Charge Pump and Voltage Multiplier Circuits

Generating a voltage higher than your available power supply or creating a negative voltage from a positive one is a common challenge in electronics. Charge pump circuits solve this elegantly by using capacitors as temporary energy storage elements and switches to transfer charge, achieving DC-DC conversion without the magnetics required by inductor-based switchers. This makes them ideal for low-current, integrated applications where size, cost, and electromagnetic interference are critical constraints.

The Core Principle: Capacitor Charge Transfer

At its heart, a charge pump operates on a simple two-phase cycle: charge and transfer. Imagine a switched capacitor circuit: during the first phase (charge), a capacitor is connected directly across a voltage source, such as your system's $V_{DD}$, and charges to that potential. During the second phase (transfer), the capacitor is disconnected from the source and reconnected in series with another capacitor or the output node. By stacking its charged voltage atop another voltage, the potential at the output can be doubled, inverted, or otherwise multiplied.

The fundamental advantage is the absence of inductors and transformers. This allows charge pumps to be built entirely on a silicon chip using capacitors and transistors acting as switches. The primary trade-off is limited output current capability and efficiency, as charge transfer is not perfectly lossless. This makes them perfect for biasing circuits, driving small displays, or powering specific IC sub-blocks that require a voltage outside the main supply rail.

Basic Architectures: Voltage Doublers and Inverters

The simplest charge pumps demonstrate the core concept. A voltage doubler typically uses two phases and two capacitors. In phase one, a "flying" capacitor charges to the input voltage. In phase two, this charged capacitor is placed in series with the input source and the output capacitor. The input source and the flying capacitor's voltage add together, theoretically providing $2\times V_{IN}$ to the output. Real-world losses from switch resistance and capacitor leakage reduce this slightly.

A voltage inverter follows a similar switched process but reconfigures the connections to create a negative voltage. The flying capacitor charges to $+V_{IN}$ during the charge phase. During the transfer phase, its positive terminal is switched to ground, and its negative terminal is connected to the output capacitor. This forces the output capacitor to charge to a voltage of $-V_{IN}$. These doubler and inverter stages are the fundamental building blocks for more complex multiplier chains.

Cascaded Multipliers: The Dickson Charge Pump

For multiplication factors greater than two, stages are cascaded. The Dickson charge pump is a classic integrated circuit implementation for generating higher DC voltages. It consists of multiple diode-connected transistors (or active switches for better efficiency) and capacitors arranged in a ladder. A multi-phase clock drives the circuit, with each stage "pumping" charge to the next.

In operation, each capacitor alternately charges and then discharges into the next capacitor in the chain. With each transfer, the voltage is incrementally boosted. For an N-stage Dickson pump, the ideal output voltage is $V_{OUT}=(N+1)\cdot V_{IN}-N\cdot V_F$, where $V_F$ is the forward voltage drop across each diode or switch. Modern designs use MOSFETs as synchronous switches to minimize this drop, dramatically improving efficiency. The Dickson topology is favored in on-chip designs for its regularity and ease of layout.

High-Voltage Generation: The Cockcroft-Walton Multiplier

When very high voltages are needed, particularly from an AC source, the Cockcroft-Walton voltage multiplier (or Greinacher circuit) is the standard architecture. Unlike the Dickson pump which is typically driven by DC and clocks, this circuit is designed to be driven by an AC input, often sinusoidal.

It is built from a ladder of diodes and capacitors. Each stage consists of a capacitor and a diode. On alternate half-cycles of the AC input, different capacitors charge and then stack in series. The beauty of this circuit is that the voltage across each capacitor, except the first, is equal to the peak AC input voltage. For an N-stage Cockcroft-Walton multiplier, the ideal no-load output voltage is the peak input voltage multiplied by the number of stages. The governing equation for a full-wave configuration is: $$V_{OUT}=2N\cdot V_{PEAK}$$ where $N$ is the number of stages and $V_{PEAK}$ is the peak voltage of the AC source. This circuit is famously used in applications like cathode-ray tube (CRT) power supplies, laser systems, and particle accelerators, where it can generate kilovolts from a lower-voltage AC transformer.

Common Pitfalls

1. Ignoring Output Ripple and Regulation: A charge pump's output is not a clean DC voltage; it has inherent ripple caused by the periodic charge/discharge cycle. Under a heavy load, this ripple increases and the average output voltage sags. Assuming a charge pump provides a rock-solid voltage like a linear regulator is a mistake. The correction is to always size the output capacitor appropriately for your load's ripple tolerance and to understand that a post-regulation low-dropout (LDO) regulator is often necessary for sensitive analog circuits.

1. Overestimating Output Current Capability: Charge pumps are not high-power devices. Their maximum output current is limited by the size of the flying capacitors, the switching frequency, and the resistance of the switches. Attempting to draw too much current causes excessive voltage droop, overheating, and failure. Always consult the device's output current vs. voltage curve and derate appropriately. For higher currents, an inductor-based switching regulator is the correct choice.

1. Neglecting Switch Losses and Clock Drive: The theoretical efficiency of a charge pump is high, but real-world losses in the switches (MOSFET $R_{DS(on)}$) and the power required to drive the switching clocks at high frequency can reduce actual efficiency to 70-85%. Using poor external capacitors with high Equivalent Series Resistance (ESR) exacerbates this. The correction is to select low-ESR capacitors and, when using discrete designs, to ensure the clock driver can switch the gates of the MOSFETs quickly and with sufficient drive strength.

1. Misapplying the Topology: Charge pumps excel at low-power (< 100mA) conversion and inversion. Using them for high-current main power rail generation is a fundamental design error. Similarly, while Cockcroft-Walton multipliers generate high voltage, they provide very low current and have poor voltage regulation under load. Match the circuit's inherent strengths and weaknesses to the application's requirements.

Summary

• Charge pump circuits use switched capacitors to transfer charge, enabling DC voltage multiplication or inversion without magnetic components, favoring integration and miniaturization.
• Basic implementations include the voltage doubler and inverter, which form the core building blocks for more complex cascaded designs like the Dickson charge pump for on-chip DC voltage generation.
• For high-voltage AC-to-DC conversion, the Cockcroft-Walton multiplier uses a ladder of diodes and capacitors to stack the peak input voltage across multiple stages.
• Key limitations include inherent output ripple, limited current sourcing capability, and efficiency losses from switch resistance and capacitor ESR, making them unsuitable for high-power applications.
• These circuits are ubiquitous in portable and integrated electronics for tasks like generating programming voltages for memory, biasing LCD displays, and providing negative supply rails for analog components.