www.ti.

com Table of Contents

Application Note
Linear Hall-Effect Sensor Angle Measurement Theory,
Implementation, and Calibration

Mitch Morse Current and Magnetic Sensing

ABSTRACT
This application note discusses how two one-dimensional linear Hall-effect sensors can be used to measure
2D angles, including both limited-angle and 360° rotation measurements. This document provides details on
some calibrated and uncalibrated implementations to help meet angle measurement accuracy requirements.
This report also covers the number of sensors needed, and the preferred magnet types for each method. For
details on how one 3D linear Hall-effect sensor can be used to measure 2D angles, see the Angle Measurement
With Multi-Axis Linear Hall-Effect Sensors application report.

Table of Contents
1 Introduction.............................................................................................................................................................................1
2 Overview..................................................................................................................................................................................2
2.1 Types of Magnetization...................................................................................................................................................... 2
2.2 Types of Magnets...............................................................................................................................................................2
3 Device Descriptions............................................................................................................................................................... 4
3.1 2.5-V to 38-V, Bipolar Hall Effect Sensor Family: DRV5053 and DRV5053-Q1.................................................................4
3.2 High-Accuracy, 3.3-V or 5-V, Ratiometric, Bipolar Hall Effect Sensor Family: DRV5055 and DRV5055-Q1..................... 4
3.3 High-Accuracy, 3.3-V or 5-V, Ratiometric, Unipolar Hall Effect Sensor Family: DRV5056 and DRV5056-Q1................... 4
4 Methods................................................................................................................................................................................... 5
4.1 Uncalibrated Implementations............................................................................................................................................6
4.2 Peak Calibrated Implementations.................................................................................................................................... 12
4.3 Lookup Table Calibration Implementations...................................................................................................................... 15
4.4 Peak Calibrated Plus Lookup Table Hybrid......................................................................................................................19
5 References............................................................................................................................................................................ 23
6 Revision History................................................................................................................................................................... 23

Trademarks
All trademarks are the property of their respective owners.
1 Introduction
Linear Hall effect sensors measure the strength of a magnetic field and output a voltage proportional to that
measurement. Based on the degree range and resolution needed, one or more linear Hall sensors can be used
to determine the magnet direction. This application report covers angle measurements using no calibration, peak
calibration, lookup table calibration, and a hybrid method of both the peak calibrated and lookup table methods.

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 1
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Overview www.ti.com

2 Overview
2.1 Types of Magnetization
The two main types of magnetization in permanent magnets are axial and diametric. This terminology makes
most sense when talking about discs, cylinders, and ring magnets. Axial magnets have north and south poles
that are on the flat surfaces of the magnet. Diametric magnets have north and south poles that are on the
rounded edges of the magnet.
Some examples of axially magnetized magnets are the two left magnets in Figure 2-1 and the two left magnets
in Figure 2-2.
Some examples of diametrically magnetized magnets are the two right magnets in Figure 2-1 and the two right
magnets in Figure 2-2.
Other magnet types are typically referred by shape, such as block and sphere magnets (Figure 2-3 and Figure
2-4), or by unique polarity, for example a multipole ring magnet (Figure 2-5).
2.2 Types of Magnets

Figure 2-1. Disc and Cylinder Magnets

Figure 2-2. Ring Magnets

Figure 2-3. Block Magnets

2 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Overview

Figure 2-4. Sphere Magnet

Figure 2-5. Multipole Ring Magnet

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 3
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Device Descriptions www.ti.com

