Resolver and Encoder Interface Circuits

In any system where precise control of a motor or shaft is required—from industrial robotics and CNC machining to aerospace actuators and electric vehicles—knowing the exact angular position is non-negotiable. Resolver and encoder interface circuits are the critical electronic bridges that translate the raw, often analog, signals from these rotary sensors into the clean digital angle measurements that a microprocessor or controller can use. Understanding how these interfaces work is key to designing reliable motion control systems that are accurate, robust, and responsive to the demands of high-performance applications.

The Resolver and Its Digital Conversion

A resolver is an analog, rotary transformer that provides absolute angular position information. It is exceptionally robust, often used in harsh environments (high temperature, vibration) where optical encoders might fail. The resolver has a single primary winding on the rotor, which is excited by a high-frequency AC reference signal, typically a sine wave of 5-10 kHz. On the stator, there are two secondary windings mechanically offset by 90 degrees.

As the rotor turns, the coupling between the windings changes. This induces two modulated output signals on the stator: $V_{sin}=A\cdot \sin (\omega t)\cdot \sin (\theta )$ and $V_{cos}=A\cdot \sin (\omega t)\cdot \cos (\theta )$. Here, $\omega$ is the excitation frequency, and $\theta$ is the mechanical shaft angle. The amplitude envelopes of these signals, $\sin (\theta )$ and $\cos (\theta )$, contain the angular information.

This is where the resolver-to-digital converter (RDC) comes in. Its primary job is to extract the angle $\theta$ from these sine and cosine envelopes. The most common and effective method employs a type-II tracking loop. This is a closed-loop control system that works by generating its own estimated angle $\varphi$, comparing it to the actual resolver signals, and driving the error to zero. Here’s the step-by-step process:

1. The RDC takes its current angle estimate $\varphi$ and computes $\sin (\varphi )$ and $\cos (\varphi )$ internally.
2. It uses these to demodulate the incoming resolver signals. A key operation is performed: $$Error=V_{sin}\cdot \cos (\varphi )-V_{cos}\cdot \sin (\varphi )$$.
3. Using trigonometric identities, this simplifies to $Error=K\cdot \sin (\theta -\varphi )$. For small errors, $\sin (\theta -\varphi )\approx (\theta -\varphi )$.
4. This error signal is fed into a type-II (proportional-integral) loop filter. The integrator in this filter is crucial—it ensures zero steady-state error even for a constantly moving shaft (i.e., it can track a velocity input without lag).
5. The filtered error drives a voltage-controlled oscillator (VCO), which updates an up/down counter, thereby adjusting the digital angle output $\varphi$ to chase and lock onto the true mechanical angle $\theta$.

This closed-loop approach makes the RDC inherently tolerant to variations in signal amplitude and provides excellent noise rejection.

Incremental Encoder Interfaces: Counting Pulses

An incremental encoder is a digital optical or magnetic device that outputs pulsed signals as the shaft rotates. It provides relative position information. The standard output consists of two square wave channels, A and B, in quadrature (90 degrees out of phase). The sequence of these pulses indicates direction, and counting them yields relative displacement from a known starting point.

The interface circuit for an incremental encoder is a quadrature decoding counter. This digital logic block, often integrated into microcontrollers or FPGAs, performs three key functions:

• Direction Detection: It examines the phase relationship between Channel A and Channel B. If A leads B, the shaft is turning clockwise; if B leads A, it is turning counterclockwise.
• Counting: It increments or decrements a digital counter based on the detected direction. Each full cycle (four edges) of the quadrature signal can generate four counts, a technique known as 4x interpolation, which effectively quadruples the base resolution of the encoder disk.
• Index Handling: Most encoders include a third "Index" or "Z" channel that pulses once per revolution, providing a homing or absolute reference point.

The primary challenge with incremental interfaces is managing count integrity. Power loss or electrical noise can cause missed or erroneous counts, leading to a permanent positional error until the system is re-homed via the index pulse.

Absolute Encoder Interfaces: Reading a Unique Code

An absolute encoder solves the "loss-of-position-on-power-up" problem of incremental encoders. Its optical or magnetic disk is patterned to produce a unique digital code at each angular position over a single rotation (single-turn) or multiple rotations (multi-turn). Common coding schemes include Gray code (where only one bit changes per position to prevent read errors) and natural binary.

