See the effect of differential technology on magnetic position sensors and stray field interference

Magnetic position sensing has proved to be popular in a range of motion- and motor-control applications in the industrial and automotive markets. Various methods of measuring flux density have evolved, leading to the development of the fully integrated magnetic position sensor IC, which incorporates the magnetic sensing element, , and . The latest generation of 3D magnetic position can sense magnetic flux in three dimensions, opening up a host of applications.

Whichever method of magnetic sensing is used, magnetic technology is more robust and reliable than optical sensing or contacting (potentiometer) methods for position sensing. That’s because it’s not affected by the dust, dirt, grease, vibration, and humidity common to harsh automotive and industrial applications.

Design engineers who use conventional magnetic position sensors are increasingly running into a problem, however. That’s the interference from stray magnetic fields, which tends to corrupt the sensor’s output or reduce the signal-to-noise ratio (SNR) to unacceptable levels. Even the known risk of malfunction due to stray magnetism is damaging to safety-critical designs, which in the automotive field must comply with the stringent risk-management requirements of the ISO26262 functional safety standard.

The increased risk has emerged as electrification in vehicles has been extended. Motors and cables carrying high currents are particularly powerful sources of stray magnetism. These can equally be found in many industrial applications. Counter-measures to protect a vulnerable magnetic position sensor from stray magnetism are cumbersome and expensive. A better method is to make the sensor highly immune to stray magnetic fields.

Protecting the sensor from stray fields

A common approach to dealing with magnetic stray fields is to shield the sensor IC. This is a blunt tool to use, for two reasons. First, the shielding material interacts not only with the magnetic stray field, but also with the field of the magnet with which the sensor is paired. This paired magnet is attached to the moving object to be measured. The static position sensor converts the changes in magnetic flux as the paired magnet moves towards or away from it into precise displacement measurements.

The shielding material may itself become magnetized, and its characteristics will also tend to change as the temperature changes. In addition, shielding materials exhibit hysteretic behavior, potentially redirecting the paired magnet’s flux lines away from the sensor. To prevent the shield’s parasitic properties from disrupting the system’s operation, it must be placed at some distance from the magnet.

This limits the system designer’s freedom to place, route, and enclose the sensor module’s components. It also makes the system larger, heavier, more complicated, more difficult to assemble, and more expensive.

A completely different approach, which requires no shielding, is to pair the position sensor with a magnet that has high remanence (Br), and to assemble it in close proximity to the sensor. The effect is to make the signal-to-stray-field ratio more favorable; it has the same effect on the overall SNR.

Unfortunately, strong , such as the NdFeB or SmCo types, are around 10X more expensive than cheap hard ferrite or plastic-bounded magnets, ruining the economic case for position sensors in many cases. In addition, this option isn’t available to the many applications that can’t accommodate the magnet close to the IC.

Built-in immunity with dual-pixel sensor ICs

A better approach is to make the sensor immune to stray magnetism. And in fact, a basic mathematical operation enables the noise from stray magnetic fields to be cancelled, provided the sensor’s supports the technique. In addition, intelligent placement of the paired magnet, as close to the IC as possible, will always help increase a sensor module’s tolerance of stray magnetism. But the only way to achieve immunity to stray fields is to use a position sensor that has this feature built-in.

The crucial hardware feature of a magnetic position sensor with stray field immunity is a dual-pixel magnetic sensing element (Figure 1). Unlike a conventional absolute , a dual-pixel type uses two pixel cells instead of one to determine the magnet’s position. This dual-pixel structure then enables the implementation of differential measurement. Each pixel cell can measure all three vectors of the magnetic field: Bx, By, and Bz. In members of ams’ AS54xx sensor family, for example, these two pixel cells are spaced 2.5 mm apart.

Structure of a dual-pixel sensor IC

1. Structure of a dual-pixel sensor IC.

To illustrate the mathematical operation simply, the following description of the sensor’s working principle focuses on a linear application. Here, only the vectors Bx and Bz are measured by the device. The sensor IC measures the following values to determine the magnet’s position:

Bx_Pix0: x vector of the magnetic field, measured by Pixel 0
Bx_Pix1: x vector of the magnetic field, measured by Pixel 1
Bz_Pix0: z vector of the magnetic field, measured by Pixel 0
Bz_Pix1: z vector of the magnetic field, measured by Pixel 1

Figure 2 shows the output curves of this application over a magnet travel of -15 to +15 mm. When the magnet is at position “0,” the magnet is exactly centered over the IC. At this position, the magnet’s north-to-south pole transition is exactly between the two pixels. Since the pixels are 2.5 mm apart, there’s is a ±1.25-mm phase shift between the Pix0 and Pix1 curves.

