Skip to content
GitHubYouTubeInstagram

Watts in, watts out, and where the rest goes.curve → coils → limits → battery → does the motor matter

A motor is rated in thousands of watts. Our battery can give the whole robot about 2,640 watts. This course follows the amps from the battery into one wheel: the motor’s curve, what the brushes or the controller do to the coils, what a current limit takes away, and what is left when you swap a CIM for a Kraken on the same robot.

5 chapters · about 30 minutes

  1. 01 · 6 minOne motor’s curveTorque, current, power and efficiency against speed, for seven motors we have used or could.
  2. 02 · 6 minBrushes, coils and FOCWhat switches the current in the coils, and what FOC does at the motor.
  3. 03 · 5 minWhat a current limit takes awayA limit clips the curve. Where it cuts, and why supply current is a different limit from motor current.
  4. 04 · 6 minThe battery is the limitA sprint and a launcher spin-up on one battery, and the most power it can give before the roboRIO shuts the motors off.
  5. 05 · 7 minDoes the motor matter?The same robot with six CIMs, four NEOs or four Krakens, geared for the same top speed.
Chapter 1 of 5 6 min

One motor’s curve

Torque, current, power and efficiency against speed, for seven motors we have used or could.

Give a motor 12 V and hold the shaft still and it draws its most current and makes its most torque. That is stall. Let it spin with nothing on the shaft and it makes no torque at all. That speed is its free speed. Between the two, torque falls along a straight line. A spinning motor generates its own voltage, the back-EMF, which pushes against the battery, so the faster it turns, the less voltage is left to push current, and current is what makes torque.

Power is torque times speed, so it is zero at both ends and peaks in the middle. For every motor here the peak is at half of free speed, and the power band runs from 28% to 72% of it. The Kraken X60 with FOC is rated 9.37 N·m at 483 A and 5,800 rpm. The model gives it 1,417 W at 2,900 rpm. CTRE publishes 1,405 W.

LabCompare seven motors at 12 V

Pick a motor, then click or drag across the graph to read it at any speed. Find where the power peaks and how efficient the motor is there. Then compare the green hump with the red curve under it, and switch to All motors to see every motor's red curve on the same axes.

Power (band shaded)Torque, % of stall (current has the same shape)EfficiencyUsable: the most it makes on its battery shareOther motors, usable

Torque—
Current—
Power out—
Efficiency—
Heat in windings—

Simplified: a steady 12 V, windings that stay the same temperature (real ones lose strength as they heat), and the no-load current treated as friction. Motor numbers are WPILib's DCMotor checked against each maker, and the Krakens are shown with FOC, as we run them. See the note on the NEO below.

Look at the efficiency line. At peak power the X60 draws 243 A and is 49% efficient: for every 1,400 W it puts on the shaft, another 1,460 W becomes heat in the windings. Its best efficiency, 86%, comes at 5,460 rpm and 30 A, close to free speed, but it makes only a fifth of its peak power there. A robot that is accelerating lives between those two points, and the winding resistance decides how much of the current is wasted on the way. The X60's is 25 mΩ. A CIM's is 90 mΩ, so at its own peak power a CIM is only 40% efficient.

The red curve is the catch. The battery can give the whole robot about 2,640 W, so each of four brushless drive motors gets at most 660 W in, and each of six CIMs 440 W. The red curve is the most each motor can put on the shaft on that share. At its peak the X60 is 49% efficient, so its 1,417 W out would cost about 2,900 W from the battery. On 660 W it makes 555 W. Most of the X60's hump is out of reach on our robot.

Switch to All motors and the gap between the motors mostly closes. The X60's unlimited hump is three times a NEO's, but on the same 660 W the X60 makes 555 W, an X44 521 W, a Vortex 468 W and a NEO 402 W. That makes the X60 about 1.4 times a NEO, not 3. The reason is torque per amp. Every brushless motor here makes between 0.015 and 0.020 N·m per amp, so at the same current they push about the same. What sets the X60 apart is its low winding resistance, 25 mΩ against a NEO's 75, and that only matters at currents the battery cannot supply. What it actually buys you is less heat and a little more power at every current. Chapter 3 shows the same thing with current limits, and chapter 4 with the battery.