3 Device Descriptions
When using linear Hall effect sensors to measure angles, a bipolar sensor is generally most practical to use,
although unipolar sensors can still be used for limited-angle measurements. Bipolar sensors respond to both
the north and south poles of a magnet, and allow for wider-angle measurements. Unipolar sensors respond to
one pole of the magnet allowing for only half of the movement range. The following subsections list some of the
linear Hall effect devices from TI.
3.1 2.5-V to 38-V, Bipolar Hall Effect Sensor Family: DRV5053 and DRV5053-Q1
The DRV5053 is a chopper-stabilized Hall effect sensor that offers a magnetic sensing solution with superior
sensitivity stability over temperature and integrated protection features.
The 0-V to 2-V analog output responds linearly to the applied magnetic flux density, and distinguishes the
polarity of magnetic field direction. A wide operating voltage range of 2.5 V to 38 V with reverse polarity
protection up to –22 V makes this device suitable for a wide range of industrial and consumer applications.
Internal protection functions are provided for reverse-supply conditions, load dump, and output short circuit or
overcurrent.
The DRV5053-Q1 is the automotive-grade version of the DRV5053.
3.2 High-Accuracy, 3.3-V or 5-V, Ratiometric, Bipolar Hall Effect Sensor Family: DRV5055 and
DRV5055-Q1
The DRV5055 is a linear Hall effect sensor that responds proportionally to magnetic flux density. This device can
be used for accurate position sensing in a wide range of applications.
The device operates from 3.3-V or 5-V power supplies. When no magnetic field is present, the analog output
drives half of VCC. The output changes linearly with the applied magnetic flux density, and four sensitivity
options enable maximal output voltage swing based on the required sensing range. North and south magnetic
poles produce unique voltages.
Magnetic flux perpendicular to the top of the package is sensed, and the two package options provide different
sensing directions.
The device uses a ratiometric architecture that can eliminate error from VCC tolerance when the external
analog-to-digital converter (ADC) uses the same VCC as a reference. Additionally, the device features magnet
temperature compensation to counteract magnet drift for linear performance across a wide –40°C to +125°C
temperature range.
The DRV5055-Q1 is the automotive-grade version of the DRV5055.
3.3 High-Accuracy, 3.3-V or 5-V, Ratiometric, Unipolar Hall Effect Sensor Family: DRV5056 and
DRV5056-Q1
The DRV5056 is a linear Hall effect sensor that responds proportionally to flux density of a magnetic south pole.
The device can be used for accurate position sensing in a wide range of applications.
The devices features a unipolar magnetic response. The analog output drives 0.6 V when no magnetic field is
present, and increases when a south magnetic pole is applied. This response maximizes the output dynamic
range in applications that sense one magnetic pole. Four sensitivity options further maximize the output swing
based on the required sensing range.
The device operates from 3.3-V or 5-V power supplies. Magnetic flux perpendicular to the top of the package is
sensed, and the two package options provide different sensing directions.
The device uses a ratiometric architecture that minimizes error from the VCC tolerance when the external
analog-to-digital converter (ADC) uses the same VCC as a reference. Additionally, the device features magnet
temperature compensation to counteract magnet drift for linear performance across a wide –40°C to +125°C
temperature range.
The DRV5056-Q1 is the automotive-grade version of the DRV5056.

4 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4 Methods
Table 4-1 shows a summary of the angle measurement methods discussed in this application report. For the column labeled Magnet Placement
Orientation Required?, the term Approximately means that the magnet must be oriented during placement, but not very precisely. For more information
about each method, see the associated sections linked in Table 4-1.

Note
When trying to achieve high accuracy and resolution with 360° rotation, the Peak + Lookup Hybrid method is easier to implement than the
standard lookup table. The Peak + Lookup Hybrid method is used for the DRV5055-ANGLE-EVM.

Table 4-1. Angle Measurement Summary

Magnet Placement Orientation # Sensors Estimated Accuracy
Accuracy
Recommended Magnet Required? Needed Peak-to-Peak Error
Calibration Method Improved by Complexity
Options (Based on measured data with a
1 Sensor 2+ Sensors Adding: < 180° 360° DRV5055)
Diametrically magnetized disc or
Uncalibrated axially magnetized cylinder or Yes Yes Sensors 1+ 2+ = 180° / #Sensors Low
block
Peak Calibrated Diametrically magnetized disc Yes Yes N/A 1 2 ≈ 8° Low
Lookup Table Diametrically magnetized disc Approximately No Calibration points 1 2 ≈ (Spacing Between Cal Points) / 8 High
Peak + Lookup Hybrid Diametrically magnetized disc Approximately No Calibration points 1 2 ≈ (Spacing Between Cal Points) / 15 Medium

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and Calibration 5
Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.1 Uncalibrated Implementations

4.1.1 Overview
4.1.1.1 General Implementation
The peak amplitude of the signal is unknown with an uncalibrated system. Therefore, the only usable information
from the sensor is whether VOUT is greater or less than VVCC / 2, as shown in Figure 4-1. This information
indicates whether the magnet is pointing towards a degree range (or region) where the sensor is sensing more
north or more south polarity. The number of regions for a system depends on the number of sensors used.

Figure 4-1. Uncalibrated Sensor Positions

4.1.1.2 Preferred Magnet Types

• Diametrically magnetized disc or cylinder
• Axially magnetized cylinder
• Block magnet
4.1.1.3 General Accuracy and Resolution
• Low accuracy
• Low resolution
• Results come in the form of general regions
• Accuracy can be improved by adding sensors
4.1.1.4 Considerations
• The magnet must be oriented to align desired regions.
• The boundary line for each sensor is at VOUT = VVCC / 2. If VVCC / 2 is measured, then either side of the
boundary line may be chosen as the measured region.
• The uncalibrated implementations discussed here are for 360° rotation. For smaller ranges of movement,
fewer regions are available.

