Guide · Engines
Horsepower vs torque: which one makes a car accelerate?
Torque is how hard an engine twists its shaft; power is that torque multiplied by how fast the shaft turns. In SI units, power in kW = torque in N·m × rpm ÷ 9549.3. In mechanical horsepower and pound-feet, hp = lb-ft × rpm ÷ 5252. They are not two rival properties of an engine: one is calculated from the other.
What accelerates the car is force at the tires, and gears can trade engine speed for wheel torque without changing power. At any road speed, the most thrust a car can have is the power it can deliver at that speed, divided by the speed. So when the gearbox can keep the engine near its power peak, power sets acceleration. The torque figure matters when you stay in one gear, and a broad torque curve keeps more power available between gear changes.
DraftThis page has not completed editorial review. It is hidden from search engines until it has.
Torque, engine speed and power: one equation
Torque is a turning moment: a force multiplied by the length of the lever arm it acts on. It is measured in newton-meters (N·m) or pound-feet (lb-ft). Power is the rate of doing work. For a rotating shaft, power equals torque multiplied by angular speed.
With engine speed in revolutions per minute, angular speed is rpm × 2π ÷ 60 radians per second, so power in watts = torque in N·m × rpm × 2π ÷ 60. Because 60,000 ÷ 2π = 9549.3, the same statement in kilowatts is the first formula below.
One mechanical horsepower is defined as 550 foot-pounds of work per second, which is 33,000 per minute. Dividing 33,000 by 2π gives 5252, which is the second formula. One consequence: on a graph that plots hp and lb-ft against the same number scale, the two curves can cross only at 5252 rpm. Where they cross says nothing about the engine; it comes from the units.
power (kW) = torque (N·m) × rpm ÷ 9549.3
power (hp) = torque (lb-ft) × rpm ÷ 5252
| Engine speed (rpm) | Torque (N·m) | Torque (lb-ft) |
|---|---|---|
| 3,000 | 350.1 | 258.3 |
| 4,500 | 233.4 | 172.2 |
| 6,000 | 175.1 | 129.1 |
| 7,500 | 140.1 | 103.3 |
Arithmetic from the definition of power, not measurements of any engine. 110 kW is the peak power of the illustrative car used below.
What pushes the car: force at the tires
Engine torque reaches the road through the gearbox and the final drive. Each one multiplies torque by its ratio and divides rotational speed by the same ratio. The tire then turns that torque into a force at the contact patch.
A gear multiplies torque, not power. Leaving losses aside, the power coming out of a pair of gears equals the power going in: torque goes up by the ratio and speed comes down by the ratio. Comparing two engines by their torque figures alone therefore says little about thrust, because thrust depends just as much on the gearing between engine and road.
thrust = engine torque × gear ratio × final drive ratio × driveline efficiency ÷ tire radius
For the low-revving engine in the experiment below, at its torque peak in second gear: 280 N·m × 2.0 × 3.1 × 0.91 ÷ 0.305 m = 5,180 N. The efficiency and the tire radius are this model’s assumptions for the illustrative car.
Why power sets the limit at every speed
Thrust multiplied by road speed is the power arriving at the wheels. Turn that around: at a road speed v, thrust cannot be more than the power the powertrain can deliver at that speed, divided by v. No choice of gear can beat this. Gearing only decides how close to it you get.
So when you are free to choose the gear, the best one at any speed is the one in which the engine makes the most power at that road speed. That is usually the gear that puts it nearest its power peak, not its torque peak. In a higher gear the engine may sit right on its torque peak, but that torque is multiplied by a smaller ratio, and the product at the wheels is smaller.
Modeled estimate
| Gear | Engine speed (rpm) | Engine torque (N·m) | Thrust (N) | Share of the power limit |
|---|---|---|---|---|
| Gear 2 | 4,315 | 243 | 4,488 | 100% |
| Gear 3 | 3,021 | 279 | 3,608 | 80% |
| Gear 4 | 2,266 | 276 | 2,685 | 60% |
| Limit: peak power × driveline efficiency ÷ speed | — | — | 4,504 | 100% |
In this model the low-revving engine makes peak torque at 2,700 rpm and peak power at 4,500 rpm. The gear that puts the engine nearest the power peak gives the most thrust, even though engine torque is lower there.
When the torque figure does matter
- When you stay in one gear
- With the ratio fixed, thrust is engine torque multiplied by a constant. The car pulls hardest at the torque peak, and the shape of the torque curve is the shape of the acceleration. In-gear acceleration without a downshift is where a large torque figure at low engine speed shows.
