Electronics Guide

DC Tachogenerators

DC tachogenerators are small permanent-magnet DC generators that produce a voltage directly proportional to rotational speed. The output polarity indicates direction of rotation, making them ideal for bidirectional speed control applications.

Key characteristics of DC tachogenerators include:

• Linearity - Output voltage varies linearly with speed, typically within 0.1% to 1% of full scale
• Voltage gradient - Expressed in volts per 1000 RPM. A gradient of a few V/kRPM keeps the signal inside a low-voltage input range at top speed, while tens of V/kRPM give a larger signal and better low-speed resolution at the cost of that headroom
• Ripple - Commutation causes AC ripple superimposed on the DC output, typically 1% to 5% of output voltage. Its frequency is the product of shaft speed and commutator segment count, so it falls with speed and the filter that removes it at low speed also limits the bandwidth of the velocity loop
• Temperature coefficient - Output varies with temperature due to magnet and winding resistance changes
• Brush wear - Mechanical brushes require periodic maintenance and limit maximum speed

DC tachogenerators are commonly used in analog servo systems, industrial drives, and applications requiring simple, reliable speed feedback without digital processing.

AC Tachogenerators

AC tachogenerators produce an alternating output and keep brushes out of the signal path. What varies with speed differs between the two types: a permanent-magnet machine varies both the frequency and the amplitude of its output, while an induction type holds the carrier frequency at the excitation frequency and varies only the amplitude. Either way the signal requires more processing than a DC tachogenerator's before a controller can use it.

Two main types exist:

• Permanent magnet AC tachometers - Generate sinusoidal output with both frequency and amplitude proportional to speed
• Drag cup tachometers - Carry an excitation winding, a thin conducting cup as the rotor, and an output winding at right angles to the excitation. The cup couples the two windings only when it turns, so the output amplitude is proportional to speed and its phase relative to the excitation gives the direction

Digital Tachometers

Digital tachometers generate pulse trains with frequency proportional to speed. They interface directly with microcontrollers and digital control systems, offering noise immunity and easy signal transmission over long distances.

Common implementations include:

• Slotted disc with optical sensor - Interrupts light beam to generate pulses
• Magnetic pickup with toothed wheel - A passing tooth changes the reluctance of a magnetic circuit and induces a voltage pulse in the pickup coil. The pulse amplitude grows with speed and vanishes at standstill, so a variable-reluctance pickup cannot report position or confirm a stopped shaft
• Hall effect sensors with magnetic targets - Detects passing magnets or ferrous teeth

Incremental Encoders

Incremental encoders generate pulses as the shaft rotates, with each pulse representing a fixed angular increment. The controller counts pulses to track relative position changes and calculates speed from pulse frequency.

Standard incremental encoder outputs include:

• Channel A - Square wave pulse train, one cycle per encoder line
• Channel B - Square wave in quadrature (90 degrees phase offset) with Channel A for direction detection
• Index (Z) - Single pulse per revolution for absolute position reference

Quadrature decoding provides four times the resolution of the encoder line count by detecting all rising and falling edges of both channels. Each line produces one full cycle on each channel, and one cycle of two signals in quadrature presents four distinct edges, so a 1000-line encoder yields 4000 counts per revolution under times-four decoding.

Key specifications include:

• Lines per revolution - The number of graduations on the disc. It equals the pulses per revolution (PPR) counted on one channel and one quarter of the counts per revolution obtained after times-four decoding, so the three figures must not be interchanged. Optical units run from about 100 lines to more than 10,000
• Maximum speed - Set jointly by the bearing rating and by the maximum output frequency of the line driver, and the frequency limit gives a hard ceiling: an encoder whose driver is rated to 300 kHz cannot exceed 300,000 cycles per second divided by its line count, so 1000 lines caps the shaft at 300 revolutions per second, or 18,000 RPM. A high line count therefore trades top speed for resolution
• Output type - TTL, open collector, differential line driver (RS-422), or push-pull
• Accuracy - Angular error between actual and indicated position, typically specified in arc-minutes or arc-seconds

Absolute Encoders

Absolute encoders provide a unique digital code for each shaft position, eliminating the need for homing routines after power-up. The position is known immediately without any shaft movement.