6 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.1.2 One Bipolar Sensor, Uncalibrated

4.1.2.1 Specific Implementation
With one sensor, as shown in Figure 4-2, the sensor output voltage takes the form shown in Figure 4-3. Because
there is no calibration phase, the peak amplitude is unknown.

Figure 4-2. One Sensor Near Magnet

3.3

2.8875

2.475
Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D001

Figure 4-3. One Sensor Uncalibrated Data

4.1.2.2 Calculating Region

To determine the region that the magnet points towards, measure to see if VOUT is greater or less than VVCC / 2,
as shown in Table 4-2.
Table 4-2. One Sensor Regions
VOUT Region
> VVCC / 2 0° to 180°
< VVCC / 2 180° to 360°

4.1.2.3 Accuracy
The accuracy for this setup is the size of each region, 180°.

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 7
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.1.3 Two Bipolar Sensors 90° Apart, Uncalibrated

4.1.3.1 Specific Implementation
With two sensors 90° apart, as shown in Figure 4-4, the sensor output voltage takes the form shown in Figure
4-5. There is no calibration phase, so the peak amplitude is unknown; therefore, an example amplitude is shown.

Figure 4-4. Two Sensors 90° Apart

3.3

2.8875

2.475
Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D002

Figure 4-5. Two Sensors 90° Apart Uncalibrated Data

4.1.3.2 Calculating Region

To determine the region that the magnet points towards, measure VOUT for each sensor to see if VOUT is greater
or less than VVCC / 2, as shown in Table 4-3.
Table 4-3. Two Sensors 90° Apart Regions
VOUT 1 VOUT 2 Region
> VVCC / 2 > VVCC / 2 0° to 90°
> VVCC / 2 < VVCC / 2 90° to 180°
< VVCC / 2 > VVCC / 2 180° to 270°
< VVCC / 2 < VVCC / 2 270° to 360°

4.1.3.3 Accuracy
The accuracy for this setup is the size of each region, 90°.

8 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.1.4 Two Bipolar Sensors n° Apart, Uncalibrated

4.1.4.1 Specific Implementation
When the two sensors are not 90° apart, the regions sizes are no longer the same, and instead depend on the
degree (n) between the sensors. With two sensors that are n° apart, as in Figure 4-6, the sensor output voltage
takes the form shown in Figure 4-7. Both of these images use n = 45° as an example. There is no calibration
phase, so the peak amplitude is unknown; therefore, an example amplitude is shown. To avoid losing the benefit
of two sensors, n cannot approximately equal either 0° or 180°.

Figure 4-6. Two Sensors 45° Apart

3.3

2.8875

2.475
Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D003

Figure 4-7. Two Sensors 45° Apart Uncalibrated Data

4.1.4.2 Calculating Region

To determine the region that the magnet points towards, measure VOUT for each sensor to see if VOUT is greater
or less than VVCC / 2, as shown in Table 4-4.
Table 4-4. Two Sensors 45° Apart Regions
VOUT 1 VOUT 2 Region for n Region at n = 45°
> VVCC / 2 > VVCC / 2 0° to (180 – n)° 0° to 135°
> VVCC / 2 < VVCC / 2 (180 – n)° to 180° 135° to 180°
< VVCC / 2 > VVCC / 2 180° to (360 – n)° 180° to 315°
< VVCC / 2 < VVCC / 2 (360 – n)° to 360° 315° to 360°

4.1.4.3 Accuracy
The accuracy for this setup depends on the size of the current region. Out of the four regions, two regions have
an accuracy of n° and the other two regions have an accuracy of (180 – n)°.

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 9
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.1.5 Three or More Bipolar Sensors, Uncalibrated

4.1.5.1 Specific Implementation
With s number of sensors n° apart, the sensor system output varies for every setup, but produces s × 2 regions.
For evenly spaced regions, place the sensors so that n = (180 / s)° apart. For example, Figure 4-8 and Figure
4-9 use s = 3 sensors and n = (180 / 3) = 60° and produces 3 × 2 = 6 regions. Region sizes can be adjusted by
changing the degree n between each sensor. There is no calibration phase, so the peak amplitude is unknown;
therefore, an example amplitude is shown. To avoid losing the benefit of each sensor, n between any given two
sensors cannot approximately equal either 0° or 180°.