- When the curve is broad
- Power at any engine speed is torque × rpm. An engine that holds its torque over a wide band of engine speeds also stays close to its peak power over a wide band. Each upshift drops the engine speed; with a broad curve it drops to a point where power is still high, so the average power used over a run is closer to the peak figure.
- What the single number cannot tell you
- Peak torque says how high the torque curve gets, not how wide it is or where it sits. Two engines with the same peak torque can have very different curves, and two engines with the same peak power can have very different peak torque.
Where to change gear
The quickest point to change up is where thrust in the next gear becomes equal to thrust in the current gear. Below that speed the current gear pulls harder; above it the next gear does. If the current gear still pulls harder all the way to the rev limit, the best point is the rev limit.
For an engine whose torque rises to a single peak, that point is always above the torque peak. At the torque peak the engine is making its largest torque and the current gear multiplies it more than the next gear would, so the next gear cannot match it there. Changing up at the torque peak gives away thrust.
This site’s model changes up when the next gear would give more thrust, or at the rev limit. When it picks that point it does not count the time lost during the change itself. In the chart below, the best change-up points are where one numbered line passes under the next.
Try it: the same peak power, two different engines
The car is the illustrative gasoline hatchback: a made-up vehicle with round numbers, not a real one. Its peak power stays at 110 kW throughout. You choose where in the rev range the engine makes that power, which sets its peak torque, and you choose the final drive ratio.
Inputs
Illustrative vehicle
0–60 mph
8.06s
from rest, changing gear
7.92–8.59 s across the assumption ranges
0–100 km/h
8.51s
from rest, changing gear
8.35–9.04 s across the assumption ranges
60–100 km/h in gear 3
4.77s
gear held · engine 2,266–3,776 rpm
4.64–4.98 s across the assumption ranges
With 280 N·m at 2,700 rpm, 110 kW at 4,500 rpm and a 3.10:1 final drive, the illustrative gasoline hatchback reaches 60 mph in 8.06 s and 100 km/h in 8.51 s from rest. Held in gear 3, it takes 4.77 s to go from 60 to 100 km/h.
The engine these inputs describe
- Peak power
- 110 kW at 4,500 rpm
- Peak torque
- 280 N·m at 2,700 rpm
- Rev limit
- 5,000 rpm
The model builds the torque curve from these points by a fixed method. It is not a measured curve.
Something to try
Load the high-revving engine, then lower its final drive ratio from 4.90 to 3.10, the gearing the low-revving engine uses. Held in gear 3, 60 to 100 km/h goes from about 5.3 s to about 8.8 s, while 0–100 km/h moves by about 0.3 s. With the same gearing, the engine with less torque has less thrust at any road speed that both engines can reach in a given gear; from rest, the gearbox can still keep it near its power peak.
Thrust in each gear
- Best gear (solid line, round marker)
- Peak power ÷ speed (dashed line, square marker)
- Thin numbered lines: thrust in each gear, from idle to the rev limit.
Drag, or use the arrow keys, to read the lines at a chosen speed.
Show the data as a table
| Speed (km/h) | Gear 1 | Gear 2 | Gear 3 | Gear 4 | Gear 5 | Gear 6 | Best gear | Peak power ÷ speed |
|---|---|---|---|---|---|---|---|---|
| 10 | 7,131 | — | — | — | — | — | 1 | 36,036 |
| 20 | 8,653 | 4,238 | — | — | — | — | 1 | 18,018 |
| 30 | 9,059 | 4,761 | 3,010 | 2,075 | — | — | 1 | 12,012 |
| 40 | 8,529 | 5,075 | 3,271 | 2,257 | 1,717 | — | 1 | 9,009 |
| 50 | 7,096 | 5,181 | 3,461 | 2,409 | 1,834 | 1,426 | 1 | 7,207 |
| 60 | — | 5,105 | 3,579 | 2,531 | 1,934 | 1,505 | 2 | 6,006 |
| 70 | — | 4,874 | 3,626 | 2,623 | 2,019 | 1,575 | 2 | 5,148 |
| 80 | — | 4,488 | 3,608 | 2,685 | 2,088 | 1,636 | 2 | 4,504 |
| 90 | — | 3,843 | 3,536 | 2,716 | 2,140 | 1,688 | 2 | 4,004 |
| 100 | — | — | 3,412 | 2,718 | 2,177 | 1,731 | 3 | 3,604 |
| 110 | — | — | 3,234 | 2,696 | 2,197 | 1,765 | 3 | 3,276 |
