Example with a study case using COMSOL AC/DC (source: comsol.com paper on modeling PCB-based inductive position sensors with the COMSOL AC/DC module).

Target surface current distribution and coil-plane magnetic field distribution

Figure 1. Target surface current distribution and coil-plane magnetic field distribution.

Context

In a perfect world without mechanical or physical interferences, inductive position sensors (IPS) would operate with minimal error. However, real-world applications always involve non-ideal conditions. Various factors such as additional metallic parts, feeding wires, multilayer PCB designs and nearby ground planes can introduce accuracy errors. Evaluating and optimizing these factors is essential for accurate IPS performance.

This study focuses on three key factors affecting IPS accuracy using COMSOL simulations:

  1. Number of PCB layers (2 vs. 4)
  2. Feeding wires connecting Rx coils to the IC
  3. Nearby metallic ground planes

Methodology

Starting from an ideal configuration, non-idealities are progressively introduced. The following protocol determines the accuracy for each configuration:

  • Simulations: perform simulations with different air gaps between coils and target, recording the Rx signal amplitude.
  • Angle calculation: calculate the sensor's angle using Clarke's transformation on the Rx signals followed by an arctangent calculation.
  • Accuracy assessment: compare the calculated angle with the target's reference position.
Simulated Rx coil signal amplitude vs target position over one electrical period

Figure 2. Simulated Rx coil signal amplitude vs. target position over one full electrical period.

Results

1. Ideal case: 4-layer PCB. An ideal 4-layer PCB with 0.3 mm spacing between planes shows nearly perfect accuracy across various air gaps.

PCB design of an IPS in an ideal 4-layer case and the associated accuracy

Figure 3. PCB design of an IPS in an ideal case, 4 layers, and the associated accuracy.

2. Introducing a ground plane. Adding a metallic ground plane introduces a first-harmonic error, increasing with the ground plane's proximity to the coils.

PCB design including a ground plane, 4 layers, and the associated accuracy

Figure 4. PCB design including a ground plane, 4 layers, and the associated accuracy.

3. Feeding-wire configuration. Feeding wires introduce parasitic surfaces. Two configurations are tested, 60 deg and 120 deg electrical separation.

PCB design with 60 degree electrical feeding configuration, 4 layers, and accuracy

Figure 5. 60 deg electrical feeding configuration, 4 layers, and its accuracy.

PCB design with 120 degree electrical feeding configuration, 4 layers, and accuracy

Figure 6. 120 deg electrical feeding configuration, 4 layers, and its accuracy. The 120 deg separation demonstrates superior accuracy.

4. 2-layer PCB configurations. For 2-layer PCBs with 1.6 mm spacing:

PCB design with 60 degree electrical feeding configuration, 2 layers, and accuracy

Figure 7. 60 deg electrical feeding configuration, 2 layers, and its accuracy.

PCB design with 120 degree electrical feeding configuration, 2 layers, and accuracy

Figure 8. 120 deg electrical feeding configuration, 2 layers, and its accuracy. Both 2-layer configurations show significant errors due to larger horizontal parasitic surfaces from feeding lines, which are minimized in the 4-layer design.

Conclusion

Optimizing PCB design, especially the number of layers and the arrangement of feeding wires, is vital for minimizing accuracy errors in inductive position sensors. This study underscores the importance of careful PCB layout to ensure high-performance sensor operation in real-world applications.