Figure 4-8. Three Sensors 60° Apart

3.3

2.8875

2.475
Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D004

Figure 4-9. Three Sensors 60° Apart Uncalibrated Data

10 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.1.5.2 Calculating Region

To determine the region that the magnet points towards, measure VOUT for each sensor to see if VOUT is greater
or less than VVCC / 2 while considering the degree n between each sensor. The regions from the example where
s = 3 sensors and n = (180 / 3) = 60° are shown in Table 4-5. To calculate the regions when using any other
value for n, adapt the equations shown in Table 4-4 to adjust for the number of sensors (s) and spacing (n).
Table 4-5. Three Sensors 60° Apart Regions
VOUT 1 VOUT 2 VOUT 3 Region
> VVCC / 2 > VVCC / 2 > VVCC / 2 0° to 60°
> VVCC / 2 > VVCC / 2 < VVCC / 2 60° to 120°
> VVCC / 2 < VVCC / 2 < VVCC / 2 120° to 180°
< VVCC / 2 < VVCC / 2 < VVCC / 2 180° to 240°
< VVCC / 2 < VVCC / 2 > VVCC / 2 240° to 300°
< VVCC / 2 > VVCC / 2 > VVCC / 2 300° to 360°

4.1.5.3 Accuracy
The accuracy for this setup depends on the current region and the value of n. When using evenly spaced
regions, where n = (180 / s)°, the accuracy for each region is (180 / s). To determine the accuracy when using
any other value for n, adapt the method described in Section 4.1.4.3 to adjust for the number of sensors (s) and
spacing (n).

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 11
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.2 Peak Calibrated Implementations

4.2.1 Overview
4.2.1.1 General Implementation
With a peak-calibrated system, bipolar sensor data can be normalized to ±1 for use with the arctan2 (two
sensors) or arcsin (one sensor) function in order to determine the angle. Arctan2 must be used instead of arctan,
because arctan2 accounts for which of the two values are negative. The process for calibration is:
1. Find the min and max values from each sensor by continuously reading voltages while rotating the magnet
360°. One full rotation is required, but more rotations help make sure that more accurate min and max
values are found.
2. Then, during normal operation, each new measured voltage can be normalized to ±1 using Equation 1.
Vmeasured − Vmin − Vamplitude V −V
NORM = Vamplitude wℎere Vamplitude = max 2 min (1)

3. The normalized data is then put directly into the arctan2 (two sensors, 0° to 360° output) or arcsin (one
sensor, 0° to ±90° output) function in order to get the angle of the magnet.
4.2.1.2 Preferred Magnet Types
• Diametrically magnetized disc or cylinder
4.2.1.3 General Accuracy and Resolution
• Good accuracy, ≈ 8° max error peak-to-peak (found experimentally using the DRV5055)
• High resolution is possible (depending on ADC)
• Results come in degrees
• Accuracy is affected by physical setup and magnet selection
4.2.1.4 Considerations
• The magnet must be oriented to align the degree output to the desired physical location.
• The sensors and magnet must be placed so that the sensor voltage output is not clipped or railed at either
the north or south pole.
• If writing code that uses both the arctan2 and arcsin functions, consider using the identity in Equation 2. This
identity saves on program space because the arctan2 function already uses the arctan function.

x
arcsin x = 2 × arctan (2)
1 + 1 − x2

• Most arctan2 and arcsin functions output the angle in radians. The angle can be converted to degrees using
Equation 3.

A° = Ar × 180
π (3)

• While a ±90° range is possible with one sensor, the voltage measurement accuracy and sensor noise limit the
angle to a value less than ±90°.

12 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.2.2 One Bipolar Sensor, Peak Calibrated

4.2.2.1 Specific Implementation
With one sensor, as in Figure 4-2, the sensor output voltage takes the form shown in Figure 4-10. With a peak
calibration phase, both the min and max voltage values are known.
3.3
VOUT, amplitude V = 1.2
2.8875

2.475

Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D005

Figure 4-10. One Sensor Peak Calibrated Data

4.2.2.2 Calculating Angle

To determine the angle that the magnet points towards, input the normalized data into the arcsin function, as
outlined in Section 4.2.1.1. The identity in Equation 2 may be used instead of arcsin if desired.
4.2.2.3 Accuracy
The output from the arcsin function has a swing of ±90°. However, the accuracy generally decreases when the
angle is too close to 90° or –90° because there is not much variance in the output voltage for those regions. In
general, the accuracy is usually within 8° peak-to-peak when operating at a swing of ≈±80°, based on datasheet
parameters and experimental data taken with the DRV5055. Figure 4-11 shows an example of a possible error
curve for this setup.
3
2
1
0
-1
Error (q)

-2
-3
-4
-5
-6
-7
-90 -70 -50 -30 -10 10 30 50 70 90
Angle (q) D007

Figure 4-11. One Sensor Peak Calibrated Error

4.2.3 Two Bipolar Sensors 90° Apart, Peak Calibrated

4.2.3.1 Specific Implementation
With two sensors 90° apart, as in Figure 4-4, the sensor output voltage takes the form shown in Figure 4-12.
With a peak calibration phase, both the min and max voltage values are known for each curve.