| 120 | — | — | 3,001 | 2,652 | 2,202 | 1,790 | 3 | 3,003 |
| 130 | — | — | 2,623 | 2,586 | 2,193 | 1,806 | 3 | 2,772 |
| 140 | — | — | — | 2,497 | 2,173 | 1,813 | 4 | 2,574 |
| 150 | — | — | — | 2,386 | 2,140 | 1,812 | 4 | 2,402 |
| 160 | — | — | — | 2,251 | 2,096 | 1,804 | 4 | 2,252 |
| 170 | — | — | — | 2,052 | 2,040 | 1,789 | 4 | 2,120 |
| 180 | — | — | — | — | 1,972 | 1,768 | 5 | 2,002 |
| 190 | — | — | — | — | 1,892 | 1,740 | 5 | 1,897 |
| 200 | — | — | — | — | 1,797 | 1,706 | 5 | 1,802 |
Modeled estimate
Two engines, the same 110 kW
| Engine | Peak torque | Final drive | 0–60 mph | 0–100 km/h | 60–100 km/h, gear 3 |
|---|---|---|---|---|---|
| Low-revving engine | 280 N·m | 3.10 | 8.06 s | 8.51 s | 4.77 s |
| High-revving engine | 168 N·m | 4.90 | 8.75 s | 9.23 s | 5.25 s |
Each loaded engine comes with a final drive ratio chosen so that every gear reaches the rev limit at about the same road speed. Both engines are given the same rotational inertia, so the shorter gearing of the high-revving engine makes that inertia count for more in each gear.
Assumption
Assumptions used in this run
These inputs are not part of the car’s stated specifications, so the model chooses them. The run above uses the “value used” column; the range line under each time shows how far the result moves when every assumption is varied across its range.
| Assumption | Value used | Range considered |
|---|---|---|
| Idle speed (rpm) | 800 | 700 – 900 |
| Torque at idle ÷ peak torque | 0.75 | 0.60 – 0.88 |
| Power lost between peak-power speed and the rev limit | 0.08 | 0.02 – 0.20 |
| Engine speed held at launch (rpm) | 2,700 | 1,890 – 3,510 |
| Gear-change time (s) | 0.35 | 0.25 – 0.50 |
| Engine rotational inertia (kg·m²) | 0.20 | 0.10 – 0.35 |
| Drag coefficient (Cd) | 0.30 | 0.26 – 0.36 |
| Frontal area ÷ (width × height) | 0.84 | 0.78 – 0.88 |
| Rolling-resistance coefficient | 0.011 | 0.008 – 0.014 |
| Loaded tire radius ÷ nominal radius | 0.965 | 0.950 – 0.980 |
| Driveline efficiency | 0.91 | 0.88 – 0.94 |
| Share of weight on the front axle | 0.61 | 0.57 – 0.64 |
| Center-of-gravity height (m) | 0.57 | 0.52 – 0.61 |
| Rotational inertia per axle (kg·m²) | 2.2 | 1.5 – 3.5 |
| Grip multiplier (tire and surface variation) | 1.00 | 0.92 – 1.08 |
Conditions
- Level road, no wind
- Air density 1.2 kg/m³
- No occupants or cargo added to the vehicle mass
- Closed-course, straight-line run from a standstill
- Thrust lines show what the powertrain can deliver after driveline losses, before the grip limit
- In-gear time: gear 3 held from 60 km/h to 100 km/h, same road and air
- Tire–road grip μ = 1.00
- Model version 0.3.0
The range is how far the result moves when the assumptions move. It is not a statistical confidence interval against measured results.
Electric motors: constant torque, then constant power
This site’s model treats an electric motor as the smaller of two limits: a constant maximum torque, available from a standstill, and a constant maximum power. At low speed the torque limit applies and thrust is flat. Above the speed where maximum torque × motor speed equals maximum power, the power limit applies and thrust falls in proportion to 1 ÷ speed.
The illustrative electric sedan has 600 N·m, 300 kW and a single 9.0:1 reduction gear. Its two limits meet at 4,775 rpm, which with its tires is 65 km/h (40 mph). Below that speed, more power with the same torque would not accelerate it any harder. Above it, more torque with the same power would not either.
With a single gear there are no change-up points to choose. The model leaves out battery and temperature limits, which a real electric car also has.
What this experiment leaves out
The torque curves here are not measured. The model draws a curve through the peak torque point, the peak power point and the rev limit by a fixed method, with two shape assumptions that are listed in the table above.
The model has no turbocharger lag, no torque-converter multiplication and no traction-control logic. A real engine’s response to the throttle, and a real driver’s gear changes, are not in it.