Q9Advanced Power Electronics

(a) Control Circuits and Voltage Mode Control

Control circuits in a switch-mode power converter are the feedback and gate-drive-generation circuitry responsible for continuously sensing the converter's actual output condition (typically output voltage, sometimes also output or inductor current), comparing it against the desired reference/setpoint value, and generating the appropriately-timed switching (gate-drive) signals needed to regulate the actual output to closely track the desired reference despite variations in input line voltage, load current demand, and component parameter drift/tolerance — essentially forming the same fundamental negative-feedback control-loop structure found throughout control engineering, specifically adapted to the pulse-width-modulated or frequency-modulated switching nature of power electronic converters.

Voltage mode control is the simplest and most traditional control scheme for switch-mode converters: the actual output voltage is sensed (often through a resistive divider and/or isolation amplifier for isolated converter topologies), compared against a stable reference voltage using an error amplifier, and the resulting error signal directly sets the duty cycle of the PWM comparator (by comparing this error voltage against a fixed-frequency sawtooth/ramp waveform) — a single feedback loop (voltage loop only) is used, with the error amplifier's compensation network (typically a combination of proportional, integral, and sometimes derivative/lead-lag compensation elements) shaped to achieve the desired closed-loop stability margin and transient response for the specific converter's power-stage transfer function.

Voltage mode control's key advantages are its relative simplicity (only one feedback variable, output voltage, needs to be sensed and processed) and generally good noise immunity in the feedback path (compared to inductor-current sensing, which can be more susceptible to switching noise pickup); its key disadvantages are that its dynamic response to sudden input-voltage-line disturbances is comparatively slower (since the control loop only 'sees' and reacts to the resulting output voltage disturbance after that disturbance has already propagated through the entire power stage's dynamics, rather than directly and immediately sensing and compensating for the input disturbance itself), and the compensation network design must carefully account for the converter power stage's own transfer function characteristics (which vary with operating point and can include a troublesome right-half-plane zero in certain boost/buck-boost-derived topologies), making voltage-mode compensator design somewhat more involved than for the alternative current-mode control scheme, which inherently simplifies the power stage's effective transfer function by incorporating an inner, faster current-control loop.

Typical compensation network structure: a practical voltage-mode error amplifier compensation network is commonly implemented as a Type II or Type III compensator (built from a single operational amplifier with a specific arrangement of resistors and capacitors forming its feedback network), providing high DC gain (for accurate steady-state output voltage regulation with minimal residual error), a controlled bandwidth (crossover frequency, typically set to a fraction of the switching frequency to maintain adequate control-loop stability margin while still providing reasonably fast transient response), and appropriately placed compensation zeros and poles to counteract the phase lag introduced by the power stage's own output LC filter resonance — designing this compensation network correctly, so that the overall closed loop achieves adequate phase margin (typically targeted around 45-60°) and gain margin across all expected operating conditions (input voltage range, load current range), is the central practical challenge of voltage-mode controller design, generally requiring either careful analytical loop-gain calculation or iterative simulation/bench-top tuning by the power supply designer.

(b) Resonant DC Power Supplies

Resonant DC power supplies apply the resonant switching principles discussed for series and parallel resonant inverters (Q.4 above) specifically to isolated DC-DC power conversion applications: a switching bridge generates a high-frequency square-wave voltage, which is applied to a resonant tank circuit (series, parallel, or the increasingly popular LLC combination resonant tank) coupling through an isolation transformer to a high-frequency rectifier and output filter, ultimately producing a regulated, isolated DC output voltage.

By operating the switching devices at (or appropriately relative to) the resonant tank's natural resonant frequency, resonant DC power supplies achieve zero-voltage or zero-current switching for the primary-side switches (and often also for the secondary-side rectifier diodes/synchronous rectifiers), essentially eliminating the switching losses that would otherwise limit the maximum practical switching frequency of a conventional hard-switched converter — this allows resonant DC power supplies to operate at substantially higher switching frequencies than equivalent hard-switched designs, in turn allowing the use of smaller, lighter, less costly magnetic components (transformer, inductors) and filter capacitors for a given power rating, giving resonant DC power supplies a significant power-density and (at sufficiently high power levels) efficiency advantage over conventional hard-switched converter designs, at the cost of the more complex resonant-tank design and frequency-modulation-based control scheme discussed earlier, and are consequently widely used in modern high-power-density switch-mode power supplies for computing, telecommunications, and other applications where power density and efficiency are at a premium.

Current-mode control as a complementary/alternative scheme: as a further point of comparison to the voltage-mode control described in part (a), current-mode control adds an inner, faster feedback loop directly sensing the switch or inductor current (in addition to the outer voltage loop), comparing this sensed current against a reference set by the voltage loop's error amplifier output, and terminating each switching cycle once the sensed current reaches this reference level — this inner current loop effectively simplifies the power stage's dynamic behavior as seen by the outer voltage loop (reducing the effective order of the power-stage transfer function and eliminating the problematic right-half-plane zero issue in certain topologies), provides inherent, cycle-by-cycle current limiting for improved overload/short-circuit protection, and offers faster transient response to input line disturbances (since a sudden change in input voltage is immediately reflected as a change in the rate of rise of the sensed inductor current, allowing much faster corrective action than voltage-mode control's comparatively slower output-voltage-only feedback path can provide) — trade-offs that are frequently weighed against voltage-mode control's simpler single-loop design when selecting the most appropriate control scheme for a given switch-mode converter application.

Practical implementation of resonant DC supply control: since output regulation in a resonant DC power supply is achieved through switching-frequency variation rather than duty-cycle variation, the associated control circuit typically employs a Voltage-Controlled Oscillator (VCO) whose output frequency is directly set by the error amplifier's output signal (representing the sensed output voltage error), rather than the fixed-frequency PWM comparator used in conventional hard-switched voltage-mode or current-mode control schemes — this frequency-modulation-based control approach requires the associated magnetic components (transformer, resonant inductor) and EMI filter design to correctly and efficiently accommodate the full required switching-frequency operating range (rather than being optimized for only a single, fixed switching frequency as in conventional PWM-controlled converters), an additional design consideration specific to resonant, frequency-modulated converter topologies.