System context & use-case map

A servo power stage does not live in isolation. It sits between the front-end PSU and the motor, and shares responsibilities with the control board, protection, safety and sensing subsystems. This section anchors the servo power stage in that stack so each topic is covered on the right page without overlap.

User stories & line-voltage tiers

Servo drives for different line voltages and power levels place very different demands on the power stage. Three representative scenarios help define the design space.

• 48 V robot joint: 24–48 V DC bus, typically fed from a low-voltage PSU. Phase currents are in the 3–10 A continuous range with peaks that can reach 2–3 times rated current. Axes sometimes share a DC link in compact multi-joint modules.
• 230 V single-axis servo: 230 Vac mains through a rectifier or PFC front-end producing roughly 325–400 V DC. Phase currents often sit around 5–15 A continuous with short overloads during acceleration, and each axis may have its own DC link.
• 400/480 V multi-axis cabinet: 400/480 Vac three-phase mains feeding a PFC/rectifier that produces a DC link in the 560–800 V range. Individual axes can see 10–40 A continuous and share a large DC link with common brake and protection hardware.

Servo drive stack split

The complete servo drive can be split into a small set of hardware blocks. The servo power stage is only one of them and should not absorb responsibilities that are handled better on dedicated pages.

• Front-end PSU (PFC / rectifier + DC-link): AC mains interface, rectification, power-factor correction and bulk energy storage on the DC link. Detailed PFC and AC/DC design is covered under the front-end PSU topics.
• Servo power stage (this topic): Three-phase bridge legs, power devices, DC-link hooks, gate-drive circuits and the fast protection chain directly around the bridge.
• Control / FOC MCU board: PWM generation, position and speed loops, observers and signal processing for encoders, resolvers and current feedback.
• Protection, brake & STO: Pre-charge, brake chopper, safe torque off paths and higher-level safety logic governed by standards such as IEC 61508 or ISO 13849.
• Sensing, logging, HMI & comms: Current, voltage and temperature AFEs, local human-machine interface, event logging, predictive maintenance and industrial communications.

The servo power stage section links to those other topics where appropriate so gate-drive and bridge layout discussions stay focused and do not duplicate sensing, safety or PSU content.

Interfaces of the servo power stage

The servo power stage touches several critical interfaces. Each interface has a defined scope on this page and a handover point to other topics.

• DC-link interface: Voltage range, ripple and peak current drawn by the three-phase bridge. This page focuses on how the bridge and its layout relate to the DC-link; sizing and control of the front-end PSU are covered elsewhere.
• Motor-side interface: U, V and W terminals, cable length, dv/dt stress and common-mode behaviour. The power stage view highlights how these factors affect device choice and protection needs.
• Control interface: PWM inputs, enable lines, fault outputs and analog or digital feedback paths for current and bus voltage. Detailed loop design and algorithms are handled under control and FOC topics.
• Protection and safety interface: Trip signals, brake connections and STO-related hooks that originate near the bridge but extend into dedicated protection and safety modules.
• Auxiliary supplies and sensing: Local supplies for gate drivers, isolated bias rails and temperature-sensing points on power modules or heat sinks, which are later consumed by thermal and logging blocks.

Power-switch platform, ratings & stress

Selecting devices for a servo power stage begins with line voltage, DC-link levels and realistic current envelopes. The goal is to converge on a sensible MOSFET, IGBT, IPM or SiC platform before detailed loss calculations and layouts start.

Line voltage, DC-link and device class

A few voltage tiers cover most industrial servo applications. For each tier, the DC-link level and margin drive the choice of device voltage rating and technology.