Hold an X60 at stall on 12 V and it draws 483 A and makes 5.8 kW of heat. That is why every motor here needs a current limit, and why the next chapters spend so long on limits and the battery.

The one thing to remember Every 12 V motor is a straight line of torque and a hump of power that peaks at half of free speed, and at that peak half of what it draws is heat.

Next: Brushes, coils and FOC ↓

Chapter 2 of 5 6 min

Brushes, coils and FOC

What switches the current in the coils, and what FOC does at the motor.

A motor turns because two magnetic fields pull on each other. One comes from magnets and the other from coils. The pull is strongest when the two fields are 90° apart, weakens with the sine of the angle between them, and is zero when they line up. As the rotor turns the angle shrinks, so the coil field has to keep moving ahead of it. Switching the current from coil to coil to make that happen is called commutation. The motors on this page differ in what does the switching and how finely it can aim.

Brushed: metal does the switching

In a CIM or MiniCIM the coils are on the rotor and the magnets are on the stator. Carbon brushes press on a ring of copper segments, the commutator, and flip the current from coil to coil as the rotor turns. The timing is fixed in the metal, so a brushed motor needs no sensor and no logic. A Talon SRX only sets the voltage.

What you pay: the brushes wear, they drop voltage that turns into heat, the coils spin in the middle of the motor where they are hardest to cool, and the switching happens in jumps. That is part of why the model's CIM peaks at 63% efficient and the Kraken X60 at 86%.

Brushless: a controller does the switching

Turn the motor inside out and the coils are on the stator, fixed to the case, and the magnets are on the rotor. With no brushes, something electronic has to know the rotor's angle and switch the coils. For the NEOs that is a SPARK controller reading sensors in the motor. For a Kraken it is the Talon FX built into it, reading the motor's own encoder. There is nothing to wear out or drop voltage, and the coils touch the case, so their heat can leave through it.

The controller can do the switching two ways. In trapezoidal commutation it uses six coil pairs per electrical turn, so the field can point in only six directions and the angle between the fields wanders up to 30° either side of the best 90°. In FOC it works out the current in every coil from the exact rotor angle, so the field stays 90° ahead. The Kraken supports both. A SPARK MAX or SPARK Flex drives the NEOs six-step, from three hall sensors in the motor. REV publishes no FOC mode.

Electrical turns and shaft turns

The coils don't care where the shaft is. They care which magnet pole is passing them. Picture a rotor with one north and one south magnet. Each coil sees north, then south, then north again once per shaft turn, and the controller runs through its six steps once. That cycle is one electrical turn. Now give the rotor seven north-south pairs around its rim, as a NEO has. Each coil sees a full north-south cycle every seventh of a turn, so the controller runs through its six steps seven times per shaft turn. That is 42 switches.

So the steps are counted per electrical turn, and that is also what sets the ripple: each switch is 60 electrical degrees, whatever the pole count. More pole pairs make the switches finer on the shaft, but no shallower. The lab draws the real magnets: one pair for the brushed drawing, seven for a NEO, all turning at the same shaft speed.

LabWatch the coil field chase the rotor

Green and gray are the magnets and orange is the coils and their field. In a brushless motor the magnets spin; in the brushed one the coils do. Every mode turns the shaft at the same speed, so compare how often each one switches in a turn. Pause it near a tick and read the torque.

Magnet (north)Coils and coil fieldTicks: every switch in one shaft turn. In the plot, the other two ways

Torque per amp now—
Average—
Worst angle—
Field can point—

Idealised: constant current and torque exactly proportional to the sine of the angle between the fields, so the lab shows torque ripple and nothing else. The brushed motor is drawn with one magnet pair and 12 steps, which is not a CIM's real count. Six-step and FOC are drawn on a NEO's rotor, seven magnet pairs.

