How to Test E-Bike Hall Sensors (and When It Means a New Motor)

An opened hub motor on a light wood bench showing copper stator windings, a bundle of red, yellow, green, blue and black wires emerging from the axle, and the aluminium side cover lying beside it.

A hall sensor fault is one of the few e-bike problems that feels dramatic and is still cheap to diagnose. The motor shudders, lurches, or refuses to spin, and the cause ranges from a corroded connector to a motor that needs replacing. That spread is too wide to guess at, and a multimeter closes it in about ten minutes.

What the hall sensors actually do

A brushless hub motor has no brushes to switch current mechanically, so the controller must know where the rotor is before energizing the right winding. Three hall-effect sensors sit inside the motor, spaced around the stator, each reporting magnet-present or magnet-absent as the wheel turns. Lose one and the timing goes wrong several times per revolution — which is exactly what a stutter is.

They live in the thin wire bundle, not the thick one. A standard hub motor leaves the axle with three heavy phase wires carrying drive current, plus five thin sensor wires: red for the supply, black for ground, and yellow, green and blue for the three signals. A ninth wire, when present, is usually a temperature sensor — see the phase and hall wiring guide.

Symptoms that point here

Hall faults have a signature worth recognizing before you open anything:

  • Strong but jerky. The motor clearly has power, but it shudders, knocks, or stutters instead of pulling smoothly.
  • Rough from a standstill, fine once rolling. Some controllers fall back to a sensorless mode when signals go missing: it works at speed but is weak and lumpy off the line.
  • Uneven lurching pulses. A motor that surges in irregular bursts rather than accelerating smoothly.
  • Dead in one wheel position. Nudge the wheel a few inches and it suddenly runs, which points at one sensor rather than all three.

If your display shows a fault code, write it down first. Codes are not standardized across brands, but many displays have a code that points at motor hall faults — the SW900 code table is one example.

Rule out the cheap causes first

Two things fail far more often than the sensors themselves, and both cost a fraction of a motor.

The first is the connector. Water in a multi-pin plug corrodes the small signal pins long before it bothers the fat phase pins, so a bike that lived outdoors deserves an inspection before a diagnosis. The second is wiring order: if the stutter started right after fitting a different motor or controller, you likely have a mismatched combination rather than a dead part, and the fix is the 36-combination procedure, not a new motor.

The test: five wires, one meter, one slow turn

Work with the wheel off the ground and the bike powered on but not in gear-engaging mode. Set the meter to DC volts. Back-probe the sensor connector on the controller side so the circuit stays connected while you measure.

  1. Check the supply. Black probe on the sensor ground, red probe on the supply wire. Expect a steady low-voltage rail — commonly around 5 V, though not every system uses the same one, so confirm your own rather than assuming. No supply means stop here: the fault is upstream.
  2. Watch each signal. Leave the black probe on ground, put the red probe on the first signal wire, and turn the wheel by hand slowly — one revolution over several seconds.
  3. Look for the toggle. A working sensor swings between roughly zero and roughly the supply voltage, several times per revolution. You are looking for movement, not a specific number.
  4. Repeat for the other two signal wires. All three should toggle, and they should toggle a similar number of times per revolution.

Reading what you get

One wire stuck at zero or stuck high while the other two toggle: that sensor has failed. All three flat while the supply is present: suspect the sensor ground or a break in the bundle before condemning the motor, because three simultaneous failures are rare. All three flat with no supply present: the problem is the controller or the harness, not the motor. Everything toggles cleanly: the halls are fine, and the stutter is coming from somewhere else — phase wiring, the controller, or a battery that sags under load.

When it does mean a new motor

The sensors are inside the motor. Replacing them means pulling the wheel, opening the side cover, and working on the stator with a soldering iron — real work, but not exotic work. Whether it is worth doing depends on labor cost against the price of a replacement wheel, and on whether the winding itself is healthy.

There is also a middle path: some controllers run a sensorless mode that ignores the hall signals entirely. It is a genuine option when the sensors are dead but the windings are good, with a real trade-off — starts are lumpy and low-speed pull is weak, while higher-speed running is normal. For a rider who mostly cruises that can be an acceptable repair; for someone starting on hills with cargo it usually is not.

What the meter buys you is the difference between those choices: ten minutes of probing separates a corroded plug from a failed sensor from a healthy motor with a wiring fault — usually the difference between buying one part and buying three.