Measurement outputs of a two-pixel sensor IC

2. Measurement outputs of a two-pixel sensor IC.

From these four values, the sensor IC calculates two differential signals, called Bi (for the x vector) and Bj (for the z vector):

Bi = Bx_Pix0 – Bx_Pix1
Bj = Bz_Pix0 – Bz_Pix1

Imagine a stray field, Bs, applied to the device being measured. The stray field’s source is usually further from the sensor IC than its paired magnet. This means that the designer can assume that the same stray field vector is applied to both pixel cells. Here, then, are the same Bi and Bj formulas, but with stray field Bs applied to them:

It’s easy to see that the value of Bs has no influence on the values of Bi and Bj. Bs can simply be eliminated from the calculation, to produce accurate position measurements without any interference from stray fields, as shown in Figures 3 and 4. This is the differential principle of position measurement. The magnet’s position (MPos) can then be calculated from the values of Bi and Bj by an ATAN2 function.

MPos = ATAN2( - Bj ; Bi )

The sin and cos signal, as calculated by the sensor IC

3. The sin and cos signal, as calculated by the sensor IC.

The magnet’s position, as calculated by the sensor IC

4. The magnet’s position, as calculated by the sensor IC.

A demonstration of stray-field immunity

The superior performance of a dual-pixel magnetic position sensor with differential sensing has been demonstrated in the lab. The test described below compares the measurement results from an automotive position sensor module containing a dual-pixel sensor with another automotive module which contains a conventional single-pixel sensor.

The modules measured a magnet’s movement in an arc above the sensor IC, and a Helmholtz coil applied a stray field to the modules (Figure 5). The modules’ output voltage was measured with an oscilloscope. The IC’s output voltage changes in relation to changes in the magnet’s position (Figure 6). The coil was configured to generate a stray field of known strength in the vectors Bx, By, or Bz. This kind of measurement would typically be required in an application such as measuring the movement of a car’s brake, accelerator, or clutch pedal.

This test measured the movement of a magnet in an arc

5. This test measured the movement of a magnet in an arc.

Shown is the output characteristic of the two modules

6. Shown is the output characteristic of the two modules.

The captured data in Figure 7 shows that the error of the single-pixel sensor IC is more than 30 times greater than the error of the dual-pixel IC when exposed to a stray field in the z direction. DC stray fields appear as an offset superposed on the desired signal. AC stray fields appear as noise; the frequency of the stray field is superposed on the desired signal.

Conditions of test:

Magnet position: 4 V
Stray field direction: z
Stray field frequency: 50 Hz
Stray field strength: 2500 A/m

Displaying output voltages in the presence of a stray field, Channel 1 shows the results from the dual-pixel sensor, and Channel 2 shows the single-pixel sensor

7. Displaying output voltages in the presence of a stray field, Channel 1 shows the results from the dual-pixel sensor, and Channel 2 shows the single-pixel sensor.

Figure 8 also shows a difference between the two sensor types. The ±1% error limit is a typical requirement in automotive motion-sensing applications. The test measured all noise sources, including the stray magnetic field. Integral non-linearity and temperature drift are application-dependent, and so their values aren’t included in this diagram.

Here is the error comparison between the two modules when subject to AC and DC stray fields. The noise value is shown in blue, and offset in yellow

8. Here is the error comparison between the two modules when subject to AC and DC stray fields. The noise value is shown in blue, and offset in yellow.

Dual-pixel products on the market

The dual-pixel method of differential sensing is implemented in all of the AS54xx series of automotive-qualified position sensors. They can operate in a temperature range of -40°C to +150°C, with no temperature compensation. In addition, they can operate in an input range from 5 to 100 mT. When combined with the high tolerance of magnetic stray fields, this allows the use of small and cheap magnets. Reliable operation in the presence of stray magnetism helps automotive system designers to comply with ISO26262.

In the automotive arena, stray field immunity is going to become an increasingly important attribute of magnetic position sensors as the drivetrain of vehicles becomes partially or wholly electrified. New standards such as ISO11452-8 add to the challenge.

In this electromagnetically and mechanically harsh environment, 3D provide a means for designers to achieve robust performance and to provide for compliance with the most exacting functional safety standards, without the need for complicated and expensive magnetic shielding.

David Schneider, Application Engineer, graduated with a Bachelor of Science degree from the FH Joanneum Kapfenberg University of Applied Sciences. On graduating, he joined ams AG as an application engineer, specializing in 3D magnetic position sensors. Marcel Urban graduated with a Master degree in microelectronics at FH Technikum Kärnten. He is currently working at ams AG as manager of marketing and product management.