• 24/48 V servo axes: DC links in the 24–60 V range with overshoot margins point toward 80–100 V MOSFET families. Low-voltage MOSFETs offer low R_DS(on) and are well suited to compact robot joints and small actuators.
• 120 Vac systems: Rectified 120 Vac produces roughly 170–200 V DC. Devices in the 200–250 V class give comfortable margin for transient and ringing effects in compact servo drives.
• 230 Vac servo drives: Rectified 230 Vac is typically managed as a 325–400 V DC link. Devices rated 600–650 V are common, and both MOSFET and IGBT options exist in this band depending on power and switching-frequency targets.
• 400/480 Vac multi-axis cabinets: 400/480 Vac mains yield DC links in the 560–800 V range. Devices in the 1200 V class are typical, and SiC solutions become attractive where efficiency, switching frequency or thermal headroom are tight.

Overvoltage and regeneration events often justify DC-link margins around 1.15–1.20 times the highest planned operating voltage. Device ratings should be selected with those conditions in mind rather than nominal values alone.

Conduction and switching loss planning

Servo workloads include long periods of partial load, bursts of acceleration and frequent regeneration. Loss planning therefore needs more than a single operating point.

• Estimate continuous and peak phase-current envelopes from torque-speed profiles rather than from motor nameplate data alone.
• Use those envelopes to approximate conduction loss from R_DS(on) or V_CE(sat) values, taking temperature and parallel devices into account.
• Combine switching frequency, dv/dt, snubber strategy and expected operating duty cycles to estimate switching losses across key operating regions.

As a rule of thumb, when conduction losses consume most of the thermal budget, lower R_DS(on) or a module-based approach is often more effective than further changing switching conditions. When switching losses dominate, device technology, topology and PWM strategy become the main tuning levers.

Discrete devices, modules and IPMs

Several hardware integration levels are available for servo power stages. Each level trades freedom in layout and protection strategy against simplicity and robustness.

• Discrete MOSFETs or IGBTs: Offer maximum flexibility in layout, device mix and current-sharing schemes. They demand careful attention to loop inductance, Kelvin-source routing and thermal spreading.
• Half-bridge and three-phase modules: Provide optimized internal layouts and well-characterised thermal paths. They reduce board-level parasitics but constrain device combinations and sometimes footprints.
• Intelligent power modules (IPMs): Combine power devices, gate drives and basic protections in a single package. These are attractive for small and medium-power servo axes and multi-axis stacks, but many protection behaviours become fixed by the module family.

IPMs simplify early designs and compact axes, while discrete and module-based approaches remain important for high-power systems or projects with non-standard protection and sensing requirements.

SOA and I²t coordination with upstream protection

Short-circuits and overloads stress power devices far beyond normal operating points. Device safe operating area (SOA) curves and I²t capabilities must align with fuses or eFuses upstream.

• Check device SOA for the specific DC-link voltage and fault durations that may occur in the servo system, including in-cabinet and motor-cable faults.
• Compare the energy let-through of upstream fuses or eFuses with device I²t margins so faults are cleared before devices are driven beyond safe limits.
• Use this coordination as a boundary condition for later decisions on DESAT thresholds, overcurrent comparators and gate-drive fault responses.

Detailed selection of fuses and eFuses is handled in the dedicated protection and distribution topics; the servo power stage view concentrates on ensuring that device ratings and upstream protection are consistent.

Phase-leg layout, parasitics & DC-link hooks

A servo phase leg is more than a schematic symbol. Physical placement of devices, DC-link capacitors and current shunts defines the parasitic inductances and resistances that shape switching behaviour, overshoot, losses and protection margins.

Canonical phase leg structure

A canonical servo phase leg combines a high-side device, a low-side device, freewheel paths, a short link to the DC-link rails and a phase node that feeds the motor cable. One or more current shunts can sit on the DC- side, the DC+ side or in-line with the phase output.

• High-side and low-side switches share the same DC-link, forming the core bridge leg.
• Body diodes or external diodes provide freewheel paths for inductive motor currents.
• The phase output node is kept compact and connects to the motor cable and filters.
• Current shunts may sit at low-side return, on the DC-link or in-line with the phase.

Low-side shunts are simple and referenced to ground, while in-line shunts expose phase current more directly. High-side and DC-link shunts support protection and diagnostics but typically require dedicated current-sensing AFEs covered in the current-sensing topics.