Six steps cost 4.5% of the torque per amp on average, and 13% at the worst angle, right before the controller switches. That ripple is all the lab models. It is not the whole difference between the two, because CTRE's published X60 numbers change by more when it goes from trapezoidal to FOC: stall torque from 7.09 to 9.37 N·m, stall current from 366 to 483 A, and peak power from 1,108 to 1,405 W. Torque per amp is the same in both, 0.0194 N·m/A. FOC lets the motor reach more current at the same 12 V. What that is worth on our robot is a different question, and the FOC chapter of the CTRE course races it: about 1% in a 5 m sprint, with less heat.

So is a NEO worse than a CIM?

No. The lab shows only the ripple, and on that one number a brushed motor does look smoother. It gives up far more elsewhere. The brushes drop about a volt and a half, the rotor coils are hard to cool, and a CIM's windings have more resistance. The published numbers already include each motor's own switching, because they were measured on its controller. They include the NEO's six-step ripple, and the NEO still comes out ahead. It makes 0.0187 N·m per amp against the CIM's 0.0182, and peaks at 76% efficient against 63%. It does that at a third of the weight, 0.43 kg against 1.27. The steps are also finer than they look. The lower plot in the lab shows why that matters. Each six-step dip is as deep as a one-pair motor's would be, but a NEO has seven times as many, and each one is over in a seventh of the time. The shaft's own inertia smooths a short dip much more than a long one, so what the shaft feels is a speed wobble that shrinks as the dips get faster. At the same shaft speed, the NEO's 42 dips of 13% wobble it about 1.1 times as much as the brushed drawing's 12 dips of 3%. With one magnet pair it would have been about 8 times.

From worst to best: brushed, then six-step, then FOC. Six-step to FOC is the Kraken's edge, and it is a few percent. Brushed to brushless is the big step.

MotorCoils are onWhat switches themWhere the rotor angle comes from
CIM, MiniCIMThe rotorBrushes on the commutatorNot needed: the brushes sit at fixed angles
NEO, NEO 2.0, NEO VortexThe statorA SPARK MAX or SPARK Flex (six-step)Three hall sensors in the motor
Kraken X44, X60The statorThe Talon FX built into the motor (trapezoidal or FOC)The motor's built-in encoder

The one thing to remember A motor needs its coil field held about 90° ahead of the rotor. Brushes do it in fixed jumps, six-step does it in six, and FOC does it at every angle.

Next: What a current limit takes away ↓

Chapter 3 of 5 5 min

What a current limit takes away

A limit clips the curve. Where it cuts, and why supply current is a different limit from motor current.

A current limit cuts the top off the torque line. Torque follows current, so a 60 A limit holds an X60 to 1.1 N·m, about an eighth of the 9.3 it makes at stall. Where the line is already below the limit, nothing changes, so a limit takes a lot of torque at low speed and none near free speed. It also flattens the power hump, and it does that much more to a strong motor than to a weak one.

At 60 A a CIM keeps 98% of its peak power, a NEO 93%, a Vortex 79%, an X44 58% and an X60 42%. With no limit those five range from 330 W to 1,417 W. With the limit they range from 324 W to 598 W.

LabCut a motor's curve with a current limit

Pick one of our real setups from the menu, then drag Limit lower and watch the solid curve pull away from the dashed one. Try the same limit on a CIM and an X60.

Power with the limitTorque with the limitWithout the limit (dashed)

Peak power—
Stall torque—
Motor amps at stall—
Battery amps at stall—

Simplified: a straight limit at every speed. REV lets you lower the limit toward free speed and we do not model that. The battery is held at 12 V here, and the next chapter lets it sag. Motor numbers are from chapter 1.