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 13
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

3.3
VOUT 2, amplitude V = 1.2
2.8875 VOUT 1, amplitude V = 1

2.475

Sensor Output (V)

2.0625

1.65

1.2375

0.825

0.4125

0
0 30 60 90 120 150 180 210 240 270 300 330 360
Magnet Angle to Sensor (q) D006

Figure 4-12. Two Sensors Peak Calibrated Data

4.2.3.2 Calculating Angle

To determine the angle that the magnet points towards, input the normalized data into the arctan2 function, as
outlined in Section 4.2.1.1.
4.2.3.3 Accuracy
The output from the arctan2 function goes from 0° to 360°. In general, the accuracy is usually within 8°
peak-to-peak, based on datasheet parameters and experimental data taken with the DRV5055. Figure 4-13
shows an example of a possible error curve for this setup.
5

2
Error (q)

-1

-2

-3
0 30 60 90 120 150 180 210 240 270 300 330 360
Ange (q) D008

Figure 4-13. Two Sensors Peak Calibrated Error

14 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.3 Lookup Table Calibration Implementations

4.3.1 Overview
4.3.1.1 General Implementation
With a lookup table calibrated system, sensor voltage data for known angles are recorded, and then the angle
for any measured voltage is taken from a linear interpolation between the known voltages. The process for
calibration is:
1. For each desired calibration angle, rotate the magnet to the angle, and record the measured voltage for each
sensor.
2. Then, during normal operation, measured voltages for each sensor fall between two of the previously
recorded voltages, referenced as Vabove and Vbelow. When using two sensors, make sure that the Vabove and
Vbelow for each sensor are associated with the same calibration angle.
3. The measured angle is then taken as a ratio of those two voltages and the respective known angles using
Equation 4:
V − Vbelow
ANGLEnew = measured
Vabove − Vbelow × ANGLEabove − ANGLEbelow + ANGLEbelow (4)

Note
It is important to note that calibration regions where Vabove – Vbelow ≈ 0 do not work in Equation 4, and
therefore must not be used for this method.

4.3.1.2 Preferred Magnet Types

• Diametrically magnetized disc or cylinder
4.3.1.3 General Accuracy and Resolution
• High accuracy is achievable, but depends significantly on the number of calibration points used. A good
starting point to estimate the spacing (in degrees) needed between calibration points for a desired peak-
to-peak accuracy (in degrees) is found using Equation 5, based on experimental data collected with the
DRV5055. This equation uses peak-to-peak error, so a peak-to-peak accuracy of 1° gives an approximate
error of ±0.5°. With a lookup table calibration, the spacing between calibration points must not be greater than
30°, or the error may be unpredictable.
Spacing ≈ Accuracy * 8 (5)
• High resolution is possible (depending on ADC)
• Results come in degrees
• Accuracy is also affected by physical setup (when using one sensor) and magnet selection

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 15
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.3.1.4 Considerations
• The magnet does not need to be oriented when using two sensors, and only needs to be roughly oriented
when using one sensor because a 0° point can be set during calibration.
• The sensors and magnet must be placed so that the sensor voltage output is not clipped or railed at either
the north or south pole.
• 0° and 360° are the same angle; therefore, use 0 as ANGLEbelow and 360 as ANGLEabove in Equation 4.
• Although it is possible to measure voltages that are out of the range of the lookup table (either above the max
or below the min recorded voltage values), the absolute min and max values are unknown. Therefore, these
measurements are unusable for linear interpolation.
• The lookup table calibration method can be more difficult to implement when using two sensors than methods
that use the arctan2 function for the following reasons:
– Exceptions must be coded to account for when Vabove – Vbelow ≈ 0 V in order to avoid dividing by 0.
– Data from the nonlinear regions of each sensor output must be avoided.
– Calibration data for each sensor must be stored (instead of storing the arctan2 output); therefore:
• It is harder to determine which calibration region to use because the voltage from each sensor appears
in two different regions of the respective lookup tables.
• It is possible that near a calibration boundary line, the data from each sensor is on either side of that
boundary.