High di/dt loop minimization

The dominant high di/dt loop in a phase leg includes the high-side device, the low-side device, the DC-link capacitor connections and the short section of the phase node between them. In industrial servo systems this loop can switch tens to hundreds of amperes at several kilovolts per microsecond.

• DC-link film capacitors should be placed very close to the device DC terminals.
• The path between high-side and low-side devices is kept short and compact.
• The phase copper directly above or below the devices should not fan out before leaving the bridge area.

Sensitive signals such as gate-drive traces, current-sense leads and encoder or communication lines should not pass through this loop. Routing these signals around the loop reduces common-mode coupling and decreases the risk of false switching events or measurement disturbances.

DC-link capacitors and regeneration paths

The DC-link capacitor bank acts as the local energy reservoir for the servo bridge and absorbs regeneration energy from the motor. Bulk capacitors provide energy storage, while film capacitors support the high-frequency switching currents directly at the phase legs.

• Bulk electrolytic capacitors can sit slightly further from the bridge but share the same DC path.
• Film capacitors are mounted close to the device DC terminals and close the high di/dt loop.
• Regeneration current flows from the motor, through the bridge and into the DC-link and brake path.

During braking and regeneration, power devices, DC-link capacitors and brake chopper components share the thermal and electrical stress. Detailed brake chopper sizing and protection planning is handled in the dynamic braking topic; this section focuses on how DC-link connections interact with the phase leg.

Multilayer layout and shunt placement

Multilayer PCBs allow power and sensing paths to be separated while keeping high-current loops compact. The copper stack should reserve clear roles for top, inner and bottom layers and give current shunts well-defined Kelvin connections.

• Top layers usually carry device pads and the most critical high di/dt copper shapes.
• Inner layers can provide DC-link and return planes but should not spread the high di/dt loop excessively.
• Bottom layers can help close current loops and move small-signal traces away from power areas.

Shunt resistors need Kelvin sense connections that route separately from the main power copper. The power pads carry the servo current, while sense pads connect directly to the current-sensing AFE or ADC reference without sharing high-current paths. This layout discipline is essential for low-noise current feedback and robust protection signals.

Typical patterns include phase legs placed in a straight row, dual-leg boards for two axes and module-based layouts with power modules central and gate-drive circuitry at the edge or on a daughtercard. Each pattern must respect the same loop and shunt rules.

Gate-drive topology, dead-time & fast protection chain

Gate-drive choices link directly to device ratings, switching dynamics and short-circuit behaviour. This section connects bootstrap and isolated drive topologies, dead-time planning and the fast protection chain around the servo phase leg.

Bootstrap and isolated gate drivers

Bootstrap drivers and isolated drivers both appear in servo power stages. The right choice depends on DC-link voltage, switching frequency, motor cable length and EMC constraints.

• At DC-link levels below 100 V with short cables and moderate frequencies, bootstrap drivers are often sufficient.
• In the 300–400 V range, bootstrap solutions remain common, but long cables, high dv/dt and strict EMC requirements favour isolated drivers.
• At higher DC-link levels above 600–700 V, isolated gate drivers and modules with robust isolation and high CMTI ratings are usually preferred.

Bootstrap drivers are limited by their maximum on-time and the need for regular refresh pulses. Isolated drivers rely on isolated supplies but provide clear isolation boundaries and high CMTI ratings, which are important for long motor cables and aggressive dv/dt edges.

Gate-drive timing and dead-time planning

Dead-time sits between high-side and low-side switching events and must be based on device timing information and real measurements. Excessive dead-time increases distortion and torque ripple; insufficient dead-time risks shoot-through.

• Extract turn-on and turn-off timing (minimum, typical and maximum) from device datasheets across temperature and current ranges.
• Translate those timings into PWM timer dead-time settings in the MCU or DSC, adding margin for worst-case switching behaviour.
• Ensure that only one element in the chain inserts dead-time so that the MCU timer and gate driver do not overlap or amplify delay.