Two limits, and REV's one

A Talon FX has two, and the CTRE course covers how they apply. The stator current limit caps the current in the windings, so it is the one that clips the torque line. The supply current limit caps the current from the battery. Switch the lab to Battery amps and the X60 at a 40 A limit still pushes 139 A through its windings at stall, and makes 2.7 N·m, because at low speed the windings take little voltage and so little battery power. Near the peak the two limits do about the same thing: 40 A motor amps and 40 A battery amps both leave the X60 about 410 W.

A SPARK MAX or SPARK Flex has one. REV's smartCurrentLimit lowers the controller's output voltage so the motor current stays under the number, and REV's locked-rotor tests describe holding constant phase current. So the battery button is off for every motor without a Talon FX. REV's parameter table lists an 80 A default at stall.

The model agrees with the makers where they give a number to check. CTRE lists the X60 at 413 W with 40 A, and the model gives 412 W. REV lists 375 W for the Vortex at 40 A, and it gives 363 W.

The one thing to remember A current limit costs the strongest motor the most, so at a tight limit a CIM and an X60 are much closer together than their datasheets say.

Next: The battery is the limit ↓

Chapter 4 of 5 6 min

The battery is the limit

A sprint and a launcher spin-up on one battery, and the most power it can give before the roboRIO shuts the motors off.

Two mechanisms want the battery at the same moment. The drive wants to cross the field, and the launcher wants a flywheel at 3,790 rpm before the first ball. Fitted to our Q36 match log, the battery is 12.38 V with nothing connected, and every amp we draw takes 14.4 mΩ of it: V = 12.38 − 0.0144 × amps. That is the internal resistance. A 100 A draw costs 1.4 V, and at 390 A the voltage is down to 6.75 V, where the roboRIO cuts power to the motors. That is a brownout.

Power is voltage times current, and voltage falls as current rises, so the battery has a peak. The most it can give before the brownout is 6.75 V × 391 A, about 2,640 W for the whole robot. Four X60s are rated for 1,400 W each and six CIMs for 330 W each. The battery can give the robot less than half of what four X60s are rated for, whatever motor sits in the wheel.

LabSprint 5 m and spin up the launcher at once

Run our 2026 limits, then pick No battery limit and No limits at all. Watch the voltage against the brownout line, and the power against the ceiling. Then turn the launcher off and see what it costs the drive.

DriveLauncherEverything else (10 A)Brownout and power ceiling
  • Stay above 6.75 V—No brownout at any moment of the run.
  • Launcher at speed by 0.5 s—Within 200 rpm of 3,790.
Lowest voltage—
Most power—
Time to 5 m—
Launcher ready—
Set the limits yourself

Amps in each drive motor's windings. 500 means no limit.

Amps each drive motor may draw from the battery. 130 means no limit.

Simplified: straight line, perfect traction up to the slip current, the offseason launcher spinning from idle to 3,790 rpm on its own logged current limits, and everything else on the robot at a steady 10 A. One battery, the one in our Q36 log. Drive: 60 kg, four X60s at 6.03:1, 1.964-inch wheels, from 2026-Spectrum. The bus follows the draw over a few milliseconds, and below 6.75 V every motor is off for 20 ms.

With our real limits the drive alone bottoms out at 8.5 V. Add the launcher and the lowest voltage is 6.94 V, 0.19 V above the brownout, with the battery giving 2,629 W of its 2,639. The two loads land on top of each other because the drive is at its hardest for the first 0.3 s and the launcher, at its 80 A limit, is spinning up over the same 0.3 s. Take away the battery limit and it browns out twice. Take away all the limits and it browns out about 40 times, because every motor tries to draw its stall current from a battery that has 391 A to give.

Once the launcher is at speed it draws about 18 A, the drive's draw dies away as the robot reaches speed, and the battery recovers to 11.9 V. The trouble is the first 0.3 s, when both mechanisms are at their limits together. The current limits are what divide the battery's 2,640 W between them.

The one thing to remember The battery can give the whole robot about 2,640 W before the roboRIO cuts the motors, and every mechanism's current limit is a share of that.