16 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.3.2 One Bipolar Sensor, Lookup Table Calibrated

4.3.2.1 Specific Implementation
With one sensor, as in Figure 4-2, the sensor output voltage takes the form shown in Figure 4-10. With a lookup
table calibration phase, specific voltages for various angles are known, but the min and max voltage values are
not known.
4.3.2.2 Calculating Angle
To determine the angle that the magnet points towards, find the estimated angle between two lookup table
points, as outlined in Section 4.3.1.1.
4.3.2.3 Accuracy
With just one sensor, only the linear region of the curve in Figure 4-10 is usable. This span is ≈140°; therefore,
the magnet must be roughly positioned so that the desired measurement range falls within the 140° linear
region. The accuracy for this method largely depends on the spacing between the calibration points, and is
estimated using Equation 5.
Figure 4-14, Figure 4-15, and Figure 4-16 each show an example of a possible error curve for this setup with a
different number of calibration points, calibrated between ±70°.

3 3

2 2

1 1
Error (q)

Error (q)

0 0

-1 -1

-2 -2

-3 -3
-90 -70 -50 -30 -10 10 30 50 70 90 -90 -70 -50 -30 -10 10 30 50 70 90
Angle (q) D009
Angle (q) D010

Figure 4-14. One Sensor Lookup Table Calibrated Figure 4-15. One Sensor Lookup Table Calibrated
Error, 7 Cal Points, ≈ 4° Peak-to-Peak Error Error, 10 Cal Points, ≈ 2° Peak-to-Peak Error
3

1
Error (q)

-1

-2

-3
-90 -70 -50 -30 -10 10 30 50 70 90
Angle (q) D011

Figure 4-16. One Sensor Lookup Table Calibrated Error, 16 Cal Points, ≈ 1° Peak-to-Peak Error

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 17
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.3.3 Two Bipolar Sensors ≈ 90° Apart, Lookup Table Calibrated

4.3.3.1 Specific Implementation
With two sensors, as in Figure 4-4, the sensor output voltage takes the form shown in Figure 4-12. These
sensors do not need to be 90° apart to have unique data for the lookup table. However, best practice is to have
the sensors close to 90° apart so that the unusable peaks of one signal are at the linear regions of the other
signal, allowing the peak data to be ignored. To avoid losing the benefit of two sensors, the spacing cannot
be approximately equal to either 0° or 180°. With a lookup table calibration phase, specific voltages for various
angles are known, but the min and max voltage values are not known.
4.3.3.2 Calculating Angle
To determine the angle that the magnet points towards, find the estimated angle between two lookup table
points, as outlined in Section 4.3.1.1. This process must be done for each sensor, and then the two results are
averaged to find the angle associated with the magnet. Areas where Vabove – Vbelow ≈ 0 (the peaks) are not
usable in this method; therefore, the best practice is to use the value from the other sensor in these regions.
4.3.3.3 Accuracy
The accuracy for this method largely depends on the spacing between the calibration points, and can be
estimated using Equation 5.
Figure 4-17, Figure 4-18, and Figure 4-19 each show an example of a possible error curve for this setup with a
different number of calibration points.

3 3

2 2

1 1
Error (q)

Error (q)

0 0

-1 -1

-2 -2

-3 -3
0 30 60 90 120 150 180 210 240 270 300 330 360 0 30 60 90 120 150 180 210 240 270 300 330 360
Angle (q) D012
Angle (q) D013

Figure 4-17. Two Sensors Lookup Table Calibrated Figure 4-18. Two Sensors Lookup Table Calibrated
Error, 14 Cal Points, ≈ 4° Peak-to-Peak Error Error, 26 Cal Points, ≈ 2° Peak-to-Peak Error
3

1
Error (q)

-1

-2

-3
0 30 60 90 120 150 180 210 240 270 300 330 360
Angle (q) D014

Figure 4-19. Two Sensors Lookup Table Calibrated Error, 45 Cal Points, ≈ 1° Peak-to-Peak Error

18 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.4 Peak Calibrated Plus Lookup Table Hybrid

4.4.1 Overview
4.4.1.1 General Implementation
With a peak calibrated plus lookup table hybrid system, bipolar sensor data is first normalized to ±1 for use with
the arctan2 (two sensors) or arcsin (one sensor) function to determine a preliminary angle (Ap). Arctan2 must be
used instead of arctan because arctan2 accounts for which of the two values are negative. Then the calculated
Ap for various known ideal angles (Ai) are recorded as calibration angles (Ac), and a linear error adjustment is
done to all future Ap based on the error between the recorded Ac and the known Ai. The process for calibration
is:
1. Find Ap by inputting the normalized data into the arctan2 or arcsin function, as outlined in Section 4.2.1.1).
2. Rotate the magnet to the 0° point and use this Ap as a zero-offset for all other Ap calculations.
3. Rotate the magnet to the desired calibration Ai and record the Ap as Ac. (The Ac for 0° is 0° because of the
zero-offset adjustment done previously).
4. Then, during normal operation, each new Ap falls between two of the previously recorded Ac; the angle just
above Ap (AcA) and the angle just below Ap (AcB).
5. The error for any point between AcA and AcB is estimated from a linear approximation of the known error
from AcA to AiA and Acb to AiB in the form of y = Mx + b, where:
A −A
M = A iA − AiB , and B = AiA − AcA × M (6)
cA cB