Encoding methods include:

• Gray code - Only one bit changes between adjacent positions, preventing ambiguous readings during transitions
• Binary code - Standard binary representation, requires careful timing to avoid glitches
• BCD (Binary Coded Decimal) - Simplifies display but uses more bits than pure binary

Absolute encoders are available as:

• Single-turn - Provides a unique code within one revolution, typically 10 to 16 bits, or 1,024 to 65,536 distinct positions
• Multi-turn - Tracks multiple revolutions using gear-driven secondary encoders or battery-backed counting, offering 12 to 16 bits of turn counting in addition to single-turn resolution

Communication interfaces for absolute encoders include parallel outputs, SSI (Synchronous Serial Interface), BiSS, EnDat, and various industrial fieldbus protocols.

Optical Encoder Technology

Optical encoders use light-based sensing to achieve high resolution and accuracy. Key components include:

• Code disc - Glass or plastic disc with precisely etched opaque and transparent patterns
• Light source - LED with collimating optics provides parallel light beam
• Photodetectors - Phototransistors or photodiodes detect transmitted or reflected light
• Signal conditioning - Comparators convert analog signals to digital outputs

Transmissive encoders pass light through the disc, while reflective encoders detect light reflected from the disc surface. Transmissive designs generally offer better performance but require more precise alignment.

Magnetic Encoders

Magnetic encoders use magnetized targets and magnetic field sensors to detect position. They offer superior resistance to contamination, temperature extremes, and shock compared to optical encoders.

Technologies include:

• Magnetoresistive sensors - Detect field direction from multipole magnetic ring
• Hall effect sensors - Respond to field strength from magnetic targets
• Inductive sensors - Detect eddy current variations in conductive targets

While magnetic encoders traditionally offered lower resolution than optical types, modern designs reach 16 bits per revolution and beyond. Resolution is not the same as accuracy here: the accuracy of a magnetic encoder over a full turn depends on how uniformly its target ring is magnetized and on how well the sensor is centered on it, so a high bit count does not by itself close the gap to a comparable optical encoder.

Hall Sensor Fundamentals

The Hall effect produces a voltage across a current-carrying conductor when exposed to a perpendicular magnetic field. In motor applications, Hall sensors detect the position of permanent magnets on the rotor.

Hall sensors are available as:

• Linear Hall sensors - Output voltage proportional to field strength
• Hall switches - Digital output changes state at defined field thresholds
• Hall latches - Toggle between states at opposite polarity thresholds

Three-Phase Hall Configurations

BLDC motors typically use three Hall sensors spaced 120 electrical degrees apart. Electrical angle is the mechanical angle multiplied by the number of pole pairs, so for a motor with p pole pairs the sensors sit 120/p mechanical degrees apart. Three binary outputs allow eight combinations, but the all-high and all-low patterns never occur in a correctly built and correctly wired motor. The remaining six states divide the 360 degrees of each electrical cycle evenly, so each state marks a sector of 60 electrical degrees and the rotor position is known to within that sector.

Hall sensor placement considerations include:

• Electrical versus mechanical degrees - A 4-pole motor has two pole pairs and therefore two electrical cycles per mechanical revolution, which by the relation above puts its three Hall sensors 60 mechanical degrees apart
• Sensor alignment - Sensors must be precisely positioned relative to stator windings
• Temperature compensation - Hall sensitivity varies with temperature, affecting switching thresholds

The six-step (trapezoidal) commutation pattern derived from Hall sensors is simple to implement but produces torque ripple. Higher-performance systems use encoder feedback for sinusoidal commutation.