Practical checks include verifying propagation delay mismatch between high-side and low-side channels and measuring phase and gate waveforms on the bench to confirm that dead-time is sufficient yet not excessive across operating conditions.

Short-circuit modes and DESAT / overcurrent detection

Servo drives must survive faults such as phase-to-phase shorts, phase-to-DC shorts and wiring errors. Short-circuit detection combines fast DESAT monitoring near the devices with shunt-based overcurrent detection in the current-sensing path.

• DESAT detection is tightly coupled to the gate driver and offers very fast turn-off in response to short-circuit currents.
• Shunt-based overcurrent detection supports both protection and control functions such as current limiting and diagnostics.
• Trip thresholds must align with device SOA and I²t margins so that protection reacts before power devices are overstressed.

Soft turn-off strategies reduce di/dt and overvoltage during fault clearing, while hard turn-off prioritises immediate current interruption. The choice depends on device robustness and system-level protection goals.

Fault propagation and reset strategy

A complete protection chain must propagate fault information from the devices to the control and safety logic and define when and how the system is allowed to restart.

• Gate drivers aggregate DESAT and overcurrent events into fault outputs exposed to the control board.
• Control and safety logic decide whether to disable a single phase leg, a full bridge or multiple axes.
• Startup and reset sequences must avoid inadvertent turn-on during brownout, power-up or noisy conditions.

Repeated automatic restarts can create chattering fault behaviour and additional stress on devices. Many designs therefore use latching schemes that require an explicit reset action from the control or safety domain before gate drives are re-enabled.

Sensing hooks & cooperation with control algorithms

A servo power stage must expose the right sensing hooks to support FOC, fast protection, diagnostics and predictive maintenance. This section maps current and voltage sensing points, separates fast hardware protection from slower ADC-based loops and identifies signals that should feed logging and PdM engines.

Phase and DC-link current sensing points

Servo drives commonly combine low-side shunts, in-line shunts and DC-link sensing elements such as shunts, ΣΔ current transformers or isolated transducers. The choice affects FOC sampling windows, protection coverage and multi-axis DC-link monitoring.

• Low-side shunt: simple, ground-referenced measurement suitable for cost-sensitive and low-voltage designs.
• In-line phase shunt: direct phase current measurement that supports higher accuracy and overload capability.
• DC-link sensing: shunts, ΣΔ CTs or transducers give total current for bus protection and multi-axis monitoring.

FOC sampling must align with device conduction intervals so that ADC windows fall within stable current plateaus. Low-side shunts are tied to specific switching states, while in-line shunts support valid samples in more PWM sectors. DC-link current sensing helps supervise shared DC buses, regeneration energy and cabinet-level protection.

Bus voltage sensing and filtering

Bus voltage sensing typically uses a resistor divider between DC+ and DC- near the DC-link capacitor nodes. The divider feeds an ADC input or comparator path and must balance leakage, power dissipation and accuracy, especially at high DC-link voltages.

An RC filter on the sense node defines bandwidth. The filter is slow enough to smooth PWM ripple and spikes, yet fast enough to follow regeneration and load-step events on a millisecond time scale. Overvoltage and undervoltage comparators can sit directly on the power board for fast local protection or on the control board for centralised safety logic. Detailed amplifier, ADC and comparator selection is handled in the bus and phase voltage sensing topics.

Fast hardware comparators and slower ADC-based loops

Protection schemes combine very fast hardware actions with slower software supervision. Nanosecond to microsecond responses come from DESAT detectors and high-speed current or voltage comparators, while millisecond-scale supervision relies on ADC sampling and FOC or control loops.

• Fast hardware paths interrupt short-circuit events before devices leave their SOA.
• ADC-based loops track overload, imbalance and long-term stress conditions.
• Protection design should not rely solely on firmware; device-level hardware action is the first line of defence.