6. Then, all new angle values (An) can be calculated using Equation 7:
An = M × Ap + B (7)

4.4.1.2 Preferred Magnet Types

• Diametrically magnetized disc or cylinder
4.4.1.3 General Accuracy and Resolution
• High accuracy is achievable, but significantly depends on the number of calibration points used. A good
starting point to estimate the spacing (in degrees) needed between calibration points for a desired peak-
to-peak accuracy (in degrees) is found using Equation 8, based on experimental data collected with the
DRV5055. This equation uses peak-to-peak error, so a peak-to-peak accuracy of 1° gives an approximate
error of ±0.5°.
Spacing ≈ Accuracy × 15 (8)
• High resolution is possible (depending on ADC).
• Results come in degrees.
• Accuracy is also affected by physical setup and magnet selection.

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 19
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.4.1.4 Considerations
• When using two sensors, the magnet does not need to be oriented. When using one sensor, the magnet only
needs to be roughly oriented because a 0° point can be set during calibration.
• The sensors and magnet must be placed so that the sensor voltage output is not clipped or railed at either
the north or south pole.
• 0° and 360° are the same angle; therefore, use 0 as AiB and use 360 as AiA in Equation 6.
• If writing code that uses both the arctan2 and arcsin functions, consider using the identity in Equation 9,
which saves on program space because the arctan2 function already uses arctan.

x
arcsin x = 2 × arctan (9)
1 + 1 − x2

• Most arctan2 and arcsin functions output the angle in radians. This angle can be converted to degrees using
Equation 10:

– A° = Ar × 180
π (10)

• While a ±90° range is possible with one sensor, the voltage measurement accuracy and sensor noise limit the
angle to a value less than ±90°.

20 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com Methods

4.4.2 One Bipolar Sensor, Hybrid Calibrated

4.4.2.1 Specific implementation
With one sensor, as shown in Figure 4-2, the sensor output voltage takes the form shown in Figure 4-10. With
a lookup table plus peak calibration phase, specific voltages for various angles, as well as the min and max
voltage values, are known.
4.4.2.2 Calculating Angle
To determine the angle that the magnet points towards, input the normalized data into the arcsin functionand
calibrating for accuracy using a lookup table, as outlined in Section 4.4.1.1. The identity in Equation 9 may be
used instead of arcsin, if desired.
4.4.2.3 Accuracy
The output from the arcsin function has a swing of ±90°, but the accuracy generally decreases when too
close 90° or -90°, as there is not much variance in the output voltage for those regions. In general, this
leaves an operating region of ≈±80°. Therefore, the magnet needs to be roughly positioned so that the desired
measurement range falls within the ±80° region. The accuracy for this method largely depends on the spacing
between the calibration points, which can be estimated using Equation 8.
Figure 4-20, Figure 4-21, and Figure 4-22 each show an example of a possible error curve for this setup with a
different number of calibration points, calibrated between ±80°.

3 3

2 2

1 1
Error (q)

Error (q)

0 0

-1 -1

-2 -2

-3 -3
-90 -70 -50 -30 -10 10 30 50 70 90 -90 -70 -50 -30 -10 10 30 50 70 90
Angle (q) D015
Angle (q) D017

Figure 4-20. One Sensor Peak Calibrated Plus Figure 4-21. One Sensor Peak Calibrated Plus
Lookup Table Error, 3 Cal Point, ≈ 4° Peak-to-Peak Lookup Table Error, 5 Cal Points, ≈ 2° Peak-to-Peak
Error Error
3

1
Error (q)

-1

-2

-3
-90 -70 -50 -30 -10 10 30 50 70 90
Angle (q) D018

Figure 4-22. One Sensor Peak Calibrated Plus Lookup Table Error, 9 Cal Points, ≈ 1° Peak-to-Peak Error

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 21
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
Methods www.ti.com

4.4.3 Two Bipolar Sensors 90° Apart, Hybrid Calibrated (Recommended High Accuracy Method)

Note
This is the method that is used on the DRV5055-ANGLE-EVM.

