Hall effect sensor: how it works, types and uses

A Hall effect sensor is a small chip that detects a magnetic field and turns it into a signal a microcontroller can read. Put a magnet on a door, a wheel or a rotor, put the sensor next to it, and the chip tells you whether the magnet is near, how strong its field is, or which way it points. Because nothing touches, it has no contacts to wear, it can be sealed away from water and dirt, and it can count fast. This guide explains the physics in plain words, what is inside the chip, the main types, how to wire each one, where they are used, and how they compare with a reed switch. It sits with the sensors and hardware guides, next to the overview of IoT sensors and the microcontroller choice guide.

The Hall effect in plain words

Pass a current through a thin, flat plate of conductive material, then hold a magnet so its field passes through the plate at right angles. The field pushes the moving charges sideways, so they crowd toward one edge of the plate. One edge ends up with extra negative charge and the opposite edge with less, and that difference is a voltage across the width of the plate. It is called the Hall voltage, and it grows with the strength of the field.

Allegro MicroSystems’ Hall-effect IC applications guide describes the effect as discovered by Hall in 1879. Edwin Hall reported it in “On a New Action of the Magnet on Electric Currents”, published in the American Journal of Mathematics in 1879. Texas Instruments’ introduction, What Is a Hall-effect Sensor?, gives the same picture: a small voltage develops across the conductor as current is forced through it and a magnetic field is applied in an orthogonal direction.

The catch is size. TI notes that the output of a bare Hall element is tiny, sometimes in the microvolt range, which is why nobody wires a raw element to a microcontroller.

What is inside a Hall sensor chip

A Hall sensor IC (integrated circuit) puts the element and its support circuit on one piece of silicon. Allegro’s guide lists the parts of a digital Hall switch: a Hall voltage generator, a signal amplifier, a Schmitt trigger and an output transistor, which is usually an open-collector NPN or an N-channel FET. Linear versions replace the trigger with an analog output stage and add a stable amplifier and voltage regulator.

The Schmitt trigger is a comparator with two thresholds. The output switches on when the field rises past the operate point, called BOP, and switches off only when the field falls below a lower release point, BRP. The gap between the two is the hysteresis, and TI’s article explains that it stops the output bouncing back and forth when the field sits near a threshold.

The simplest parts have three pins, for supply, ground and output, in a small surface-mount SOT-23 or a through-hole TO-92 package. Direction matters: Allegro says the field must cross the active area at right angles, and TI’s DRV5055 linear sensor senses flux perpendicular to the top of the package. TI notes that most datasheets give the expected direction of the field relative to the package surface.

Types of Hall effect sensor

TI’s article groups position sensors into three families, and two more types matter for IoT projects.

Unipolar switch. Turns on when one magnetic pole, usually south, comes close enough to pass BOP, and turns off when it moves away past BRP. Allegro describes a unipolar part as having its switch points entirely in the south range. The other pole does nothing.

Omnipolar switch. The same, but either pole turns it on, which means you do not have to care which way round the magnet is fitted. TI’s DRV5032 comes in omnipolar and unipolar versions.

Latch. Turns on with a south pole and stays on after the magnet leaves. According to Allegro, it releases only when a north pole comes past. That makes latches suit alternating poles, such as a ring magnet on a shaft.

Linear. Gives an analog voltage that follows the field. TI’s article explains that with no field the output sits at half the supply voltage, then moves toward the supply for one pole and toward ground for the other until it saturates. The DRV5055 is ratiometric: its datasheet says the resting voltage and the sensitivity scale with the supply voltage.

3D and angle sensors. Several Hall elements on one chip measure the field along three axes and report numbers over a digital bus. TI’s TMAG5273 has three Hall sensors in the X, Y and Z axes, a 12-bit analog-to-digital converter (ADC) and an angle engine that gives a full 360 degrees of rotation, read over I2C.

Current sensors. A conductor carrying current makes its own magnetic field, so a Hall chip can measure the current from that field. Allegro’s guide calls Hall devices excellent for limiting or measuring current, and notes that the conductor does not have to be broken. Allegro says its current sensor ICs deliver galvanic isolation. The metering circuit in a smart plug uses a low-value shunt resistor or a current transformer instead.

How to wire a Hall sensor to a microcontroller

Start with the datasheet for the exact part, because pin order and output type vary between parts that look identical.

Digital switch or latch. Connect supply and ground, and take the output to a GPIO (general-purpose input/output) pin. The DRV5032 datasheet shows the usual behavior: when the field exceeds BOP the output goes low, and it stays low until the field falls below BRP. Some versions have push-pull outputs that drive the pin high or low themselves. Others have open-drain outputs, which only pull low, and for those the datasheet says an external pull-up resistor must be used. The same datasheet points out that an open-drain output leaks current through the pull-up while it is low, and that push-pull gives the lowest system power. Because the output is a clean edge, you can attach an interrupt to it and let a battery node sleep until the magnet moves, as the guide to powering IoT sensors describes.

Linear analog. Take the output to an ADC input. The DRV5055 datasheet says the ratiometric design can eliminate error from supply tolerance when the ADC uses the same supply as its reference, so power the sensor from the same rail as the microcontroller’s ADC. That also keeps the output inside the range the ADC pin can read on a 3.3 V board, since the output swings with the supply. Its datasheet gives the linear output range as 0.2 V up to 0.2 V below the supply.