The interface for a parallel absolute encoder is relatively straightforward: it reads the state of multiple digital input lines (e.g., 12 or 16 bits) representing the code word. For serial absolute encoders (using protocols like SSI, BiSS, or Ethernet-based), the interface involves generating a clock signal and reading back a synchronized serial data stream. The received code is then mapped directly to an absolute angle, with no counting required. The design focus shifts to ensuring reliable, high-speed digital communication and correctly parsing the specific protocol.

Signal Conditioning and Noise Mitigation

Whether dealing with the low-level analog outputs of a resolver or the digital pulses of an encoder, signal conditioning is vital for accurate high-resolution position measurement. This stage addresses real-world imperfections before the conversion or counting logic.

For resolvers, conditioning involves:

• Amplification: Bringing the millivolt-level stator signals up to a robust voltage range for the RDC.
• Filtering: Applying band-pass filtering around the excitation frequency to remove out-of-band noise and DC offsets. The type-II tracking loop itself acts as a powerful low-pass filter for in-band noise.
• Ratio-metric Operation: The RDC’s tracking error is based on the ratio of Sine/Cosine, making it immune to variations in the excitation signal amplitude or the transformer coupling.

For encoders, conditioning focuses on the digital domain:

• Noise Filtering: Using Schmitt-trigger inputs or digital debouncing circuits to clean up noisy edges that could cause false counts.
• Signal Integrity: Employing differential line receivers (for RS-422 type outputs) to reject common-mode noise over long cable runs.
• Interpolation: As mentioned, 4x quadrature decoding is a fundamental form of interpolation. High-end counters may use analog interpolation of the sinusoidal encoder signals (if available) to achieve even higher resolution, such as subdividing a single sine wave period into thousands of counts.

Common Pitfalls

1. Ignoring Signal Integrity for Encoders: Running single-ended encoder signals over long cables in an electrically noisy industrial environment is a recipe for count errors. Correction: Always use differential transmission lines (twisted pair with a shield) and differential receivers. For resolvers, use shielded twisted pairs for each winding and avoid routing cables near motor power lines.

1. Mismatching Resolver Excitation and RDC Reference: An RDC requires a clean, synchronous reference signal that matches the frequency and phase of the resolver excitation. Using a poor-quality or mismatched reference introduces scaling and offset errors. Correction: Derive the RDC reference signal directly from the excitation oscillator or use a dedicated synthesizer within a monolithic RDC chip.

1. Overlooking Bandwidth and Latency: Both RDC tracking loops and encoder count rates have limits. Selecting an interface with too slow a tracking bandwidth for a high-speed motor will result in lag and dynamic error. Similarly, a quadrature counter with a low maximum clock speed will be overwhelmed by high pulse rates. Correction: Calculate the required bandwidth and pulse frequency based on maximum motor speed and system response requirements, then choose components with a significant safety margin.

1. Assuming Absolute Encoders are Error-Proof: While absolute encoders don't lose position on power cycling, they are still susceptible to communication errors, electromagnetic interference, or physical damage to the code disk. A corrupt serial read can still yield a wildly incorrect position. Correction: Implement protocol checks (parity, CRC), read-back verification, and sanity checks in software (e.g., is the position change physically possible between two samples?).

Summary

• Resolvers provide robust, analog absolute position via sine/cosine envelopes and are converted to digital angles using a closed-loop type-II tracking loop in a Resolver-to-Digital Converter (RDC).
• Incremental encoders provide relative position through quadrature A/B pulses, which are decoded and counted by a digital interface to track movement, often using 4x interpolation for higher resolution.
• Absolute encoders output a unique code at each angular position, allowing for instant, non-volatile position knowledge; their interfaces focus on reliable digital communication.
• Signal conditioning—including amplification, filtering, and noise rejection—is essential for both resolver and encoder systems to ensure accurate high-resolution position measurement in real-world electrically noisy environments.
• Successful implementation requires careful attention to signal integrity, component bandwidth, and system-level error checking to avoid common pitfalls that lead to measurement inaccuracy or failure.