1,000 kW. Small converter-fed BLDC machines fall outside that framework, so any “IE class” claim on a 40 W gear motor is informal. Specify measured efficiency at stated load points and conditions instead.
Related Reading
- BLDC Planetary Gear Motor: Types, Features and Functions
- How to Read a Motor Datasheet: Stall Torque, No-Load Speed and the Rated Point
- BLDC Motor Controller Selection: Trapezoidal vs Sinusoidal Commutation
- Motor Duty Cycles: S1, S2 and S3 Explained for Gear Motors
If you want the efficiency numbers behind your application instead of a catalogue peak, contact our engineering team with your duty cycle — measured points at your load conditions are part of every serious quotation.
Shenzhen DongMing Motor Co., Ltd https://szdmmotor.com