Integrated Hall Sensor Solutions

Modern motor designs often integrate Hall sensors directly into the motor housing or stator assembly. This simplifies installation but limits field replacement options. Some integrated circuits combine multiple Hall sensors with signal conditioning and interface logic in a single package.

Back-EMF Fundamentals

When a motor rotates, its windings cut through magnetic flux lines, inducing a voltage proportional to speed. In BLDC motors, the back-EMF waveform contains information about rotor position that can be used for commutation timing.

The back-EMF voltage is described by:

• Magnitude - Proportional to shaft speed through the motor voltage constant K_e, so the induced voltage is K_e × ω_m, where ω_m is the mechanical angular velocity. Data sheets quote K_e in volts per 1000 RPM or in volts per radian per second, and per phase or line to line, so the definition must be checked before the figure is used
• Frequency - Equals the electrical frequency, which is the mechanical speed in revolutions per second multiplied by the number of pole pairs
• Phase - Related to rotor position, enabling commutation timing

Zero-Crossing Detection

The most common back-EMF sensing method detects when the back-EMF voltage crosses zero in the undriven motor phase. During six-step commutation, one phase is always inactive and can be monitored.

Implementation challenges include:

• PWM noise - Switching transients obscure the back-EMF signal, requiring filtering or synchronous sampling
• Low-speed operation - Back-EMF is too small at low speeds for reliable detection
• Starting - No back-EMF exists at standstill, requiring open-loop starting sequences
• Voltage dividers - Motor voltages must be scaled for controller input ranges

Advanced Sensorless Techniques

Beyond zero-crossing detection, sophisticated algorithms extract position information across the entire speed range:

• Observer-based methods - State estimators predict rotor position from measured currents and voltages
• High-frequency injection - Injects test signals to detect rotor position through saliency effects, enabling operation at zero speed
• Direct back-EMF integration - Integrates back-EMF to calculate flux linkage and position
• Model reference adaptive systems (MRAS) - Compare motor model with measurements to estimate speed and position

These advanced techniques enable sensorless operation of PMSM and induction motors with performance approaching that of encoder-based systems.

Resolver Operating Principle

A resolver is essentially a rotating transformer with one primary winding on the rotor and two secondary windings on the stator, displaced 90 electrical degrees from one another. In a single-speed resolver, which completes one electrical cycle per mechanical revolution, that displacement is also 90 mechanical degrees; in a multi-speed resolver it is 90 divided by the number of speeds. The rotor is excited with a sinusoidal reference signal, and the stator windings return amplitude-modulated signals whose envelopes vary as the sine and the cosine of the rotor angle.

With the rotor driven by an excitation V_ref × sin(ωt), the output signals are:

• Sine output - V_s = T × V_ref × sin(θ) × sin(ωt)
• Cosine output - V_c = T × V_ref × cos(θ) × sin(ωt)

Here θ is the rotor angle, ω is the angular frequency of the excitation in radians per second and ωt is its instantaneous phase, and T is the resolver's transformation ratio, the ratio of secondary to primary voltage at the angle of maximum coupling. T is below unity in a standard resolver and is stated in its data sheet, so the stator signals are smaller than the excitation and the converter must be scaled for them.

The rotor angle follows from the ratio of the two demodulated envelopes, but a single-argument arctangent returns a value only between −90 and +90 degrees. Taken alone, arctan(V_s / V_c) therefore places the shaft within half a revolution and leaves an ambiguity of 180 degrees. To recover absolute position over a full revolution the converter must also use the individual signs of V_s and V_c to select the quadrant, which is the two-argument arctangent that software libraries expose as atan2(V_s, V_c). A tracking converter reaches the same result implicitly, by driving the error between the measured signals and an internal angle estimate to zero across all four quadrants.