Digital 3D sensor. Connect it to the I2C bus like any other digital sensor, with the usual pull-up resistors on the two bus lines; the I2C vs SPI guide explains how the bus works. The TMAG5273 datasheet lists a 1.7 to 3.6 V supply, a maximum I2C clock of 1 MHz, factory-programmed and user-configurable addresses, and an interrupt signal when the field or temperature crosses a threshold you set.

What Hall effect sensors are used for

Door, lid and cover detection. TI’s article lists switches in laptops and refrigerator doors, and its application note on replacing reed switches is written around battery-operated door and window sensors for home security. A magnet in the door and a switch in the frame report open or closed.

Flow meters. A common type of small water flow sensor turns a wheel in the flow and senses it with a Hall sensor. DATAQ Instruments’ flow sensor datasheet describes a pinwheel in the water line and a Hall sensor that outputs a pulse with every revolution, sealed from the water. Count the pulses with an interrupt and convert to volume with the factor from the datasheet.

Wheel, fan and shaft speed. Allegro lists latches in shaft encoders, speedometers and tachometers. Count pulses over a fixed time and divide by the number of pulses the sensor gives per turn to get the speed.

Brushless motor commutation. A brushless motor’s controller has to know the rotor position to switch the windings at the right moment, and both Allegro and TI name motor commutation as a common job for latches.

Joysticks and keyboards. TI’s joystick and lever design note says magnets and Hall sensors remove contact-based components that wear out over time. Its keyboard application note says mechanical key switches degrade from friction and can only give an on or off response, while a Hall sensor sees the magnet move closer as the key goes down.

To detect people in a room, see the mmWave vs PIR guide.

Hall effect sensor vs reed switch

A reed switch is the older way to sense a magnet. TI’s reed switch note describes it as a glass tube filled with inert gas holding two ferromagnetic blades, which touch when a magnet comes near. Both do the same job in a door sensor and differ in most other ways.

Power is the most argued point. Magnetics Magazine explains that a reed switch that receives no magnetic pulse does not consume any power, while a Hall sensor needs a constant supply of power. TI’s note adds that a reed circuit can still draw current through its pull-up or pull-down resistor, depending on how it is wired. Sampled Hall switches narrow the gap: the DRV5032 averages 0.54 µA (microamps) in its 5 Hz version, according to its datasheet.

The same article, quoting a Standex product manager, puts the Hall sensor ahead for high-speed use, in the range of 1 kHz and above, where reed contacts reach their physical limits. It also says even the smallest reed switches can isolate voltages of up to 1,000 volts. A reed switch is a contact, so it can switch a circuit directly, while a Hall output drives a logic input. TI’s tests found a low-power Hall switch had a tighter, more symmetrical detection range than the reed switch in its comparison. For a battery door contact either works; for counting, motors or anything that moves all day, the Hall sensor is the better fit.

Downsides of Hall effect sensors

It needs a magnet, placed with care. TI’s article shows how fast a magnet’s field decays with distance, and advises keeping the working range in the region where the field changes most. A misaligned magnet or an extra millimeter of gap can leave a switch that never operates.

Temperature. Allegro notes that extremes of temperature shift the operate and release points, and its table gives negative temperature coefficients for most of the magnet materials it lists, so those magnets weaken as they warm. Some linear parts correct for this: the DRV5055’s A versions compensate for magnet temperature drift and its Z versions do not.

Stray fields. Any other magnet near the sensor counts as a signal. TI’s reed switch note looked at how easily an external magnet could fool a door sensor, and found that 3D sensors can detect tamper attempts and stray fields, because their three-axis reading moves away from its normal resting value when another magnet comes near.

Constant supply current for most parts. Honeywell’s SS49E linear sensor draws 6 mA typical at 5 V, and the DRV5055 datasheet gives 3 mA typical at 5 V. Micropower parts get down to microamps by sampling the field at a low rate: the DRV5032 updates its output at 80 Hz, 20 Hz or 5 Hz depending on version. A 5 Hz part can miss a magnet that passes quickly. The TMAG5273 lists 2.3 mA active and 1 µA in its wake-up and sleep mode.

The original ESP32’s built-in Hall sensor

The original ESP32 had a Hall sensor on the chip, so a project could sense a magnet held near the module with no extra parts. Espressif’s ESP-IDF v4.4 ADC documentation says reading it used channels 0 and 3 of ADC1 (GPIO 36 and 39), through a function called hall_sensor_read(), and that nothing else should be connected to those pins.

ESP-IDF 5.0 dropped it. The v5.0 peripherals migration guide says the hall_sensor_read API on ESP32 has been removed and the Hall sensor is no longer supported. If a tutorial calls that function, it was written for an older framework. For current projects on any ESP32, including the ESP32-C6, use an external Hall sensor wired as above.

How to pick one for a project

1. Decide what you need to know. Present or absent: a switch. Rotation with alternating poles: a latch. Position or distance: a linear sensor. Angle or position in more than one direction: a 3D sensor.
2. Match the threshold to the magnet and gap. Switches come with different operate points. The DRV5032 is sold in versions with maximum BOP values from 3.4 mT (millitesla) to 63 mT, and a sensitive version lets a smaller magnet work from further away.
3. Check the power budget. On a battery, pick a sampled micropower switch or a 3D sensor with a sleep mode, and check that the sampling rate is fast enough for what you are watching.
4. Check the output type and supply voltage against your board: open-drain needs a pull-up, analog needs an ADC on the same supply, I2C needs a free address.
5. Try the real gap before fixing anything. Check the magnet at the distances and angles it will see in use, including the worst case.

For more sensor types and how their readings leave the board, go back to the sensors and hardware guides or read the IoT sensors overview and the guide to powering IoT sensors.