Hardware protection provides immediate gate turn-off and a fault indication, while software decides about derating, restart policies and error logging. The sensing architecture must clearly separate these roles in both timing and signal paths.

Diagnostic signals for PdM and logging

Several signals from the power stage are valuable for predictive maintenance and data logging. These include phase currents, DC-link voltage and ripple, key temperature points and fault event counters from gate drivers and comparators.

• Phase current channels for imbalance detection, overload history and torque estimation.
• DC-link voltage and ripple as indicators of loading, regeneration and capacitor health.
• Module, heatsink and capacitor temperatures to monitor thermal stress over time.
• Latched fault flags and counters for overcurrent, overvoltage and DESAT events.

These signals are exported to logging and PdM subsystems, where algorithms convert raw measurements into lifetime estimates, fault trends and service indicators. The servo power stage defines which hooks are available and how cleanly they reach the control and monitoring domains.

Thermal, mechanics & multi-axis integration

Thermal design, mechanical integration and multi-axis cabinet layout determine whether a servo power stage survives real operating conditions. This section traces thermal paths, highlights creepage and clearance considerations and shows how individual axes connect to shared DC-link, braking and safety infrastructure.

Thermal paths and sensor placement

Heat flows from semiconductor junctions into leadframes or substrates, through copper, thermal interface materials and heatsinks, and finally into the cabinet air and environment. Modules and IPMs often concentrate hot spots near chip clusters, screw locations or terminal blocks.

Temperature sensors such as NTCs, RTDs or integrated sensors should monitor at least the main device thermal path, DC-link capacitors and critical heatsink or airflow regions. Measurements support both fast overtemperature protection and long-term stress analysis.

PCB stack-up and creepage / clearance

Servo power stages for 230 V and 400/480 V lines must respect minimum creepage and clearance distances between high-voltage and low-voltage domains. PCB stack-up, slots and coating techniques are used to achieve the required distances without excessive board size.

• Keep high-voltage copper and low-voltage copper on separated layers or well-defined regions.
• Use slots, cut-outs and routed channels to extend creepage paths across the board surface.
• Consider conformal coating, barriers and mechanical shields in areas with tight spacing.

Detailed clearance numbers come from applicable standards. Design reviews should examine creepage and clearance in three dimensions, not only from the top view, especially around terminals, heatsinks and mounting hardware.

Multi-axis cabinet patterns

Industrial motion cabinets often host several servo axes sharing a common DC-link, braking unit and safety functions. Power stages may appear as individual axis boards, dual or triple-axis assemblies or plug-in axis modules on a shared backplane.

• Single-axis boards connected to a shared DC-link busbar inside the cabinet.
• Multi-axis boards combining two or three bridges on one PCB with a common DC-link input.
• Plug-in modules that slide into a backplane providing DC-link, control, brake and STO lines.

Each axis module must define clear interfaces for DC-link in and out, brake connections, STO or axis-enable signals and fault reporting. The servo power stage layout should reflect how these interfaces will connect to cabinet busbars and backplanes.

Wiring, terminals and EMC side effects

Terminal layout defines motor and DC-link cable routing and has a strong impact on EMC behaviour. Motor terminals and protective earth, DC-link terminals and brake connections must be positioned with insulation, routing and filtering in mind.

• Keep motor power terminals and PE grouped and separated from low-voltage control connectors.
• Route motor cables and control cables on different paths to limit coupling from high dv/dt edges.
• Reserve locations for common-mode chokes, filters or dV/dt filters where EMC functions are not on the power board.

A dedicated EMC subsystem topic covers filter structures and compliance testing. The servo power stage focuses on providing robust mechanical mounting, terminal placement and routing options that support those EMC strategies.

Design Checklist (System-to-Layout Flow)

The checklist follows the natural flow from system ratings into device selection, phase-leg layout, gate drive and protection, sensing hooks and multi-axis integration. Each block can be walked through during design reviews so gaps can be caught before prototypes are built.