Resolver Types

Different resolver configurations serve various applications:

• Brushless resolvers - Use rotary transformers to couple excitation to rotor, eliminating wear items
• Variable reluctance resolvers - Rotor has shaped laminations rather than windings, improving reliability
• Multi-speed resolvers - Provide multiple electrical cycles per revolution for increased resolution

Resolver-to-Digital Converters

Resolver-to-digital converters (RDCs) process resolver signals to produce digital position and velocity outputs. Key functions include:

• Excitation generation - Provides the sinusoidal reference signal, typically 2 kHz to 20 kHz
• Signal demodulation - Extracts position-dependent amplitudes from modulated signals
• Angle calculation - Computes arctangent using tracking converter or direct computation
• Velocity derivation - Calculates speed from position changes or internal tracking loop

Modern RDCs provide selectable 10-bit to 16-bit resolution, and the setting is a trade-off rather than a free choice. Each added bit halves the angular step the tracking loop must resolve, so the maximum rate at which the converter can follow the shaft falls as resolution rises. Analog Devices gives its AD2S1210 variable-resolution converter a headline maximum tracking rate of 3125 revolutions per second, or 187,500 RPM; that figure belongs to the lowest of its four resolution settings, and selecting 16 bits lowers the ceiling by more than an order of magnitude. Integrated RDC chips fold excitation generation, demodulation, and the tracking loop into one device, which simplifies resolver interface design considerably.

Resolver Advantages

Resolvers offer significant benefits for demanding applications:

• Temperature range - Operate reliably from −55 to +155 degrees Celsius, and specialized designs beyond that
• Contamination immunity - Sealed construction resists oil, dirt, and moisture
• Shock and vibration - No fragile components, withstand severe mechanical stress
• EMI immunity - High-level analog signals resist electromagnetic interference
• Absolute position - Position known immediately after power-up
• Long cable runs - High-level differential signals survive tens of meters of shielded cable, far beyond what a logic-level encoder output tolerates. Cable capacitance does shift the phase of the returned signals relative to the excitation, so a long run requires the converter or its firmware to compensate for that shift

Thermistors

Thermistors are temperature-sensitive resistors commonly embedded in motor windings. Two types are used:

• PTC (Positive Temperature Coefficient) - Resistance increases sharply at a defined temperature, providing simple overtemperature detection. Trip points are chosen to suit the thermal class of the winding insulation, with 130, 140, and 150 degrees Celsius among the common values
• NTC (Negative Temperature Coefficient) - Resistance decreases continuously with temperature, enabling analog temperature measurement across a range

PTC thermistors are typically used with relay circuits that disconnect motor power when resistance exceeds a threshold. NTC thermistors interface with analog-to-digital converters for continuous temperature monitoring and thermal modeling.

Resistance Temperature Detectors

RTDs use the predictable resistance change of metals with temperature. Platinum RTDs (Pt100, Pt1000) offer excellent accuracy and stability for precision temperature measurement.

RTD characteristics include:

• Linearity - More linear than thermistors over wide temperature ranges
• Accuracy - IEC 60751 defines the Class A tolerance as ±(0.15 + 0.002|t|) degrees Celsius, where t is the measured temperature. That is ±0.15 degrees Celsius at 0 degrees Celsius and ±0.35 degrees Celsius at 100 degrees Celsius, so the tolerance widens with temperature rather than holding at a single figure
• Stability - Maintain calibration over years of service
• Self-heating - Measurement current causes slight temperature errors

RTDs require more sophisticated interface circuits than thermistors, typically using three-wire or four-wire connections to compensate for lead resistance.

Thermocouples

Thermocouples generate voltage proportional to temperature difference between the sensing junction and reference junction. While less common in motors than thermistors or RTDs, thermocouples offer:

• Wide temperature range - Type K covers roughly −200 to +1260 degrees Celsius, wider than any thermistor or RTD. The upper figure applies to heavy-gauge wire; fine-gauge assemblies, and the sheath and insulation around them, are rated for continuous service well below it
• Self-powered - No excitation current required
• Small size - Junction can be very compact

Thermocouple signals require cold junction compensation and amplification, adding complexity to the interface circuit.