4.4.3.1 Specific Implementation

With two sensors at 90° apart as in Figure 4-4, the sensor output voltage will take the form shown in Figure 4-12.
Note that with a lookup table plus peak calibration phase, specific voltages for various angles are known, as well
as the min and max voltage values.
4.4.3.2 Calculating Angle
To determine the angle that the magnet points towards, input the normalized data into the arctan2 function and
calibrating for accuracy using a lookup table, as outlined in Section 4.4.1.1.
4.4.3.3 Accuracy
The accuracy for this method largely depends on the spacing between the calibration points, and can be
estimated using Equation 8.
Figure 4-23, Figure 4-24, and Figure 4-25 each show an example of a possible error curve for this setup with a
different number of calibration points.

3 3
2.5
2 2
1.5
1 1
0.5
Error (q)

Error (q)

0 0
-0.5
-1 -1
-1.5
-2 -2
-2.5
-3 -3
0 30 60 90 120 150 180 210 240 270 300 330 360 0 30 60 90 120 150 180 210 240 270 300 330 360
Angle (q) D019
Angle (q) D020

Figure 4-23. Two Sensors Peak Calibrated Plus Figure 4-24. Two Sensors Peak Calibrated Plus
Lookup Table Error, 8 Cal Points, ≈ 4° Peak-to-Peak Lookup Table Error, 14 Cal Points, ≈ 2° Peak-to-
Error Peak Error
3

1
Error (q)

-1

-2

-3
0 30 60 90 120 150 180 210 240 270 300 330 360
Angle (q) D021

Figure 4-25. Two Sensors Peak Calibrated Plus Lookup Table Error, 25 Cal Points, ≈ 1° Peak-to-Peak
Error

22 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and SLYA036B – JULY 2018 – REVISED NOVEMBER 2021
Calibration Submit Document Feedback
Copyright © 2021 Texas Instruments Incorporated
www.ti.com References

5 References
• Texas Instruments, Overview Using Linear Hall Effect Sensors to Measure Angle application brief
• Texas Instruments, Breakout Adapter for SOT-23 and TO-92 Hall Sensor Evaluation Module tools page
• Texas Instruments, DRV5055 Evaluation Module tools page
• Texas Instruments, DRV5055-ANGLE-EVM tools page
• Texas Instruments, Angle Measurement With Multi-Axis Linear Hall-Effect Sensors application report
• Texas Instruments, E2E forums at https://e2e.ti.com/

6 Revision History
NOTE: Page numbers for previous revisions may differ from page numbers in the current version.

Changes from Revision A (August 2018) to Revision B (November 2021) Page

• Updated the numbering format for tables, figures, and cross-references throughout the document..................1
• Updated the Abstract.......................................................................................................................................... 1
• Updated the References...................................................................................................................................23

Changes from Revision * (July 2018) to Revision A (August 2018) Page

• Changed document title......................................................................................................................................1

SLYA036B – JULY 2018 – REVISED NOVEMBER 2021 Linear Hall-Effect Sensor Angle Measurement Theory, Implementation, and 23
Submit Document Feedback Calibration
Copyright © 2021 Texas Instruments Incorporated
 IMPORTANT NOTICE AND DISCLAIMER
TI PROVIDES TECHNICAL AND RELIABILITY DATA (INCLUDING DATA SHEETS), DESIGN RESOURCES (INCLUDING REFERENCE
DESIGNS), APPLICATION OR OTHER DESIGN ADVICE, WEB TOOLS, SAFETY INFORMATION, AND OTHER RESOURCES “AS IS”
AND WITH ALL FAULTS, AND DISCLAIMS ALL WARRANTIES, EXPRESS AND IMPLIED, INCLUDING WITHOUT LIMITATION ANY
IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT OF THIRD
PARTY INTELLECTUAL PROPERTY RIGHTS.
These resources are intended for skilled developers designing with TI products. You are solely responsible for (1) selecting the appropriate
TI products for your application, (2) designing, validating and testing your application, and (3) ensuring your application meets applicable
standards, and any other safety, security, regulatory or other requirements.
These resources are subject to change without notice. TI grants you permission to use these resources only for development of an
application that uses the TI products described in the resource. Other reproduction and display of these resources is prohibited. No license
is granted to any other TI intellectual property right or to any third party intellectual property right. TI disclaims responsibility for, and you
will fully indemnify TI and its representatives against, any claims, damages, costs, losses, and liabilities arising out of your use of these
resources.
TI’s products are provided subject to TI’s Terms of Sale or other applicable terms available either on ti.com or provided in conjunction with
such TI products. TI’s provision of these resources does not expand or otherwise alter TI’s applicable warranties or warranty disclaimers for
TI products.
TI objects to and rejects any additional or different terms you may have proposed. IMPORTANT NOTICE

Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265
Copyright © 2022, Texas Instruments Incorporated