System and ratings

• DC-link nominal voltage, worst-case variation and regeneration margin are documented.
• Motor current envelopes for continuous, peak and short-term overload operation are understood.
• Shared DC-link usage across axes and the expected diversity factor are defined.
• Line-voltage tiers and isolation requirements are aligned with safety and standards targets.

Device platform and topology

• The chosen device class (MOSFET, IGBT, SiC, IPM) matches DC-link tier and power range.
• Conduction and switching loss splits are evaluated at representative operating points.
• Short-circuit withstand time and SOA limits are coordinated with upstream fuses or breakers.
• Module or IPM selections support required thermal paths, mounting options and service strategy.

Phase-leg and layout

• High di/dt loops are explicitly drawn on the layout and kept compact.
• DC-link film capacitors are placed close to device terminals along the high di/dt path.
• Current shunt locations are consistent with the chosen FOC sampling scheme.
• Creepage and clearance distances reflect the line-voltage tier and cabinet environment.

Gate-drive and fast protection

• The bootstrap versus isolated gate-driver decision is documented against voltage and cable constraints.
• Dead-time settings are derived from worst-case device timings and validated on the bench.
• DESAT and shunt-based overcurrent thresholds align with SOA and fuse or breaker characteristics.
• Fault signals from drivers reach control and safety logic with clear polarity and timing.

Sensing and logging hooks

• Phase and DC-link current sensors are positioned so FOC loops and fast protections do not conflict.
• Bus voltage sensing bandwidth and comparators cover control and protection requirements.
• Key temperatures and fault indicators are routed to control and logging domains.
• Signal integrity for sensing lines is maintained around high di/dt and high dv/dt regions.

Thermal and multi-axis integration

• Temperature sensors monitor the main device thermal path and DC-link capacitors.
• Shared heatsinks or cabinet volumes have defined derating and fan control strategies.
• Shared DC-link and brake wiring are reviewed for worst-case regeneration events.
• Axis modules have clear terminals and harnessing plans for cabinet integration.

Vendor and IC Role Mapping

Instead of listing part numbers, this mapping groups IC roles used in servo power stages and highlights what to look for in vendor portfolios. Detailed selection happens in dedicated topics for gate drivers, current sensing, voltage sensing and power management.

High-speed isolated gate drivers with DESAT

For high-voltage servo stages and long motor cables, isolated drivers with reinforced isolation, high CMTI ratings and integrated DESAT detection are a primary focus. Vendor lines labelled for inverters, servo drives or motor control at 400 V and 600 V DC-link levels are most relevant.

Bootstrap half-bridge drivers

For low to mid-voltage tiers with moderate cable lengths, bootstrap half-bridge drivers offer compact solutions. Suitable families specify clear maximum on-time limits, support the intended switching frequency and provide negative gate drive and robust undervoltage lockout behaviour.

Servo-focused IPM families

Integrated power modules that bundle power devices, gate drive and protection are attractive for compact servo axes. Portfolios that cover common line-voltage tiers, several current ratings and provide thermal and fault feedback interfaces map well to multi-axis cabinets and modular designs.

Current-sense amplifiers and sigma-delta modulators

Servo-oriented current-sense solutions emphasise bandwidth, noise and CMRR under fast PWM edges. Shunt amplifiers and sigma-delta modulators that support shunt, in-line and DC-link sensing, with automotive or industrial ratings, align well with servo power-stage requirements.

Temperature monitors and fan controllers

Temperature monitor ICs and fan controllers close the loop between measured thermal data and airflow control. Suitable families provide multiple sensor inputs, support NTC or RTD interfaces and offer PWM or voltage outputs for fans, matching shared heatsink and cabinet cooling strategies in multi-axis systems.

FAQs on Servo Power-Stage Planning and Selection

These twelve questions distil the main decisions around device choice, layout, gate drive, protection, sensing and cabinet integration. Answers stay compact so they are easy to reuse in design reviews, vendor discussions and structured data for search and documentation.