Thermal Modeling

Beyond direct measurement, advanced motor controllers implement thermal models that estimate internal temperatures from current, speed, and ambient temperature. These models protect against thermal runaway in applications where direct sensing is impractical or during transient overloads.

Accelerometers

Accelerometers measure vibration acceleration and are the most common sensors for motor condition monitoring. Technologies include:

• Piezoelectric accelerometers - A piezoelectric crystal produces a charge proportional to acceleration, which a built-in or external charge amplifier converts to a voltage. General-purpose industrial units cover roughly 1 Hz to 10 kHz within a few decibels, while small, high-resonance designs extend to several tens of kHz. They are rugged enough for permanent installation but cannot read a steady acceleration
• MEMS accelerometers - Micromachined silicon sensors provide lower cost for integrated monitoring systems. They typically cover DC to several kHz with adequate sensitivity for most applications
• Capacitive sensing - Most MEMS devices measure the displacement of their proof mass as a change in capacitance against fixed electrodes. Unlike a piezoelectric element, a capacitive sensor responds down to DC, so the same device reads steady acceleration and shaft tilt as well as vibration

Key accelerometer specifications include:

• Sensitivity - 100 mV/g is the most common rating for general-purpose industrial accelerometers; 10 mV/g units are chosen where high amplitudes would otherwise saturate the amplifier
• Frequency range - Must capture bearing defect frequencies, typically requiring response to at least 10 kHz
• Mounting - Stud mounting provides best frequency response; adhesive and magnetic mounts trade convenience for reduced bandwidth

Velocity Sensors

Velocity sensors measure vibration velocity, which correlates well with vibration severity in rotating machinery. Moving coil velocity sensors use a coil suspended in a magnetic field, generating voltage proportional to velocity.

Velocity sensors offer:

• Direct velocity output - No integration required for velocity-based analysis
• Self-generating - No power supply needed
• Limited frequency range - Typically 10 Hz to 1000 Hz, adequate for general machinery monitoring but short of the bearing defect frequencies an accelerometer reaches

Proximity Probes

Eddy current proximity probes measure shaft displacement directly, detecting shaft runout, bearing wear, and rotor eccentricity. They are essential for monitoring large machines with fluid film bearings.

Proximity probe systems include:

• Probe - Coil that induces eddy currents in the shaft
• Driver electronics - Provides excitation and signal conditioning
• Installation gap - Probes require precise positioning relative to shaft

Vibration Analysis Techniques

Raw vibration signals are processed to extract diagnostic information:

• Overall level - RMS or peak vibration indicates general machine condition
• Spectrum analysis - FFT reveals frequency components related to specific faults (imbalance at 1x, misalignment at 2x, bearing defects at characteristic frequencies)
• Envelope analysis - Detects modulation caused by bearing and gear defects
• Time waveform - Reveals impacting and looseness not apparent in spectra

Strain Gauge Torque Sensors

Strain gauge sensors measure the twist of a shaft under torque. Strain gauges bonded to the shaft deform with the shaft surface, changing their electrical resistance proportionally to applied torque.

Configurations include:

• Gauge orientation - Pure torsion puts its principal tensile and compressive strains at +45 and −45 degrees to the shaft axis, so torque gauges are bonded along those two directions and not along or around the axis
• Full bridge - Four active gauges, two on each principal direction, add the torsional strains around the bridge while cancelling the bending and axial strains that act on all four alike. The arrangement also cancels the first-order effect of a uniform temperature change and gives the maximum sensitivity
• Dual bridge - Two complete bridges on the same shaft section, read independently for redundancy, for a second measuring range, or to average out residual bending
• Reaction torque sensors - Measure torque through stationary housing, simplifying signal connection
• Rotating torque sensors - Inline sensors require slip rings or wireless telemetry for signal transmission

Rotary Torque Transducers

Purpose-built rotating torque transducers incorporate strain gauges, signal conditioning, and data transmission in an integrated assembly. Signal transmission methods include:

• Slip rings - Sliding contacts transfer power and signals but add friction and require maintenance
• Rotary transformers - Contactless inductive coupling for power and analog signals
• Wireless telemetry - Digital transmission eliminates all mechanical coupling, modern sensors achieve high bandwidth and accuracy

Magnetoelastic Torque Sensors

Magnetoelastic sensors detect torque through changes in magnetic properties of ferromagnetic materials under stress. A magnetic field is applied to the shaft, and torque-induced stress changes the field pattern, which is detected by sensors surrounding the shaft.