Next: Does the motor matter? ↓

Chapter 5 of 5 7 min

Does the motor matter?

The same robot with six CIMs, four NEOs or four Krakens, geared for the same top speed.

Chapter 4 said the battery gives the whole robot about 2,640 W, whatever motor is in the wheel. This chapter tests that: the same 60 kg robot, with each of the seven motors in place of our four X60s.

Each motor gets the fair version of our robot. It is geared so its free speed is the robot's 5.0 m/s top speed. The push is capped where the wheels slip, at the grip CTRE's 90 A slip current implies for our robot (a friction coefficient of 1.43), so no motor gets more push than the carpet allows. The brushed motors use six, as we ran them, and the brushless ones four. The Krakens keep our 65 A battery limit and the others have none, because a SPARK has only a motor limit. The robot's mass changes with the motors: six CIMs weigh 7.6 kg, and four X60s 2.2 kg.

LabRace seven motors 5 m on the same robot

Watch the seven lanes, then read the table. Try a slippery floor and a grippy one. Then switch to 4 of each and take away the Krakens' battery limit.

Motor counts and battery limits

MotorTime to 5 mLowest batteryBecame speed

Simplified: straight line, perfect traction up to the slip limit, no friction beyond the motors' own, motor weight only (a CIM's gearbox and a Kraken's are not counted), the same battery as chapter 4, and ideal gear ratios that real gearboxes only come near. Motors from chapter 1. Robot from 2026-Spectrum: 60 kg with four X60s, 6.03:1, 1.964-inch wheels.

On our floor every drive pulls between 2,290 W and 2,660 W from the battery at its peak, though the motors are rated from 210 W to 1,400 W each. That is chapter 4's ceiling showing up. The race is still not a tie: the X60 drive reaches 5 m in 1.23 s and the MiniCIM drive in 1.44 s, 17% slower, on about the same power.

The difference is how much of the energy reaches the wheels. 79% of what the X60 drive took from the battery became the robot's speed, against 49% for the MiniCIMs. The rest is heat in the windings: 200 J over the X60 run and 650 J over the MiniCIM run. That is chapter 1's efficiency curve, seen in a race. A lower winding resistance wastes less of the same amps.

Among the brushless motors the gap is smaller. From the X60 to the NEO it is 8% on our floor, and on a slippery floor under 4%, because the floor decides how hard the wheels can push before the motor gets a say. Six CIMs with only a traction limit take the battery down to 6.8 V and brown out twice in 5 m, which is the drive we used to run. The drive code in 2018 (Infrared-2018) and 2019 (X-Ray-2019) both set current limits.

Motor choice left the power the robot could draw where it was, between 2,290 W and 2,660 W in every run. It changed how much of that reached the wheels, and here that was worth 17% over 5 m from the MiniCIM to the X60, and less than 8% among the brushless motors.

The one thing to remember All seven drives took about the same power from the battery, so the gaps came from how much of it reached the wheels: 17% over 5 m from a MiniCIM to an X60, and less than 8% among the brushless motors.

QuizCheck yourself
  1. 1. At what speed does a DC motor on 12 V make its most power?

  2. 2. At its peak power the X60 is about 49% efficient. Where does the other half of the power go?

  3. 3. What does FOC change inside a Kraken?

  4. 4. How does a CIM switch current from coil to coil with no sensor and no controller logic?

  5. 5. You put a 60 A motor limit on each of a CIM and an X60. Which loses more peak power?

  6. 6. An X60 is stalled against a wall with a 40 A battery (supply) limit. How much current is in its windings?

  7. 7. About how much electrical power can our battery give the whole robot before the roboRIO browns out?

  8. 8. With our real limits, the drive sprints 5 m while the launcher spins up. What happens to the battery?

  9. 9. The X60 and MiniCIM drives draw about the same power from the battery, but the X60 finishes 5 m 17% sooner. Why?

That's the Motors and Power course.

Next up: CTRE Control Modes, or head back to all courses.