Advantages include:

• Non-contact operation - No physical connection to rotating shaft
• High overload capacity - No delicate strain gauges to damage
• Potentially lower cost - Shaft can serve as the sensing element

Optical Torque Measurement

Optical methods measure the angular twist between two points on a shaft under torque. A light beam or encoder tracks the relative rotation between shaft sections, with the angle proportional to applied torque.

This approach enables:

• High bandwidth - No mechanical elements limit dynamic response
• Non-contact sensing - No loading of the shaft
• Integration with encoders - Same shaft can provide position and torque

Torque Estimation from Motor Current

Motor current is proportional to torque in many motor types, enabling torque estimation without dedicated sensors. For DC motors, torque is K_t × I_a, the torque constant times the armature current. In SI units the torque constant in newton meters per ampere and the voltage constant in volt seconds per radian are numerically equal for an ideal machine, so one measurement calibrates both. For AC motors, the torque-producing current component can be calculated using vector control techniques.

Current-based torque estimation offers:

• Zero additional hardware - Uses existing current sensors
• Lower accuracy - Affected by motor parameters and temperature
• Practical for many applications - Adequate where high precision is not required

Performance Requirements

• Resolution - Determines minimum detectable position change and velocity measurement at low speeds
• Accuracy - Affects positioning precision and velocity control quality
• Bandwidth - Must exceed control loop requirements for stable operation
• Repeatability - Critical for applications requiring return to defined positions

Environmental Factors

• Temperature range - Optical encoders are typically limited to about −40 to +100 degrees Celsius by their LEDs, photodetectors, and bearing lubricants, while resolvers hold the −55 to +155 degrees Celsius range noted earlier and specialized designs go further
• Contamination - Sealed or magnetic sensors for dirty environments
• Shock and vibration - Resolvers and magnetic encoders outperform optical types
• EMI exposure - Shielding and differential signals for high-interference environments

System Integration

• Interface compatibility - Match sensor output to controller input requirements
• Cable length - Consider signal degradation over distance; differential outputs for long runs
• Mounting constraints - Available space, shaft access, and mechanical attachment method
• Power requirements - Supply voltage and current consumption

Cost Considerations

• Sensor cost - Spans orders of magnitude. Hall switches are the cheapest position feedback available, incremental encoders cost substantially more, and precision resolvers with their converters and multi-turn absolute encoders sit at the top of the range
• Interface electronics - Some sensors require significant signal conditioning
• Installation cost - Sensorless techniques eliminate sensor installation entirely
• Maintenance - Brushed tachometers and slip ring sensors require periodic service

Encoder Issues

• Intermittent counts - Check connections, inspect disc for contamination, verify signal integrity with oscilloscope
• Position drift - Indicates missed counts from noise, speed exceeding capability, or failing sensor
• Noisy signals - Verify proper shielding and grounding, check for damaged cables
• Channel imbalance - Quadrature signals should be symmetric; asymmetry indicates alignment or contamination issues

Hall Sensor Problems

• Motor runs rough - One or more Hall sensors may have failed; turn the shaft slowly and confirm that all six valid state combinations appear in order, and that the all-high and all-low patterns never do
• Motor fails to start - Verify Hall signals present and correct timing relationship
• Temperature-related failures - Hall sensors may drift or fail at temperature extremes

Resolver Diagnostics

• Position jitter - Check excitation signal quality and amplitude
• Offset errors - Verify resolver alignment and RDC calibration
• Loss of tracking - Speed may exceed RDC tracking rate capability